CN101137070A - Video input equipment gamma characteristic correcting method and apparatus in video communication - Google Patents

Abstract

The method includes operations:(1) Divide the RGB color space (RCS) into multi-cells, (2) linearize the gamma feature function (GFF) of the video input device (VID) cell by cell, (3) correspondingly map all cells of RCS into the YUV color space (YCS) to convert YCS into cells, (4) build linear GFF having unified format and different parameters, (5) use linear GFF and linear gamma correct function (LGCF) to relate the corrected YCS vector with the non-distortion RGB color vector (NRCV) to ensure the related vector and NRCV in the same signal space, (6) use the related vector and NRCV to set a mathematic optimizing subject, (7) solve this subject and obtain parameters of LGCF of all cells of YCS, (8) finally cell by cell gamma-correct the video signal of YCS output from VID according to the gamma correct function. This invention reaches the aims about the gamma correct: easy to use, extensive application range and low cost.

Description

Video input apparatus gamma characteristic correcting method and device in the video communication

Technical field

The present invention relates to the video communication technical field, be specifically related to video input apparatus gamma characteristic correcting method and device in a kind of video communication.

Background technology

Video communication, especially multi-party video communication at present along with the developing rapidly of broadband network, obtain increasingly extensive application.At home and in the world, video conference and visual telephone service are becoming the basic service on the NGN (Next Generation Network, next generation network).The telecom operators of various countries also pay much attention to this market opportunity.Be expected in the coming years, the video communication business will become the important business growth point of operator.

A key issue of development video communication business is: the user experience (UserExperience or Quality of Experience) that improves end-to-end (End-to-end).In the user experience except parameters such as the QoS of network such as packet loss, delay, shake, the R factor, for video, because the distortion to luminance signal (Distortion) that Gamma (gamma) nonlinear problem that each link causes causes also is to influence the key factor that the end user experiences.

At present, mainly concentrate on assurance network QoS pre-process and post-process (Pre-processing, post-processing) aspect relevant for method that improves end-to-end user experience and technology with video compression coding.For the brightness distortion problem that the Gamma characteristic causes, lack the solution of concern and system.But the importance of this problem has caused the concern of some international big telecom operators.

It is the ubiquitous problem that needs solution in the multi-media communication information system that Gamma proofreaies and correct.On root, Gamma in the video communication proofreaies and correct and originates from CRT (Cathode-Ray Tube cathode ray tube) display, because between the output light signal brightness of CRT monitor and the input voltage pumping signal amplitude is not linear dependence, but a kind of non-linear relation, i.e. Gamma characterisitic function relation.In order on video display apparatus, to reach high-quality display effect, obtain good user experience (UserExperience or Quality of Experience), must carry out Gamma and proofread and correct.

The universal model of single link Gamma characteristic as shown in Figure 1.

Among Fig. 1, the nonlinear relation of input luminance signal and output luminance signal can be expressed as: L _Out=G (L _In), wherein, L _OutBe output luminance signal, L _InBe input luminance signal, function G (.) is a nonlinear function.

Typical Gamma characteristic example as shown in Figure 2.

The numeral that marks in each square among Fig. 2 is a brightness value, and the gray scale of square is represented the bright degree of luminance signal.Among Fig. 2 (a), above the brightness of delegation's gray scale square be linear increment, promptly be incremented to 1.0 from 0.1, below the brightness of delegation's gray scale square increase progressively according to the power function rule, that is to say, below the brightness of delegation's gray scale square passed through the nonlinear distortion effect of Gamma.What provide among Fig. 2 (b) is Gamma characteristic with curve representation.

In actual applications, the Gamma characteristic of video input apparatus such as video camera/camera is:

$L_{\mathrm{out}}={L_{\mathrm{in}}}^{0.45}---\left(1\right)$

The form of this Gamma characteristic is a power function (Power Function).Of particular note, the input and output luminance signal here all is to have carried out normalization (Normalized), i.e. 0≤L in coordinate space separately _Out≤ 1,0≤L _In≤ 1.

The Gamma problem of proofreading and correct can be abstract for as shown in Figure 3.

Fig. 3 has increased a Gamma correction module on the basis of Fig. 2, the Gamma property list of Gamma correction module is shown Gc (.), and it is the distortion that Gg (.) causes that this Gamma correction module can be proofreaied and correct by the Gamma characteristic.For a Gamma link such as video input apparatus, its Gamma characteristic is given to be L _Out=Gp (L _In), like this, can use another one Gamma correction link L _Out=Gc (L _In) and it carry out cascade, make final Gamma characteristic input/output relation become real linear relationship.

In the multimedia video communication process, the transmission course of vision signal in communication network as shown in Figure 4.

Among Fig. 4, the light/power conversion device in the video input apparatus is converted to the signal of telecommunication/digital signal with video input apparatus at the light signal that the RGB color space collects, and output.The space of the signal of telecommunication/digital signal of light/power conversion device output is the RGB color space, and the color transform module in the video input apparatus is converted to the signal of telecommunication/digital signal of YUV color space with the signal of telecommunication/digital signal of RGB color space, and exports.The signal of telecommunication of YUV color space/digital signal transfers to the other side's video communication terminal through compressed encoding and network delivery module by communication network.

At present, the method that Gamma proofreaies and correct mainly contains two kinds, below in conjunction with accompanying drawing 4, accompanying drawing 5, accompanying drawing 6 shortcoming of these two kinds of Gamma bearing calibrations and existence is separately described.

Method one: carry out Gamma at the RGB color space and proofread and correct.

Carry out the Gamma timing at the RGB color space, the Gamma characterisitic function is three Gamma characterisitic functions of respectively corresponding three passages in fact, that is to say, and R, G, the B three primary colors each have Gamma characterisitic function separately, that is:

r _d＝gm _r(r _r)；

g _d＝gm _g(g _r)；(2)

b _d＝gm _b(b _r)；

Wherein, gm _rGm _gGm _bRepresent R respectively, B, G component Gamma characterisitic function separately.r _dExpression is through the R component signal of Gamma distortion, r _rRepresent original R component signal; g _dExpression is through the G component signal of Gamma distortion, g _dRepresent original G component signal; b _dExpression is through the B component signal of Gamma distortion, b _rRepresent original G component signal; Subscript r represents raw (original), and subscript d represents distorted (distortion).

Fig. 5 shows the Gamma characteristic curve of redness, green and blue component signals correspondence, i.e. function gm _r, gm _g, gm _bCurve.As can be seen from Figure 6, R, G, each self-corresponding Gamma characterisitic function of B, though very similar in form, the concrete parameter difference of each Gamma characterisitic function, therefore, these three Gamma characterisitic functions are not a function.If the employing vector form is represented the RGB component of not process Gamma distortion and has been passed through the RGB component of Gamma distortion, can obtain:

$c_{\mathrm{rRGB}}=\left[\begin{array}{c}{r}_r\\ g_r\\ b_r\end{array}\right],$ $c_{\mathrm{dRGB}}=\left[\begin{array}{c}{r}_d\\ g_d\\ b_d\end{array}\right].$

Because proofreading and correct, the Gamma of each passage of RBG component is independent of other passage, so, in the RGB color space, directly to carry out Gamma and proofread and correct, its implementation procedure is fairly simple.

From Fig. 4, Fig. 5 as can be known, between video display apparatus and video input apparatus, vision signal all exists with the YUV form, so, if in the RGB color space, proofread and correct, just mean that the Gamma correction module must be arranged in video display apparatus or video input apparatus.

For video input apparatus, because popular communicating requirement video input apparatus is cheap, therefore, often can only select the low side video input apparatus for use, and not having Gamma, the low side video input apparatus do not proofread and correct this Premium Features.

For video display apparatus, generally all do not carry out the function that Gamma proofreaies and correct voluntarily.It all is to rely on to introduce in the video input apparatus and Gamma characteristic that video display apparatus reverses mutually that early stage Gamma proofreaies and correct, and allows both cancel out each other, thereby reaches the purpose of Gamma correction.Early stage video display apparatus all is a bussiness class, and as video display apparatus of using in the broadcast television industry etc., the video display apparatus of these bussiness class belongs to high-end devices, and certain Gamma calibration function can accurately be provided.But along with extensively quoting of the popularizing of popular multimedia communication, a large amount of low side devices, most video input apparatus does not possess the ability that this Gamma proofreaies and correct.Like this, under a lot of practical situations, can not accomplish in the RGB color space, to carry out Gamma and proofread and correct.

Method two: carry out Gamma at the YUV color space and proofread and correct.

Carry out principle that Gamma proofreaies and correct as shown in Figure 6 at the YUV color space.As can be seen from Figure 6, proofread and correct if carry out Gamma at the YUV color space, then the Gamma correction module must be arranged between video input apparatus and the video display apparatus certain a bit on.

The concrete trimming process of the Gamma correction module among Fig. 6 is: at first, the YUV color space is carried out an inverse transformation, get back to the RGB color space, then, carry out Gamma at the RGB color space and proofread and correct, again through a direct transform, conversion is got back in the YUV color space after the correction.

In method two, the rounding error that Gamma correction meeting causes owing to computing causes the loss of vision signal, and, a large amount of multiplyings, its expense is very big, and especially the consumption aspect the software realization is very big, and is also more much bigger than the consumption that the Gamma correction self adopts look-up table to proofread and correct.

Problems such as from foregoing description as can be known, present Gamma bearing calibration has the cost of realization height, the scope of application is poor, resource consumption is serious.

Summary of the invention

The objective of the invention is to, video input apparatus gamma characteristic correcting method and device in a kind of video communication are provided, realized directly the vision signal of video input apparatus output being carried out the purpose of Gamma correction at the YUV color space, simplified the implementation procedure of Gamma correction, reduce the realization cost of Gamma correction, improved the ease for use of Gamma correction.

For achieving the above object, the invention provides video input apparatus gamma characteristic correcting method in a kind of video communication, comprising:

A, the RGB color space is divided into a plurality of cell elements;

B, according to the gamma characteristic function of gamma link in the RGB color space that video input apparatus comprised, obtain the linearisation of described gamma characteristic function on each cell element of RGB color space and represent, promptly obtain the linear gamma characteristic function of each cell element;

C, each cell element of RGB color space is mapped to the YUV color space one to one, makes the YUV color space be divided into a plurality of cell elements;

D, the linear Gamma correction function parameters representation of each cell element in the YUV color space is set;

E, utilize described linear gamma characteristic function and linear Gamma correction function to the vector sum of proofreading and correct the YUV color space not distortion RGB color vector carry out related, make the vector after the association be positioned at same signal space with distortion RGB color vector not, utilize the vector sum after the described association based on average generalized distance minimum criteria between vector the mathematical optimization problem not to be set between the distortion RGB color vector, find the solution this mathematical optimization problem, to determine the linear Gamma correction function parameters on each cell element of YUV color space;

F, according to the linear Gamma correction function on each cell element of YUV color space the yuv video signal of video input apparatus output is pursued the Gamma correction of cell element.

The technical scheme of following method is optional technical scheme.

The RGB color space that described RGB color space is normalization, described YUV color space are normalization RGB color space by the YUV color space that described color notation conversion space generated.

Described step a comprises:

With R axle five equilibrium with obtain each cell element the length of side on the R axle or with the R axle not five equilibrium to obtain the length of side of each cell element on the R axle;

With G axle five equilibrium with obtain each cell element the length of side on the G axle or with the G axle not five equilibrium to obtain the length of side of each cell element on the G axle;

With B axle five equilibrium with obtain each cell element the length of side on the B axle or with the B axle not five equilibrium to obtain the length of side of each cell element on the B axle.

Described step b comprises:

The gamma characteristic function of gamma link in the RGB color space of supposing video input apparatus is known, and each division points based on cell element on the R axle is er _i, the R axle is divided into N _RIndividual interval, i=1,2,3 ...., N _R, each division points based on cell element on the G axle is eg _j, the G axle is divided into N _GIndividual interval, j=1,2,3 ...., N _G, each division points based on cell element on the B axle is eb _k, the B axle is divided into N _BIndividual interval, k=1,2,3 ...., N _B, then for the (i, j, k) individual cell element, the mathematical optimization problem of setting up according to the least mean-square error optiaml ciriterion is:

$\{k_r\left(i\right),b_r\left(i\right),i=\mathrm{1,2,3},....,N_R\}=\arg \underset{i=\mathrm{1,2,3},....,N_R}{\underset{k_r\left(i\right),b_r\left(i\right)\∈R,}{\min }}\underset{i=1}{\overset{N_R}{\mathrm{\Σ}}}{\∫}_{{\mathrm{er}}_{i-1}}^{{\mathrm{er}}_i}{\left|{\mathrm{gm}}_r\left(r_r\right)-k_r\left(i\right)r_r-b_r\left(i\right))\right|}^2{\mathrm{dr}}_r$

$s.t.\left\{\begin{array}{c}{k}_r\left(1\right){\mathrm{er}}_1+b_r\left(1\right)=k_r\left(2\right){\mathrm{er}}_1+b_r\left(2\right)\\ k_r\left(2\right){\mathrm{er}}_2+b_r\left(2\right)=k_r\left(3\right){\mathrm{er}}_2+b_r\left(3\right)\\ \·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\·\\ k_r(N_R-1){\mathrm{er}}_{N_{R-1}}+b_r(N_R-1)=k_r\left(N_R\right){\mathrm{er}}_{N_{R-1}}+b_r\left(N_R\right)\\ k_r\left(N_R\right)+b_r\left(N_R\right)=1\end{array}\right.;$

$\{k_g\left(j\right),b_g\left(j\right),j=\mathrm{1,2,3},....,N_G\}=\arg \underset{j=\mathrm{1,2,3},....,N_G}{\underset{k_g\left(j\right),b_g\left(j\right)\∈R,}{\min }}\underset{j=1}{\overset{N_G}{\mathrm{\Σ}}}{\∫}_{{\mathrm{eg}}_{j-1}}^{{\mathrm{eg}}_j}{\left|{\mathrm{gm}}_g\left(g_r\right)-k_g\left(j\right)g_r-b_g\left(j\right))\right|}^2{\mathrm{dg}}_r$

$s.t.\left\{\begin{array}{c}{k}_g\left(1\right){\mathrm{<figure-callout\; id="1"\; label="eg"\; filenames="A20061011157900331.png"\; state="\{\{state\}\}">eg</figure-callout>}}_1+b_g\left(1\right)=k_g\left(2\right){\mathrm{<figure-callout\; id="1"\; label="eg"\; filenames="A20061011157900331.png"\; state="\{\{state\}\}">eg</figure-callout>}}_1+b_g\left(2\right)\\ k_g\left(2\right){\mathrm{eg}}_2+b_g\left(2\right)=k_g\left(3\right){\mathrm{eg}}_2+b_g\left(3\right)\\ \&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\\ k_g(N_G-1){\mathrm{eG}}_{N_{G-1}}+b_g(N_G-1)=k_g\left(N_G\right){\mathrm{eg}}_{N_{G-1}}+b_G\left(N_G\right)\\ k_g\left(N_G\right)+b_g\left(N_G\right)=1\end{array}\right.;$

$\{k_b\left(k\right),b_b\left(k\right),k=\mathrm{1,2,3},....,N_B\}=\arg \underset{i=\mathrm{1,2,3},....,N_B}{\underset{k_b\left(k\right),b_b\left(k\right)\&Element;R,}{\min }}\underset{k=1}{\overset{N_B}{\mathrm{\&Sigma;}}}{\&Integral;}_{{\mathrm{eb}}_{k-1}}^{{\mathrm{eb}}_k}{\left|{\mathrm{gm}}_b\left(b_r\right)-k_b\left(k\right)b_r-b_b\left(k\right))\right|}^2{\mathrm{dr}}_r$

$s.t.\left\{\begin{array}{c}{k}_b\left(1\right){\mathrm{<figure-callout\; id="1"\; label="eb"\; filenames="A20061011157900331.png"\; state="\{\{state\}\}">eb</figure-callout>}}_1+b_b\left(1\right)=k_b\left(2\right){\mathrm{<figure-callout\; id="1"\; label="eb"\; filenames="A20061011157900331.png"\; state="\{\{state\}\}">eb</figure-callout>}}_1+b_b\left(2\right)\\ k_b\left(2\right){\mathrm{eb}}_2+b_b\left(2\right)=k_b\left(3\right){\mathrm{eb}}_2+b_b\left(3\right)\\ \&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\\ k_b(N_B-1){\mathrm{eb}}_{N_{B-1}}+b_b(N_B-1)=k_b\left(N_B\right){\mathrm{eb}}_{N_{B-1}}+b_b\left(N_B\right)\\ k_b\left(N_B\right)+b_b\left(N_B\right)=1\end{array}\right.;$

Wherein: s.t. represents constraints, k _r(i) and b _r(i) be (i, j, k) the R component gamma parameter of the linear gamma characteristic function of individual cell element, the k of RGB color space _g(j) and b _g(j) be (i, j, k) the G component gamma parameter of the linear gamma characteristic function of individual cell element of RGB color space; k _b(k) and b _b(k) be (i, j, k) the B component gamma parameter of the linear gamma characteristic function of individual cell element of RGB color space;

Above-mentioned mathematical optimization problem is found the solution, obtain parameter matrix K (i, j, k) and parameter vector B (i, j, k);

Repeat the aforementioned calculation process for each cell element of RGB color space, to determine the linear gamma characteristic function parameters on each cell element, the linear gamma characteristic function of each cell element of RGB color space is:

$c_{\mathrm{dRGB}}=K(i,j,k)c_{\mathrm{rRGB}}+B(i,j,k)=\left[\begin{array}{ccc}{k}_r\left(i\right)& 0& 0\\ 0& k_g\left(j\right)& 0\\ 0& 0& k_b\left(k\right)\end{array}\right]c_{\mathrm{rRGB}}+\left[\begin{array}{c}{b}_r\left(i\right)\\ b_g\left(j\right)\\ b_b\left(k\right)\end{array}\right];$

Wherein: i, j, k are cell element CER _GBCall number, c _DRGBBe the RGB color vector through the distortion of gamma link, c _RRGBFor not saving the RGB color vector of distortion through gamma loop.

Among the described step c, each cell element CE in the YUV color space _YUV(the limit k) coordinate range on Y, U, V direction respectively is for i, j:

0.299fr _i-1+0.587fg _j-1+0.114fb _k-1＝y _min(i，j，k)≤y≤y _max(i，j，k)＝0.299fr _i+0.587fg _j+0.114fb _k；-0.147fr _i-0.289fg _j+0.436fb _k-1＝u _min(i，j，k)≤u≤u _max(i，j，k)＝-0.147fr _i-1-0.289fg _j-1+0.436fb _k；0.615fr _i-1-0.515fg _j-0.1fb _k＝v _min(i，j，k)≤v≤v _max(i，j，k)＝0.615fr _i-0.515fg _j-1-0.1fb _k-1；

Wherein: fr _i, fg _j, fb _kBe the cell element CE in the RGB color space _RGB(i, j, k) boundary coordinate er _i, eg _j, eb _kPass through cell element CE respectively _RGB(i, j, k) numerical value after the last linear gamma characteristic function distortion, i.e. fr _i, fg _j, fb _kBe the point on the gamma characteristic function curve; Or described fr _i, fg _j, fb _kIt is not the point on the gamma characteristic function curve.

Step related among the described step e comprises:

Linear Gamma correction function in linear gamma characteristic function in the setting RGB color space on each cell element, the YUV color space on each cell element is respectively:

c _dRGB＝K(i，j，k)c _rRGB+B(i，j，k)，c _cYUV＝P(i，j，k)c _uYUV+Q(i，j，k)；

Wherein: c _DRGBBe the RGB color vector through the distortion of gamma link, c _RRGBFor not through the RGB color vector of gamma loop joint distortion, K (i, j, k) and B (i, j k) are each cell element linearity gamma characteristic function parameters in the RGB color space, c _CYUVBe the YUV color vector of proofreading and correct, c through the gamma link _UYUVBe the YUV color vector of proofreading and correct through the gamma loop joint, P (i, j, k) and Q (i, j k) are each cell element linearity Gamma correction function parameters of YUV color space;

Utilize linear gamma characteristic function with c _RRGBWith the YUV color vector c that does not carry out Gamma correction _URGBCarry out association:

c _uYUV＝T(K(i，j，k)c _rRGB+B(i，j，k))；

Wherein: T represents the direct transform from the RGB color space to the YUV color space:

Utilize linear Gamma correction function will proofread and correct YUV color vector and c _RGBCarry out association:

C _cYUV＝P(i，j，k)T(K(i，j，k)c _rRGB+B(i，j，k))+Q(i，j，k)；

Wherein: P (j, j, k) and Q (i, j k) are linear Gamma correction function parameters on each cell element of YUV color space, i, j, k are cell element CE _YUVCall number, i=1,2,3 ...., N _R, j=1,2,3 ...., N _G, k=1,2,3 ...., N _B, N _RBe the interval number that forms owing to the cell element division on the R axle, N _GBe the interval number that forms owing to the cell element division on the G axle, N _BBe the interval number that forms owing to the cell element division on the B axle.

The step that the mathematical optimization problem is set among the described step e and finds the solution comprises:

In the accumulation of two vector errors described in the distortion RGB color space not be:

$\mathrm{MSE}={\&Integral;}_{c_{\mathrm{rRGB}\&Element;{\left[\mathrm{0,1}\right]}^3}}{(c_{\mathrm{cYUV}}-T{c}_{\mathrm{rRGB}})}^T(c_{\mathrm{cYUV}}-T{c}_{\mathrm{rRGB}})d{c}_{\mathrm{rRGB}};$

Wherein: MSE represents mean square error;

In conjunction with c _CYUV=P (i, j, k) T (K (i, j, k) c _RRGB+ B (i, j, k))+Q (i, j k), draw:

$\mathrm{MSE}=$

${\&Integral;}_{c_{\mathrm{cRGB}\&Element;{\&lsqb;0.1\&rsqb;}^3}}{(P(i,j.k)T(K(i,j,k)c_{\mathrm{rRGB}}+B(i,j,k))+Q(i,j,k)-T{c}_{\mathrm{rRGB}})}^T$

$(P(i,j,k)T(K(i,j,k)c_{\mathrm{rRGB}}+B(i,j,k))Q(i,j,k)-T{c}_{\mathrm{rRGB}})d{c}_{\mathrm{rRGB}}$

Thereby for cell element CE _YUY(i, j, k) i=1,2,3 ...., N _R, j=1,2,3 ...., N _G, k=1,2,3 ...., N _B, obtain following mathematical optimization problem:

$\{P_{\mathrm{Opt}}(i,j,k),Q_{\mathrm{Opt}}(i,j,k)\}=$

$\underset{P(i,j,k),Q(i,j,k)}{\arg \min }{\&Integral;}_{c_{\mathrm{cRGB}\&Element;{\&lsqb;0.1\&rsqb;}^3}}{(P(i,j.k)T(K(i,j,k)c_{\mathrm{rRGB}}+B(i,j,k))+Q(i,j,k)-T{c}_{\mathrm{rRGB}})}^T$

$(P(i,j,k)T(K(i,j,k)c_{\mathrm{rRGB}}+B(i,j,k))Q(i,j,k)-T{c}_{\mathrm{rRGB}})d{c}_{\mathrm{rRGB}}$

Wherein: P _Opt(i, j, k), Q _Opt(i, j k) are the globally optimal solution of described mathematical optimization problem;

The described globally optimal solution P of described mathematical optimization problem _Opt(i, j, k), Q _Opt(i, j k) are exactly cell element CE in the YUV color space _YUV(i, j, linear Gamma correction function parameters k).

Described step f comprises:

The cell element of each sampled value of yuv video signal of determining video input apparatus output under in the YUV color space;

Each sampled value affiliated cell element in the YUV color space according to described vision signal is determined the Gamma correction function that each sampled value is corresponding respectively;

According to determined Gamma correction function each sampled value of described vision signal is carried out Gamma correction.

The present invention should provide video input apparatus gamma characteristic means for correcting in a kind of video communication, and described device comprises: memory module, linearization block, cell element module, mathematical optimization module, find the solution module and Gamma correction module;

Memory module: be used for receiving and storing the positional information of positional information, each cell element of YUV color space of each cell element of RGB color space, gamma link that video input apparatus comprised in the gamma characteristic function of RGB color space and corresponding respectively linear Gamma correction function parameters representation and the linear Gamma correction function parameters of each cell element in the YUV color space;

Linearization block: be used for calculating the parameter that the linearisation of gamma characteristic function on each cell element of RGB color space of the gamma link that video input apparatus comprised represented according to the memory module canned data, promptly obtain the linear gamma characteristic function of each cell element, and output result of calculation is stored to memory module;

The cell element module: the positional information according to each cell element of RGB color space of storing in the memory module is mapped to the YUV color space one to one with each cell element in the RGB color space, obtain the positional information of each cell element of YUV color space, and transfer to memory module and store;

The mathematical optimization module: utilize the linear gamma characteristic function stored in the memory module and linear Gamma correction function to the vector sum of proofreading and correct the YUV color space not distortion RGB color vector carry out related, make itself and distortion RGB color vector not be positioned at same signal space, utilize the vector sum after the described association based on average generalized distance minimum criteria between vector the mathematical optimization problem not to be set between the distortion RGB color vector;

Find the solution module: find the solution described mathematical optimization problem, determining the linear Gamma correction function parameters on each cell element in the YUV color space, and transfer to memory module and store;

The Gamma correction module: the positional information of each cell element in linear Gamma correction function parameters representation, linear Gamma correction function parameters and the YUV color space that is used for storing according to memory module, the yuv video signal of video input apparatus output is pursued the Gamma correction of cell element.

The RGB color space that described RGB color space is normalization; Described YUV color space is a normalization RGB color space by the YUV color space that described color notation conversion space generated.

Description by technique scheme as can be known, gamma revision method provided by the invention can directly carry out Gamma correction at the YUV color space to the vision signal of input equipment output, solved video input apparatus under the situation that does not possess the Gamma correction ability, directly carry out the problem of Gamma correction at the YUV color space, the present invention can be good at being applicable in the video communication process of low side video input apparatus; When the present invention realizes Gamma correction at the YUV color space, avoid rounding error, a large amount of problems such as resource consumption, simplified the implementation procedure of Gamma correction; Thereby reached raising Gamma correction ease for use by technical scheme provided by the invention, widened the range of application of Gamma correction, reduced the purpose of the realization cost of Gamma correction.

Description of drawings

Fig. 1 is the model schematic diagram of link Gamma characteristic;

Fig. 2 (a) is a Gamma characteristic schematic diagram one;

Fig. 2 (b) is a Gamma characteristic schematic diagram two;

Fig. 3 is the Gamma correction principle schematic diagram to a Gamma link;

Fig. 4 is the transmission course schematic diagram of vision signal in communication network;

Fig. 5 is the Gamma characteristic curve schematic diagram of redness, green and blue component signals correspondence;

Fig. 6 carries out the schematic diagram that Gamma proofreaies and correct at the YUV color space;

Fig. 7 is that the vision signal at video input apparatus flows to schematic diagram;

Fig. 8 is the schematic diagram that the RGB color space is divided into a plurality of cell elements of the embodiment of the invention;

Fig. 9 be the embodiment of the invention approach the geometric meaning schematic diagram of the Gamma characterisitic function of R component with broken line;

Figure 10 is the schematic diagram that the YUV color space is divided into a plurality of cell elements of the embodiment of the invention;

Figure 11 is the schematic diagram one that the Gamma correcting integrated circuit chip of the embodiment of the invention is realized YUV color space Gamma correction;

Figure 12 is the schematic diagram that the Gamma correcting integrated circuit chip of the embodiment of the invention is realized YUV color space Gamma correction

Embodiment

Directly the vision signal in the video image of video input apparatus output is carried out the Gamma timing at the YUV color space, y to vision signal, u, the Gamma that the v signal component is carried out proofreaies and correct all can relate to r, g, three signal components of b, like this, Y, U, it is not separate that the Gamma of three passages of V proofreaies and correct, but close-coupled together, and this has improved greatly from r, g, the difficulty of the correction function of Gamma correction is directly carried out in acquisition in the b signal component Gamma characteristic separately in the YUV color space, thereby improved at the YUV color space vision signal of video input apparatus output is carried out the difficulty that Gamma proofreaies and correct.

The present invention provides a kind of effective and feasible technical scheme for directly the vision signal in the video image of video input apparatus output being carried out the Gamma correction at the YUV color space.The present invention is directed to schematic diagram that video input apparatus carries out the Gamma Gamma correction as shown in Figure 6.As can be seen from Figure 6, the present invention directly proofreaies and correct the gamma link that video input apparatus comprised at the YUV color space.

When the present invention carried out Gamma correction at the YUV color space to the gamma link that video input apparatus comprised, the flow direction of vision signal as shown in Figure 7.

Among Fig. 7, the present invention carries out the process of Gamma correction at the YUV color space to the gamma link of video input apparatus, relates to four kinds of vision signal spaces altogether, and according to the flow direction of vision signal, these four kinds of vision signal spaces are followed successively by:

1, undistorted RGB color space, the vector representation of vision signal in undistorted RGB color space is c _RRGB.

2, the RGB color space of distortion has promptly passed through the RGB color space of Gamma link, and the vector representation of vision signal in the RGB of distortion color space is c _DRGB

3, uncorrected YUV color space, the vector representation of vision signal in uncorrected YUV color space is c _UYUV

4, carried out the YUV color space of Gamma correction, the vector representation of vision signal in the YUV color space that has carried out Gamma correction is c _CYUV

Wherein: _rRGB represents raw RGB, _dRGB represents distorted RGB, _uYUV represents uncorrected YUV, _cYUV represents corrected YUV.

Below the bearing calibration to the gamma characteristic that video input apparatus comprised provided by the invention is elaborated.

In technical scheme provided by the invention, the Gamma characterisitic function of the Gamma link that video input apparatus comprised in the RGB color space is known.The Gamma characterisitic function of the Gamma link of video input apparatus in the RGB color space can adopt present apparatus measures method to obtain, also can adopt non-blind method to obtain based on input and output luminance signal the A to Z of calculating Gamma characterisitic function, also can adopt the semi-blind method of calculating the Gamma characterisitic function based on the part knowledge and the output luminance signal of input luminance signal to obtain, can also adopt total blindness's method to obtain based on output luminance signal calculating Gamma characterisitic function.Preparation method for the Gamma characterisitic function of Gamma link in the RGB color space of video input apparatus is not described in detail in the present invention.

The present invention may further comprise the steps in the bearing calibration that the YUV color space carries out the Gamma characteristic:

Step 1, the RGB color space is divided into a plurality of cell elements, and, makes all corresponding linear Gamma characterisitic function of each cell element the Gamma characterisitic function subregion linear expression of RGB color space.

RGB color space among the present invention can represent that promptly in the RGB color space, the absolute value of r, g, b component is not more than 1 by the RGB color space of normalization, and r, g, b all get on the occasion of, i.e. 0≤r, g, b≤1.Adopt the method for expressing of normalization to represent that the RGB color space has a lot of benefits, in each processing links of vision signal process, concrete brightness or carrier chrominance signal grade can be different, as 256,64 grades etc., but, vision signal with different brackets can change to [1 by brightness or the carrier chrominance signal highest ranking divided by each processing links, 1] in the interval, like this, the RGB color space of the vision signal at different disposal link place just and the concrete brightness of processing links, the number of degrees of carrier chrominance signal had nothing to do.

If the RGB color space has been carried out the normalization processing, then the RGB color space is exactly a unit cube, i.e. [0,1] * [0,1] * [0,1] perhaps is abbreviated as [0,1] ³, wherein: the cartesian product (CartesianProduct) of multiplication sign " * " expression set.

In the following description, the RGB color space is the RGB color space of normalization.

In 3 dimension RGB color spaces, the present invention need be divided into the RGB color space a plurality of non-overlapping little cuboids, and each little cuboid is called cell element (cell).Concrete dividing mode as shown in Figure 11.

Among Figure 11, the unit interval on the R axle [0,1] is divided into N _RIndividual subinterval, division points are respectively 0=er ₀＜er ₁＜er ₂＜er _N-1＜er _N=1, like this, each subinterval on the R axle is respectively: [er ₀, er ₁], [er ₁, er ₂] ..., [er _I-1, er _i] ... [er _NR-1, er _NR], wherein: i subinterval note made SIR (i)=[er _I-1, er _i], i=1,2 ...., N _RoSIR represents English Sub-Interval of Red (red subinterval).

Need to prove, above-mentioned the R axle is divided into N _RThe mode in individual subinterval can be an any-mode, as five equilibrium, do not wait and grade.In like manner, according to the dividing mode of the unit interval on the R axle [0,1] subinterval is also carried out in the unit interval [0,1] on G, the B axle and divide, the division result on G, the B axle is as follows:

G axle unit interval [0,1] is divided into N _GIndividual subinterval, each subinterval on the G axle is respectively: [eg ₀, eg ₁], [eg ₁, eg ₂] ..., [eg _J-1, eg _j] ... [eg _NG-1, eg _NG], wherein: j subinterval note made SIG (j)=[eg _J-1, eg _j], j=1,2 ..., N _GSIG represents English Sub-Interval of Green (green subinterval).

B axle unit interval [0,1] is divided into N _BIndividual subinterval, each subinterval on the B axle is respectively: [eb ₀, eb ₁], [eb ₁, eb ₂] ..., [eb _K-1, eb _k] ..., [eb _NB-1, eb _NB], wherein: k subinterval note made SIB (k)=[eb _K-1, eb _k], k=1,2 ..., N _BSIB represents English Sub-Interval of Blue (blue subinterval).

Positive integer i in the foregoing description, j, k can be called the call number in subinterval.

Division based on this subinterval, the present invention can carry out cell element to the RGB color space and divide, each cell element is actually a three-dimensional cuboid, and this cuboid is at R, G, respectively corresponding subinterval, limit on the B direction, cuboid is at R, G, and SIR (i) is used in the subinterval of the limit correspondence on the B direction respectively, SIG (j), SIB (k) represents.A cell element in the RGB color space can be remembered and makes CE _RGB(i, j k), promptly use a cell element at R, G, and the call number in corresponding respectively subinterval is united this cell element of expression, this cell element of index on three directions of B, the i of this moment, j, k are the call number of cell element.Obviously, the RGB color space is divided into N _R* N _G* N _BIndividual cell element.

After the division of having carried out above-mentioned cell element, need to introduce the linear expression mode of Gamma characterisitic function, promptly need to be provided with the linear Gamma characterisitic function of each cell element correspondence, that is to say, in each cell element of RGB color space, all use same linear function to represent the Gamma characterisitic function, this linear function is linear Gamma characterisitic function, the linear Gamma characterisitic function difference in the different cell elements.

The Gamma characterisitic function of setting the RGB color space is following form:

r _d＝gm _r(r _r)；

g _d＝gm _g(g _r)；(2)

b _d＝gm _b(b _r)；

In general, the Gamma characterisitic function all is a nonlinear function, for nonlinear function, in order to simplify processing, can adopt the strategy of piece-wise linearization.It all is that function is the Gamma characterisitic function that the strategy that the present invention adopts piece-wise linearization is converted into nonlinear Gamma characterisitic function linear.Below still with CE _RGB(i, j k) describe linear Gamma characterisitic function for example.

Set cell element CE _RGB(k) Dui Ying linear Gamma characterisitic function is expressed as form for i, j:

r _d＝gm _r(r _r)＝k _r(i)r _r+b _r(i)；

g _d＝gm _g(g _r)＝k _g(j)g _r+b _g(j)；(3)

b _d＝gm _b(b _r)＝k _b(k)b _r+b _b(k)；

Wherein: r _dBe through the r component in the RGB color space after the Gamma link i.e. r component after the distortion, r _rFor not passing through the r component in the RGB color space of Gamma link, promptly undistorted r components number, in like manner, g _dBe the g component after the distortion, g _rBe undistorted g component, b _dBe the b component after the distortion, b _rBe undistorted b component; k _r(i), k _g(j), k _b(k), b _r(i), b _g(j), b _b(k) be cell element CE _RGB(i, j, the parameter of linear Gamma characterisitic function k).

Being write formula (3) as matrix form is:

$c_{\mathrm{dRGB}}=K(i,j,k)c_{\mathrm{rRGB}}+B(i,j,k)=\left[\begin{array}{ccc}{k}_r\left(i\right)& 0& 0\\ 0& k_g\left(j\right)& 0\\ 0& 0& k_b\left(k\right)\end{array}\right]c_{\mathrm{rRGB}}+\left[\begin{array}{c}{b}_r\left(i\right)\\ b_g\left(j\right)\\ b_b\left(k\right)\end{array}\right]---\left(4\right)$

Wherein: K (i, j k) are one 3 * 3 matrix, B (i, j k) are one 3 * 1 column vector, K (i, j, k) and B (i, j k) are cell element CE _RGB(i, j, the parameter of linear Gamma characterisitic function k), c _RRGBBe the vector of vision signal in undistorted RGB color space, c _DRGBBe the vector of vision signal in the RGB of distortion color space.

In the present invention, because the Gamma characterisitic function in the RGB color space is known, so the present invention can directly obtain the linear Gamma characterisitic function of each cell element in the RGB color space according to this known Gamma characterisitic function, employing piece-wise linearization.The method that nonlinear function is carried out piece-wise linearization has multiple, and belongs to prior art, is that example describes the linear Gamma characterisitic function that obtains each cell element in the RGB color space with the mean square error optiaml ciriterion only in the specific embodiment of the present invention.

Process based on the linear Gamma characterisitic function of R component of each cell element in the mean square error optiaml ciriterion acquisition RGB color space is as follows:

At first, the form of the linear Gamma characterisitic function of R component of each cell element is in the setting RGB color space:

Then, the present invention can determine the parameter k of formula (5) neutral line Gamma characterisitic function according to the mean square error optiaml ciriterion _r(i), b _r(i), i=1,2,3 ...., N _R, promptly followingly have the mathematical optimization problem of constraints to determine k by setting up _r(i), b _r(i), i=1,2,3 ..., N _R:

$\{k_r\left(i\right),b_r\left(i\right),i=\mathrm{1,2,3},....,N_R\}=\arg \underset{i=\mathrm{1,2,3},....,N_R}{\underset{k_r\left(i\right),b_r\left(i\right)\&Element;R,}{\min }}\underset{i=1}{\overset{N_R}{\mathrm{\&Sigma;}}}{\&Integral;}_{{\mathrm{er}}_{i-1}}^{{\mathrm{er}}_i}{\left|{\mathrm{gm}}_r\left(r_r\right)-k_r\left(i\right)r_r-b_r\left(i\right))\right|}^2{\mathrm{dr}}_r$

$s.t.\left\{\begin{array}{c}{k}_r\left(1\right){\mathrm{er}}_1+b_r\left(1\right)=k_r\left(2\right){\mathrm{er}}_1+b_r\left(2\right)\\ k_r\left(2\right){\mathrm{er}}_2+b_r\left(2\right)=k_r\left(3\right){\mathrm{er}}_2+b_r\left(3\right)\\ \&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\\ k_r(N_R-1){\mathrm{er}}_{N_{R-1}}+b_r(N_R-1)=k_r\left(N_R\right){\mathrm{er}}_{N_{R-1}}+b_r\left(N_R\right)\\ k_r\left(N_R\right)+b_r\left(N_R\right)=1\end{array}\right.---\left(6\right)$

Wherein: s.t. represents constraints, k _r(i) and b _r(i) be (i, j, k) the R component gamma parameter of the linear gamma characteristic function of individual cell element of RGB color space.

The meaning of the mathematic(al) representation of formula (6) is: under the optimum meaning of mean square error, and linear gamma characteristic function parameters k _r(i), b _r(i), i=1,2,3 ..., N _RShould make target function or cost function $\underset{i=1}{\overset{N_R}{\mathrm{\&Sigma;}}}{\&Integral;}_{{\mathrm{er}}_{i-1}}^{{\mathrm{er}}_i}{\left|{\mathrm{gm}}_r\left(r_r\right)-k_r\left(i\right)r_r-b_r\left(i\right))\right|}^2{\mathrm{dr}}_r$ Get minimum value.

Formula (6) is the analytical form of the Gamma characterisitic function gmr (.) of R color space, is a mathematical optimization problem that constraints is typically arranged, constraints with s.t. (Subject to, restricted on ...) symbol one set of equations of following later represents.

Constraints in the formula (6) has N _RIndividual, this N _RThe prerequisite that individual constraints is set up is: at the Gamma of RGB color space characterisitic function is continuous, and the Gamma characterisitic function satisfies gm _r(1)=1.These two preconditions are very loose, are very easy to satisfy.

The solution of mathematical optimization problem belongs to prior art, therefore, by the method for solving that prior art provides, can obtain separating of formula (6), promptly obtains parameter k _r(i), b _r(i), i=1,2,3 ...., N _R

In like manner, the present invention can be by finding the solution the mathematical optimization problem shown in following formula (7), (8), to obtain parameter k _g(j), b _g(j), j=1,2,3 ..., N _GAnd k _b(k), b _b(k), k=1,2,3 ...., N _B

$\{k_g\left(j\right),b_g\left(j\right),j=\mathrm{1,2,3},....,N_G\}=\arg \underset{j=\mathrm{1,2,3},....,N_G}{\underset{k_g\left(j\right),b_g\left(j\right)\&Element;R,}{\min }}\underset{j=1}{\overset{N_G}{\mathrm{\&Sigma;}}}{\&Integral;}_{{\mathrm{eg}}_{j-1}}^{{\mathrm{eg}}_j}{\left|{\mathrm{gm}}_g\left(g_r\right)-k_g\left(j\right)g_r-b_g\left(j\right))\right|}^2{\mathrm{dg}}_r$

$s.t.\left\{\begin{array}{c}{k}_g\left(1\right){\mathrm{<figure-callout\; id="1"\; label="eg"\; filenames="A20061011157900331.png"\; state="\{\{state\}\}">eg</figure-callout>}}_1+b_g\left(1\right)=k_g\left(2\right){\mathrm{<figure-callout\; id="1"\; label="eg"\; filenames="A20061011157900331.png"\; state="\{\{state\}\}">eg</figure-callout>}}_1+b_g\left(2\right)\\ k_g\left(2\right){\mathrm{eg}}_2+b_g\left(2\right)=k_g\left(3\right){\mathrm{eg}}_2+b_g\left(3\right)\\ \&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\\ k_g(N_G-1){\mathrm{eG}}_{N_{G-1}}+b_g(N_G-1)=k_g\left(N_G\right){\mathrm{eg}}_{N_{G-1}}+b_G\left(N_G\right)\\ k_g\left(N_G\right)+b_g\left(N_G\right)=1\end{array}\right.---\left(7\right)$

$\{k_b\left(k\right),b_b\left(k\right),k=\mathrm{1,2,3},....,N_B\}=\arg \underset{i=\mathrm{1,2,3},....,N_B}{\underset{k_b\left(k\right),b_b\left(k\right)\&Element;R,}{\min }}\underset{k=1}{\overset{N_B}{\mathrm{\&Sigma;}}}{\&Integral;}_{{\mathrm{eb}}_{k-1}}^{{\mathrm{eb}}_k}{\left|{\mathrm{gm}}_b\left(b_r\right)-k_b\left(k\right)b_r-b_b\left(k\right))\right|}^2{\mathrm{dr}}_r$

$s.t.\left\{\begin{array}{c}{k}_b\left(1\right){\mathrm{<figure-callout\; id="1"\; label="eb"\; filenames="A20061011157900331.png"\; state="\{\{state\}\}">eb</figure-callout>}}_1+b_b\left(1\right)=k_b\left(2\right){\mathrm{<figure-callout\; id="1"\; label="eb"\; filenames="A20061011157900331.png"\; state="\{\{state\}\}">eb</figure-callout>}}_1+b_b\left(2\right)\\ k_b\left(2\right){\mathrm{eb}}_2+b_b\left(2\right)=k_b\left(3\right){\mathrm{eb}}_2+b_b\left(3\right)\\ \&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\\ k_b(N_B-1){\mathrm{eb}}_{N_{B-1}}+b_b(N_B-1)=k_b\left(N_B\right){\mathrm{eb}}_{N_{B-1}}+b_b\left(N_B\right)\\ k_b\left(N_B\right)+b_b\left(N_B\right)=1\end{array}\right.---\left(8\right)$

Wherein: s.t. represents constraints, k _g(j) and b _g(j) be (i, j, k) the G component gamma parameter of the linear gamma characteristic function of individual cell element of RGB color space; k _b(k) and b _b(k) be (i, j, k) the B component gamma parameter of the linear gamma characteristic function of individual cell element of RGB color space.

By foregoing description as can be known, piecewise linearity is exactly to approach a continuous Gamma characteristic curve with one group of broken line.Approach with broken line the r component the Gamma characterisitic function geometric meaning as shown in Figure 8.

Curve among Fig. 8 is that the Gamma characterisitic function of R component is gm _r(r), the transverse axis among Fig. 8 is the R component of r axle for not influencing through the Gamma nonlinear distortion, and the longitudinal axis is gm _r(r) for having passed through the R component after the influence of Gamma nonlinear distortion.Comprised 5 division points in the r axle among Fig. 8, these 5 division points are respectively er ₀, er ₁, er ₂, er ₃, er ₄, these 5 division points are with gm _r(r) curve is divided into 4 sections, and the end points of line segment is respectively (er ₀, fr ₀), (er ₁, fr ₁), (er ₂, fr ₂), (er ₃, fr ₃), (er ₄, fr ₄).After these 4 sections line segments couple together, with gm _r(r) curve approaches very much.

The fr that 5 division points among Fig. 8 in the r axle are corresponding respectively ₀, fr ₁, fr ₂, fr ₃, fr ₄All are gm _r(r) point on the curve, in fact, the fr that the division points in the r axle is corresponding respectively _iAlso can be not at gm _r(r) on the curve.

As shown in Figure 8, for the r component, be set with N _RThe bar broken line just has N between these broken lines _R+ 1 end points, its coordinate is respectively (er ₀, fr ₀), (er ₁, fr ₁), (er ₂, fr ₂) ..., (er _NR-1, fr _NR-1), (er _NR, fr _NR), wherein, er _iAnd fr ₁Satisfy following relation:

fr _i＝k _r(i)er _i+b _r(i)，i＝1，2，....，N _R(9)

fr ₀＝b _r(1)

Wherein: fr _iBe the cell element CE in the RGB color space _RGB(i, j, k) boundary coordinate er _iThrough cell element CE _RGB(i, j, k) numerical value after the last linear gamma characteristic function distortion.

In like manner, for g, the b component also can approach the Gamma characterisitic function of g component, b component with broken line, that is:

For the g component, be set with N _GThe bar broken line just has N between these broken lines _G+ 1 end points, its coordinate is respectively (eg ₀, fg ₀), (eg ₁, fg ₁), (eg ₂, fg ₂) ..., (eg _NG-1, fg _NG-1), (eg _NG, fg _NG), wherein, eg _jAnd fg _jSatisfy following relation:

fg _j＝k _g(j)eg _j+b _g(j)，j＝1，2，....，N _G(10)

fg ₀＝b _g(1)

Wherein: fg _jBe the cell element CE in the RGB color space _RGB(k) boundary coordinate egj is through cell element CE for i, j _RGB(i, j, k) numerical value after the last linear gamma characteristic function distortion.

For the b component, be set with N _BThe bar broken line just has N between these broken lines _B+ 1 end points, its coordinate is respectively (eb ₀, fb ₀), (eb ₁, fb ₁), (eb ₂, fb ₂) ..., (eb _NB-1, fb _NB-1), (eb _NB, fb _NB), wherein, eb _kAnd fb _kSatisfy following relation:

fb _k＝k _b(k)eb _k+b _b(k)，k＝1，2，....，N _B(11)

fb ₀＝b _b(1)

Wherein: fb _kBe the cell element CE in the RGB color space _RGB(i, j, k) boundary coordinate eb _kPass through cell element CER respectively _GB(i, j, k) numerical value after the last linear gamma characteristic function distortion.

The present invention carries out relating to finding the solution of mathematical optimization problem in the linearizing processing procedure at the Gamma characterisitic function, and this solution procedure can realize by adopting DSP (Digital Signal Processor, i.e. digital signal processor) chip.When the mathematical optimization problem adopts so-called neural net such as Hopfield network to find the solution, can utilize the commercial neural net digital integrated circuit chip that has existed at present.

Step 2, the YUV color space is carried out cell element divide, adopt the method for subregion linearity to represent, promptly determine the linear Gamma correction function of YUV color space the Gamma correction function of YUV color space.

Owing in step 2, need to use the conversion between YUV color space and the RGB color space, for technical scheme of the present invention is more readily understood, so, before step 2 is elaborated, introduce the conversion between YUV color space and the RGB color space earlier.

Because in fact color space be exactly a kind of coordinate system, therefore, the principle about coordinate system transformation is followed in the mathematics in the conversion between YUV color space and the RGB color space.This conversion can be linear, also can be non-linear.

The RGB color space is linear to the conversion of YUV color space, is contrary if set from the YUV color space to the direction of RGB color notation conversion space, and the direction from the RGB color space to the YUV color notation conversion space is for just, and then the mathematical relationship of direct transform is:

$\left[\begin{array}{c}y\\ u\\ v\end{array}\right]=\left[\begin{array}{ccc}0.299& 0.587& 0.114\\ -0.147& -0.289& 0.436\\ 0.615& -0.515& -0.100\end{array}\right]\left[\begin{array}{c}r\\ g\\ b\end{array}\right]$

The mathematical relationship of inverse transformation is:

$\left[\begin{array}{c}r\\ g\\ b\end{array}\right]=\left[\begin{array}{ccc}1.0000& 0& 1.1398\\ 1.0000& -0.3946& -0.5805\\ 0.615& 2.0320& -0.0005\end{array}\right]\left[\begin{array}{c}y\\ u\\ v\end{array}\right]$

If set from the RGB color space to the direction of YUV color notation conversion space to just, the direction from the YUV color space to the RGB color notation conversion space is contrary, and then the mathematical relationship of inverse transformation is formula (13), and the mathematical relationship of direct transform is formula (12).

The present invention is primarily aimed at the situation that color notation conversion space is linear transformation.For linear transformation, can use transformation matrix to represent, and represent direct transform with T (Transform), R (Reverse Transform) represents inverse transformation.

The conversion of color space can not change the dimension of color space, and promptly color space all is 3 dimension spaces, and therefore, transformation matrix is the matrix of 3 * 3 sizes, that is:

$T=\left[\begin{array}{ccc}{t}_{11}& t_{12}& t_{13}\\ t_{21}& t_{22}& t_{23}\\ t_{31}& t_{32}& t_{33}\end{array}\right]$

After the conversion of having introduced between YUV color space and the RGB color space, divide below in conjunction with the cell element of 10 pairs of YUV color spaces of accompanying drawing and to describe.

As can be seen from Figure 10, the coordinate system of YUV color space is not a quadrature, the coordinate system that is the YUV color space is not a rectangular coordinate system, that is to say, be not mutually orthogonal in twos (orthogonal) or vertical (perpendicular) between three reference axis of YUV color space coordinate system.This coordinate system generally is called affine (affine) coordinate system, and general euclidean coordinates system is that rectangular coordinate system is a kind of special case of affine coordinate system.Therefore, the YUV color space that is generated by the RGB color notation conversion space is a parallelepiped in fact.Though each limit of this parallelepiped is not vertical mutually, all is parallel to reference axis separately.

In affine coordinate system, can represent the YUV color space of above-mentioned parallelepiped with following set, that is:

CSYUV＝{(y，u，V)|Y _min≤y≤Y _max，U _min≤u≤U _max，V _min≤V≤V _max}。Wherein, CSYUV is a name set, and its English is Color Space YUV.This parallelepiped CSYUV is along Y, U, and the axial length of side of V is respectively:

L _Y＝Y _max-Y _min

L _U＝U _max-U _min (15)

L _v＝V _max-V _min

Can calculate according to formula (12): Y _Max=1, Y _Min=0, U _Max=0.436, U _Min=-0.436, V _Max=0.615, V _Min=-0.615.

For each the cell element CE in the RGB color space _RGB(i, j, k) (i=1,2 ..., N _R, j=1,2 ..., N _G, k=1,2 ..., N _B), through of the conversion of RGB color space, being mapped to a little parallelepiped in the YUV color space to the YUV color space, this little parallelepiped promptly is a cell element in the YUV color space.The present invention can use CE _YUV(i, j, k) (i=1,2 ..., N _R, j=1,2 ..., N _G, k=1,2 ..., N _B) represent a cell element in the YUV color space.

Because cell element in the RGB color space and the cell element in the YUV color space are corresponding one by one, so the YUV color space is by cell elementization, and the number of cell element is identical with the number of cell element in the YUV color space in the RGB color space.For cell element CE _YUV(k), its coordinate range is respectively for i, j:

0.299fr _i-1+0.587fg _j-1+0.114fb _k-1＝y _min(i，j，k)≤y≤y _max(i，j，k)＝0.299fr _i+0.587fg _j+0.114fb _k；(16)

-0.147fr _i-0.289fg _j+0.436fb _k-1＝u _min(i，j，k)≤u≤u _max(i，j，k)＝-0.147fr _i-1-0.289fg _j-1+0.436fb _k；(17)

0.615fr _i-1-0.515fg _j-0.1fb _k＝v _min(i，j，k)≤v≤v _max(i，j，k)＝0.615fr _i-0.515fg _j-1-0.1fb _k-1；(18)

Wherein: fr _i, fg _j, fb _kBe the cell element CE in the RGB color space _RGB(i, j, k) boundary coordinate er _i, eg _j, eb _kPass through cell element CE respectively _RGB(i, j, k) numerical value after the last linear gamma characteristic function distortion, i.e. fr _i, fg _j, fb _kBe the point on the gamma characteristic function curve; Certainly, fr _i, fg _j, fb _kIt also can not be the point on the gamma characteristic function curve.

The present invention represents Gamma correction function in this cell element with the segmented line shape function in each cell element of YUV color space.The present invention represents Gamma correction function in the cell element with gc (.), and then the Gamma correction function is at each cell element CE _YUV(i, j, k) (i=1,2 ..., N _R, j=1,2 ..., N _G, k=1,2 ..., N _B) in the concrete manifestation form as follows:

$c_{\mathrm{cYUV}}=\mathrm{gc}\left(c_{\mathrm{uYUV}}\right)=\begin{array}{c}\left[\begin{array}{c}g{c}_y\left(c_{\mathrm{uYUV}}\right)\\ {\mathrm{gc}}_u\left(c_{\mathrm{uYUV}}\right)\\ {\mathrm{gc}}_v\left(c_{\mathrm{uYUV}}\right)\end{array}\right]\end{array}=$

$\left[\begin{array}{c}{p}_y(i,j,k)c_{\mathrm{uYUV}}+q_y(i,j,k)=p_y^1(i,j,k)y_u+p_y^2(i,j,k)u_u+p_y^3(i,j,k)v_u+q_y(i,j,k)\\ \begin{array}{}\end{array}{p}_u(i,j,k)c_{\mathrm{uYUV}}+q_u(i,j,k)=p_u^1(i,j,k)y_u+p_u^2(i,j,k)u_u+p_u^3(i,j,k)v_u+q_u(i,j,k)\\ \begin{array}{}\end{array}{p}_v(i,j,k)c_{\mathrm{uYUV}}+q_v(i,j,k)=p_v^1(i,j,k)y_u+p_v^2(i,j,k)u_u+p_v^3(i,j,k)v_u+q_v(i,j,k)\end{array}\right]---\left(19\right)$

If c _UYUV∈ CE _YUV(i, j, k), and i=1,2 ...., N _R, j=1,2 ...., N _G, k=1,2 ...., N _B

Wherein: 3 * 3 matrixes $P(i,j,k)=\left[\begin{array}{c}{p}_y(i,j,k)\\ p_u(i,j,k)\\ p_v(i,j,k)\end{array}\right]=\left[\begin{array}{c}{p}_y^1(i,j,k)p_y^2(i,j,k)p_y^3(i,j,k)\\ p_u^1(i,j,k)p_u^2(i,j,k)p_u^3(i,j,k)\\ p_v^1(i,j,k)p_v^2(i,j,k)p_v^3(i,j,k)\end{array}\right];$

3 * 1 column vectors $Q(i,j,k)=\left[\begin{array}{c}{q}_y(i,j,k)\\ q_u(i,j,k)\\ q_v(i,j,k)\end{array}\right].$

Step 3, proofread and correct the association process of the vector of YUV color space, the vector after the association is positioned at same signal space with distortion RGB color vector not.

For video input apparatus, because the flow direction of its vision signal is: the YUV color space of the RGB color space of undistorted RGB color space-distortion-do not carry out Gamma correction-the carried out YUV color space of Gamma correction.Therefore, can carry out association to the vector of proofreading and correct the YUV color space by following process, make itself and distortion RGB color vector not be positioned at same signal space, concrete association process is:

For each the cell element CE in the YUV color space _YUV(i, j, k), and at first, the signal vector c in undistorted RGB color space _RRGBThrough after the Gamma link of video input apparatus, produce gradation distortion, obtain the signal vector c in the RGB of distortion color space _DRGBC wherein _RRGBBelong to not distortion RGB color space, c _DRGBThe RGB color space that belongs to distortion, and c _DRGB=K (i, j, k) c _RRGB+ B (i, j, k), K (i, j, k) and B (i, j k) are the parameter of the linear Gamma characterisitic function on each cell element in the RGB color space.

Then, the signal vector c in the RGB of distortion color space _DRGBThrough of the association of RGB color space, obtain the signal vector c in the YUV of the distortion of not carrying out Gamma correction color space to the YUV color space _UYUV, and c _UYUV=T (K (i, j, k) c _RRGB+ B (i, j, k)), wherein: c _UYUVBelong to uncorrected YUV color space vector, T is the direct transform from the RGB color space to the YUV color space.

Signal vector c at the YUV color space that does not carry out Gamma correction _UYUVAfter proofreading and correct through Gamma again, obtain the signal vector c in the YUV color space of having proofreaied and correct _CYUV, and c _CYUV=P (i, j, k) T (K (i, j, k) c _RRGB+ B (i, j, k))+Q (i, j, k), wherein: c _CYUVBelong to the YUV color space of having proofreaied and correct, P (i, j, k) and Q (i, j k) are the parameter of the linear Gamma correction function of each cell element in the YUV color space.

By above-mentioned derivation as can be known, the vector of having proofreaied and correct the YUV color space is associated as:

c _CYUV=P (i, j, k) T (K (i, j, k) c _CRGB+ B (i, j, k)+(i, if j is k) c for Q _CRGB∈ CE _RGB(i, j, k), and i=1,2 ...., N _R, j=1,2 ...., N _G, k=1,2 ...., N _B(20)

The foundation of step 4, mean-square error criteria.

For video input apparatus, the most direct method of metric G amma calibration result is: weigh the vectorial C in the YUV color space after proofreading and correct _CYEVVectorial C in related back and the RGB color space that does not have under the Gamma distortion situation _RRGBThe size of average generalized distance.Average generalized distance such as mean-square error criteria, weighted mean square error criterion etc.

It is Tc that setting does not have the vector of the YUV color space under the Gamma distortion situation _RRGB, then, concrete errors of form can be defined as according to statistical signal process field mean-square error criteria commonly used:

(c _cYUV-Tc _rRGB)T(c _cYUV-Tc _rRGB)。

Like this, on whole YUV color space, the accumulation of this error can be with the formal definition as lower integral:

$\mathrm{MSE}={\&Integral;}_{c_{\mathrm{rRGB}\&Element;{\left[\mathrm{0,1}\right]}^3}}{(c_{\mathrm{cYUV}}-T{c}_{\mathrm{rRGB}})}^T(c_{\mathrm{cYUV}}-T{c}_{\mathrm{rRGB}})d{c}_{\mathrm{rRGB}}$

Wherein: MSE represents mean square error, and English is Mean Square Error.

From formula (21) as can be seen, mean square error has reflected that Gamma proofreaies and correct a kind of cumulative metric of deviation of the vector of the RGB color space that the result departs from the imaginary Gamma of process distortion.In like manner, weighted mean square error has reflected the cumulative metric of deviation too.This cumulative metric result's numerical value is more little, and it is more little to illustrate that then Gamma proofreaies and correct the deviation of vector of the RGB color space that the result departs from the imaginary Gamma of process distortion, and the fidelity that Gamma proofreaies and correct is high more.

Step 5, determine the linear Gamma correction function of each cell element in the YUV color space based on mean-square error criteria.

Because each cell element CE in the RGB color space _RGB(i, j, k) and the YUV color space in each cell element CE _YUV(i, j, k) corresponding one by one, so, be defined in CE _RGB(k) the Gamma characterisitic function on is defined within CE for i, j _YUV(Gamma correction function k) is proofreaied and correct for i, j.The Gamma characterisitic function and the Gamma correction function of other cell element do not influence CE _RGB(i, j, k) and CE _YUV(i, j, k).Therefore, the present invention draws a kind of important conclusions and is: each cell element CE _YUV(Gamma correction function k) can independently be determined, only depends on cell element CE for i, j _RGB(i, j k) go up the Gamma characterisitic function, only depend on cell element CE in other words _RGB(i, j, k) the linear Gamma characterisitic function on.

According to above-mentioned independent principle, the present invention can determine the Gamma correction function by cell element ground at the YUV color space.Because all cell elements have constituted whole YUV color space, so, as long as determined each cell element CE in the YUV color space _YUV(Gamma correction function k) just can be determined the Gamma correction function on whole YUV color spaces for i, j.

For video input apparatus, the present invention determines that the criterion of the Gamma correction function institute foundation on each cell element in the YUV color space is: the minimum mean square error criterion of formula (21) definition, i.e. the present invention need determine 3 * 3 matrixes

$P(i,j,k)=\left[\begin{array}{c}{p}_y(i,j,k)\\ p_u(i,j,k)\\ p_v(i,j,k)\end{array}\right]=\left[\begin{array}{c}{p}_y^1(i,j,k)p_y^2(i,j,k)p_y^3(i,j,k)\\ p_u^1(i,j,k)p_u^2(i,j,k)p_u^3(i,j,k)\\ p_v^1(i,j,k)p_v^2(i,j,k)p_v^3(i,j,k)\end{array}\right]$ With 3 * 1 column vectors $Q(i,j,k)=\left[\begin{array}{c}{q}_y(i,j,k)\\ q_u(i,j,k)\\ q_v(i,j,k)\end{array}\right],$

Make $\mathrm{MSE}={\&Integral;}_{c_{\mathrm{rRGB}\&Element;{\left[\mathrm{0,1}\right]}^3}}{(c_{\mathrm{cYUV}}-T{c}_{\mathrm{rRGB}})}^T(c_{\mathrm{cYUV}}-T{c}_{\mathrm{rRGB}})d{c}_{\mathrm{rRGB}}$ Minimum.

In conjunction with formula (20) and (21), can obtain:

$\mathrm{MSE}=$

${\&Integral;}_{c_{\mathrm{cRGB}\&Element;{\&lsqb;0.1\&rsqb;}^3}}{(P(i,j.k)T(K(i,j,k)c_{\mathrm{rRGB}}+B(i,j,k))+Q(i,j,k)-T{c}_{\mathrm{rRGB}})}^T$

$(P(i,j,k)T(K(i,j,k)c_{\mathrm{rRGB}}+B(i,j,k))Q(i,j,k)-T{c}_{\mathrm{rRGB}})d{c}_{\mathrm{rRGB}}$

(22)

Formula (22) can be converted into the mathematical optimization problem of classics, that is:

$\{P_{\mathrm{Opt}}(i,j,k),Q_{\mathrm{Opt}}(i,j,k)\}=$

$\underset{P(i,j,k),Q(i,j,k)}{\arg \min }{\&Integral;}_{c_{\mathrm{cRGB}\&Element;{\&lsqb;0.1\&rsqb;}^3}}{(P(i,j.k)T(K(i,j,k)c_{\mathrm{rRGB}}+B(i,j,k))+Q(i,j,k)-T{c}_{\mathrm{rRGB}})}^T---\left(23\right)$

$(P(i,j,k)T(K(i,j,k)c_{\mathrm{rRGB}}+B(i,j,k))Q(i,j,k)-T{c}_{\mathrm{rRGB}})d{c}_{\mathrm{rRGB}}$

That is to say that the present invention obtains the globally optimal solution P of formula (23) by the method that adopts mathematical optimization _Opt(i, j, k), Q _Opt(k), this globally optimal solution is exactly the Gamma correction function parameters of each cell element of YUV color space for i, j.Wherein Opt is the abbreviation of English Optimal (optimum).

Method for solving to formula (23) belongs to prior art, is not described in detail in the present embodiment.

It is the design process that Gamma proofreaies and correct that above-mentioned steps 1 to step 5 can be called, and promptly determines to carry out the process of the parameter of required use in the Gamma trimming process.In general, Gamma link for a given Gamma link or cascade, carrying out the design process that Gamma proofreaies and correct can determine according to actual needs, when the gamma link that comprises at video input apparatus changes, carries out the design process of a Gamma correction.That is to say that the variation of carrying out the design process of Gamma correction and the vision signal that video input apparatus is exported is irrelevant.

The Gamma correction that step 6, the video image of directly video input apparatus being exported in the YUV color space according to the Gamma correction function of above-mentioned acquisition pursue cell element.

For CE _YUV(i, j, k), since its coordinate range and Y, U, and the length of side on the V direction is respectively:

0.299fr _i-1+0.587fg _j-1+0.114fb _k-1＝y _min(i，j，k)≤y≤y _max(i，j，k)＝0.299fr _i+0.587fg _j+0.114fb _k；-0.147fr _i-0.289fg _j+0.436fb _k-1＝u _min(i，j，k)≤u≤u _max(i，j，k)＝-0.147fr _i-1-0.289fg _j-1+0.436fb _k；0.615fr _i-1-0.515fg _j-0.1fb _k＝v _min(i，j，k)≤v≤v _max(i，j，k)＝0.615fr _i-0.515fg _j-1-0.1fb _k-1；

Therefore, for the vision signal (y in one group of given uncorrected YUV color space _u, u _u, v _u), carry out following steps and carry out Gamma correction:

A, for each two field picture in the video sequence of video input apparatus output, judge each the vision signal (y in the image _u, u _u, v _u) belonging to which cell element in the YUV color space, the present invention adopts the method for coordinate range comparison to determine cell element under the pixel in the image, if promptly there is certain group (i ₀, j ₀, k ₀), and (i ₀, j ₀, k ₀) formula (27) is satisfied:

y _min(i ₀，j ₀，k ₀)≤y _u≤y _max(i ₀，j ₀，k ₀)

u _min(i ₀，j ₀，k ₀)≤u _u≤u _max(i ₀，j ₀，k ₀)(27)

v _mm(i ₀，j ₀，k ₀)≤v _u≤v _max(i ₀，j ₀，k ₀)

Then can determine vision signal (y _u, u _u, v _u) belong to cell element CE _YUV(i ₀, j ₀, k ₀).

The present invention can be to vision signal (y _u, u _u, v _u) adopt by cell element mode relatively, contrast the coordinate range of vision signal and cell element, when definite vision signal belongs to the coordinate range of certain cell element, stop follow-up comparison procedure.

B, the Gamma correction function that utilizes corresponding cell element carry out Gamma to this vision signal to be proofreaied and correct, and output.Continuous precedent is set vision signal (y _u, u _u, v _u) belong to cell element CE _YUV(i ₀, j ₀, k ₀), and cell element CE _YUV(i ₀, j ₀, k ₀) corresponding Gamma correction function parameters is P (i ₀, j ₀, k ₀) and Q (i ₀, j ₀, k ₀), then to vision signal (y _u, u _u, v _u) result that carries out Gamma correction is:

$c_{\mathrm{cYUV}}=\left[\begin{array}{c}{y}_c\\ u_c\\ v_c\end{array}\right]=P({\mathrm{<figure-callout\; id="0"\; label="i"\; filenames="A20061011157900321.png,A20061011157900331.png"\; state="\{\{state\}\}">i</figure-callout>}}_0,{\mathrm{<figure-callout\; id="0"\; label="j"\; filenames="A20061011157900321.png,A20061011157900331.png"\; state="\{\{state\}\}">j</figure-callout>}}_0,k_0)c_{\mathrm{cYUV}}+Q({\mathrm{<figure-callout\; id="0"\; label="i"\; filenames="A20061011157900321.png,A20061011157900331.png"\; state="\{\{state\}\}">i</figure-callout>}}_0,{\mathrm{<figure-callout\; id="0"\; label="j"\; filenames="A20061011157900321.png,A20061011157900331.png"\; state="\{\{state\}\}">j</figure-callout>}}_0,k_0)$

The Gamma correction process need of step 6 all repeats each sampled value of the vision signal of video input apparatus output in the video communication process.

In the implementation procedure of step 6, it is the Gamma correction that Gamma correcting integrated circuit chip is realized the YUV color space that the present invention can utilize hardware circuit.Gamma correcting integrated circuit chip realizes that the schematic diagram of YUV color space Gamma correction is shown in accompanying drawing 11, accompanying drawing 12.

Among Figure 11, Gamma correcting integrated circuit chip of the present invention is the correction that can support multi-channel video signal, and allow the user by its required parameter that uses in the Gamma correction process of instruction configuration, as the positional information of each cell element of YUV color space, the linear Gamma correction function parameters representation that each cell element in the YUV color space is corresponding respectively, linear Gamma correction function parameters or the like.

Among Figure 12, comprise 3 electronic circuits at Gamma correcting integrated circuit chip internal and realize Y respectively, U, the correction of V component.If with linear dimensions P, the linear Gamma correction function that Q represents converts the N on the curved surface to ³Individual point, each point be 4 coordinate components (Y, U, V, F), F representative function value wherein, like this, Gamma correcting integrated circuit chip needs 4N ³Individual parameter.If Gamma correcting integrated circuit chip is proofreaied and correct M road signal simultaneously, then the configuration parameter number of Gamma correcting integrated circuit chip is 4MN ³

The present invention also provides a kind of means for correcting of video communication gamma characteristic, comprising: memory module, linearization block, cell element module, mathematical optimization module, find the solution module and Gamma correction module.

Memory module is mainly used in to the Gamma correction module the required parameter of using in the Gamma correction process is provided, positional information as each cell element of YUV color space, linear Gamma correction function parameters representation that each cell element in the YUV color space is corresponding respectively and linear Gamma correction function parameters etc., memory module also is a linearization block, the cell element module, the mathematical optimization module provides necessary parameter, as the positional information of each cell element of RGB color space, the gamma characteristic function of the gamma link that video input apparatus comprised in the RGB color space, and the linear Gamma correction function parameters representation of each cell element difference correspondence in the YUV color space etc.Canned data comprises the information that the user imports in the memory module, gamma link gamma characteristic function in the RGB color space that is comprised as the positional information of each cell element of RGB color space, linear Gamma correction function parameters representation, video input apparatus etc.Canned data also comprises other modules input in the means for correcting of the present invention in the memory module, as the positional information of each cell element of YUV color space, linear Gamma correction function parameters etc.

Linearization block is mainly used in the parameter that the linearisation of gamma characteristic function on each cell element of RGB color space of the gamma link that the calculating video input apparatus such as positional information according to gamma characteristic function in the RGB color space of canned data such as video input apparatus comprised in the memory module gamma link, each cell element of RGB color space are comprised represented, promptly obtains the linear gamma characteristic function of each cell element.Linearization block can adopt the method for setting up the mathematical optimization problem and finding the solution to obtain the linear gamma characteristic function of each cell element.

When linearization block adopts the mathematical optimization problem and the method found the solution when obtaining the linear gamma characteristic function of each cell element, linearization block can be set up the mathematical optimization problem of describing as formula (6), (7), (8) among the method embodiment, and, linearization block can be utilized and find the solution module and come the mathematical optimization problem of its foundation is found the solution, to obtain the linear gamma characteristic function of each cell element of RGB color space.The linear gamma characteristic function of each cell element of RGB color space that linearization block obtains transfers to memory module, is stored by memory module.

The positional information that the cell element module is mainly used in according to each cell element of RGB color space of storing in the memory module is mapped to the YUV color space one to one with each cell element in the RGB color space, obtain the positional information of each cell element of YUV color space, and transfer to memory module and store; The RGB color space here can be the RGB color space of normalization, and like this, the YUV color space is a normalization RGB color space by the YUV color space that color notation conversion space generated.The positional information of each cell element of YUV color space of cell element module output is shown in formula (16), (17), (18) among the above-mentioned method embodiment.

The mathematical optimization module be mainly used in utilize the linear gamma characteristic function stored in the memory module and linear Gamma correction function to the vector sum of proofreading and correct the YUV color space not distortion RGB color vector carry out related, make itself and distortion RGB color vector not be positioned at same signal space, as the result after the association shown in formula (20) among the above-mentioned method embodiment.The mathematical optimization module is utilized formula (20) and based on average generalized distance minimum criteria between vector the mathematical optimization problem is not set between the distortion RGB color vector, as the mathematical optimization problem that is provided with shown in formula (22).

Finding the solution module can be a public general module, as commercial neural net digital integrated circuit chip etc., finds the solution module and can find the solution the mathematical optimization problem that mathematics optimal module and cell element module are set up.With the linear Gamma correction function parameters on each cell element in the linear gamma characteristic function of determining each cell element of RGB color space and the YUV color space, and transfer to memory module, store by memory module.

The Gamma correction module is mainly used in the Gamma correction that the yuv video signal of video input apparatus output is pursued cell element.Each two field picture in each video sequence of exporting is established in input for video, and the Gamma correction module need be according to each the vision signal (y in the positional information judgement image of each cell element of YUV color space of storing in the memory module _u, u _u, v _u) belong to which cell element in the YUV color space, then, utilize the Gamma correction function of corresponding cell element that this vision signal is carried out the Gamma correction, and the vision signal behind the output calibration.The Gamma correction module can be the Gamma correcting integrated circuit chip of describing among the said method embodiment, and Gamma correction module parameters needed is provided by memory module.Concrete as the description in the above-mentioned method.

Though described the present invention by embodiment, those of ordinary skills know, the present invention has many distortion and variation and do not break away from spirit of the present invention, and the claim of application documents of the present invention comprises these distortion and variation.

Claims (10)

1. video input apparatus gamma characteristic correcting method in the video communication is characterized in that, comprising:

A, the RGB color space is divided into a plurality of cell elements;

B, according to the gamma characteristic function of gamma link in the RGB color space that video input apparatus comprised, obtain the linearisation of described gamma characteristic function on each cell element of RGB color space and represent, promptly obtain the linear gamma characteristic function of each cell element;

C, each cell element of RGB color space is mapped to the YUV color space one to one, makes the YUV color space be divided into a plurality of cell elements;

D, the linear Gamma correction function parameters representation of each cell element in the YUV color space is set;

E, utilize described linear gamma characteristic function and linear Gamma correction function to the vector sum of proofreading and correct the YUV color space not distortion RGB color vector carry out related, make the vector after the association be positioned at same signal space with distortion RGB color vector not, utilize the vector sum after the described association based on average generalized distance minimum criteria between vector the mathematical optimization problem not to be set between the distortion RGB color vector, find the solution this mathematical optimization problem, to determine the linear Gamma correction function parameters on each cell element of YUV color space;

F, according to the linear Gamma correction function on each cell element of YUV color space the yuv video signal of video input apparatus output is pursued the Gamma correction of cell element.

2. the method for claim 1 is characterized in that, the RGB color space that described RGB color space is normalization, described YUV color space are normalization RGB color space by the YUV color space that described color notation conversion space generated.

3. method as claimed in claim 1 or 2 is characterized in that, described step a comprises:

With R axle five equilibrium with obtain each cell element the length of side on the R axle or with the R axle not five equilibrium to obtain the length of side of each cell element on the R axle;

With G axle five equilibrium with obtain each cell element the length of side on the G axle or with the G axle not five equilibrium to obtain the length of side of each cell element on the G axle;

With B axle five equilibrium with obtain each cell element the length of side on the B axle or with the B axle not five equilibrium to obtain the length of side of each cell element on the B axle.

4. method as claimed in claim 1 or 2 is characterized in that, described step b comprises:

The gamma characteristic function of gamma link in the RGB color space of supposing video input apparatus is known, and each division points based on cell element on the R axle is er _i, the R axle is divided into N _RIndividual interval, i=1,2,3 ...., N _R, each division points based on cell element on the G axle is eg _j, the G axle is divided into N _GIndividual interval, j=1,2,3 ...., N _G, each division points based on cell element on the B axle is eb _k, the B axle is divided into N _BIndividual interval, k=1,2,3 ...., N _B, then for the (i, j, k) individual cell element, the mathematical optimization problem of setting up according to the least mean-square error optiaml ciriterion is:

$\{k_r\left(i\right),b_r\left(i\right),i=\mathrm{1,2,3}....,N_R\}=\arg \underset{\underset{i=\mathrm{1,2,3},....,N_R}{k_r\left(i\right),b_r\left(i\right)\&Element;R,}}{\min }\underset{i=1}{\overset{N_R}{\mathrm{\&Sigma;}}}{{\&Integral;}_{{\mathrm{er}}_{i-1}}^{{\mathrm{er}}_i}|{\mathrm{gm}}_r\left(r_r\right)-k_r\left(i\right)r_r-b_r\left(i\right)|}^2{\mathrm{dr}}_r$

$s.t.\left\{\begin{array}{c}{k}_r\left(1\right){\mathrm{er}}_1+b_r\left(1\right)=k_r\left(2\right){\mathrm{er}}_1+b_r\left(2\right)\\ k_r\left(2\right){\mathrm{er}}_2+b_r\left(2\right)=k_r\left(3\right){\mathrm{er}}_2+b_r\left(3\right)\\ \&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\\ k_r(N_R-1){\mathrm{er}}_{N_R-1}+b_r(N_R-1)=k_r\left(N_R\right){\mathrm{er}}_{N_R-1}+b_r\left(N_R\right)\\ k_r\left(N_R\right)+b_r\left(N_R\right)=1\end{array}\right.;$

$\{k_g\left(j\right),b_g\left(j\right),j=\mathrm{1,2,3},....,N_G\}=\arg \underset{\underset{j=\mathrm{1,2,3},....,N_G}{k_g\left(j\right),b_g\left(j\right)\&Element;R,}}{\min }\underset{j=1}{\overset{N_G}{\mathrm{\&Sigma;}}}{\&Integral;}_{{\mathrm{eg}}_{j-1}}^{{\mathrm{eg}}_j}{|{\mathrm{gm}}_g\left(g_r\right)-k_g\left(j\right)g_r-b_g\left(j\right)|}^2{\mathrm{dg}}_r$

$s.t.\left\{\begin{array}{c}{k}_g\left(1\right){\mathrm{eg}}_1+b_g\left(1\right)=k_g\left(2\right){\mathrm{eg}}_1+b_g\left(2\right)\\ k_g\left(2\right){\mathrm{eg}}_2+b_g\left(2\right)=k_g\left(3\right){\mathrm{eg}}_2+b_g\left(3\right)\\ \&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\\ k_g(N_G-1){\mathrm{eg}}_{N_G-1}+b_g(N_G-1)=k_g\left(N_G\right){\mathrm{eg}}_{N_G-1}+b_G\left(N_G\right)\\ k_g\left(N_G\right)+b_g\left(N_G\right)=1\end{array}\right.;$

$\{k_b\left(k\right),b_b\left(k\right),k=\mathrm{1,2,3},....,N_B\}=\arg \underset{\underset{k=\mathrm{1,2,3},....,N_B}{k_b\left(k\right),b_b\left(k\right)\&Element;R,}}{\min }\underset{k=1}{\overset{N_B}{\mathrm{\&Sigma;}}}{\&Integral;}_{{\mathrm{eb}}_{k-1}}^{{\mathrm{eb}}_k}{|{\mathrm{gm}}_b\left(b_r\right)-k_b\left(k\right)b_r-b_b\left(k\right)|}^2{\mathrm{db}}_r$

$s.t.\left\{\begin{array}{c}{k}_b\left(1\right){\mathrm{eb}}_1+b_b\left(1\right)=k_b\left(2\right){\mathrm{eb}}_1+b_b\left(2\right)\\ k_b\left(2\right){\mathrm{eb}}_2+b_b\left(2\right)=k_b\left(3\right){\mathrm{eb}}_2+b_b\left(3\right)\\ \&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;\\ k_b(N_B-1){\mathrm{eb}}_{N_B-1}+b_b(N_B-1)=k_b\left(N_B\right){\mathrm{eb}}_{N_B-1}+b_b\left(N_B\right)\\ k_b\left(N_B\right)+b_b\left(N_B\right)=1\end{array}\right.;$

Wherein: s.t. represents constraints, k _r(i) and b _r(i) be (i, j, k) the R component gamma parameter of the linear gamma characteristic function of individual cell element, the k of RGB color space _g(j) and b _g(j) be (i, j, k) the G component gamma parameter of the linear gamma characteristic function of individual cell element of RGB color space; k _h(k) and b _b(k) be (i, j, k) the B component gamma parameter of the linear gamma characteristic function of individual cell element of RGB color space;

Above-mentioned mathematical optimization problem is found the solution, obtain parameter matrix K (i, j, k) and parameter vector B (i, j, k);

Repeat the aforementioned calculation process for each cell element of RGB color space, to determine the linear gamma characteristic function parameters on each cell element, the linear gamma characteristic function of each cell element of RGB color space is:

$c_{\mathrm{dRGB}}=K(i,j,k)c_{\mathrm{rRGB}}+B(i,j,k)=\left[\begin{array}{ccc}{k}_r\left(i\right)& 0& 0\\ 0& k_g\left(j\right)& 0\\ 0& 0& k_b\left(k\right)\end{array}\right]c_{\mathrm{rRGB}}+\left[\begin{array}{c}{b}_r\left(i\right)\\ b_g\left(j\right)\\ b_b\left(k\right)\end{array}\right];$

Wherein: i, j, k are cell element CE _RGBCall number, c _DRGBBe the RGB color vector through the distortion of gamma link, c _RRGBFor not saving the RGB color vector of distortion through gamma loop.

5. method as claimed in claim 1 or 2 is characterized in that, among the described step c, and each cell element CE in the YUV color space _YUV(the limit k) coordinate range on Y, U, V direction respectively is for i, j:

0.299fr _j-1+0.587fg _j-1+0.114fb _k-1＝y _min(i，j，k)≤y≤y _max(i，j，k)＝0.299fr _i+0.587fg _j+0.114fb _k；-0.147fr _i-0.289fg _j+0.436fb _k-1＝u _min(i，j，k)≤u≤u _max(i，j，k)＝-0.147fr _i-1-0.289fg _j-1+0.436fb _k；0.615fr _i-1-0.515fg _j-0.1fb _k＝v _min(i，j，k)≤v≤v _max(i，j，k)＝0.615fr _i-0.515fg _j-1-0.1fb _k-1；

Wherein: fr _i, fg _j, fb _kBe the cell element CE in the RGB color space _RGB(i, j, k) boundary coordinate er _i, eg _j, eb _kPass through cell element CE respectively _RGB(i, j, k) numerical value after the last linear gamma characteristic function distortion, i.e. fr _i, fg _j, fb _kBe the point on the gamma characteristic function curve; Or described fr _i, fg _j, fb _kIt is not the point on the gamma characteristic function curve.

6. method as claimed in claim 2 is characterized in that, step related among the described step e comprises:

Linear Gamma correction function in linear gamma characteristic function in the setting RGB color space on each cell element, the YUV color space on each cell element is respectively:

c _dRGB＝K(i，j，k)c _rRGB+B(i，j，k)，c _cYUV＝P(i，j，k)c _uYUv+Q(i，j，k)；

Wherein: c _DRGBBe the RGB color vector through the distortion of gamma link, c _RRGBFor not through the RGB color vector of gamma loop joint distortion, K (i, j, k) and B (i, j k) are each cell element linearity gamma characteristic function parameters in the RGB color space, c _CYUVBe the YUV color vector of proofreading and correct, c through the gamma link _UYUVBe the YUV color vector of proofreading and correct through the gamma loop joint, P (i, j, k) and Q (i, j k) are each cell element linearity Gamma correction function parameters of YUV color space;

Utilize linear gamma characteristic function with c _RRGBWith the YUV color vector c that does not carry out Gamma correction _URGBCarry out association:

c _uYUV＝T(K(i，j，k)c _rRGB+B(i，j，k))；

Wherein: T represents the direct transform from the RGB color space to the YUV color space;

Utilize linear Gamma correction function will proofread and correct YUV color vector and c _RRGBCarry out association:

c _cYUV＝P(i，j，k)T(K(i，j，k)c _rRGB+B(i，j，k))+Q(i，j，k)；

Wherein: P (i, j, k) and Q (i, j k) are linear Gamma correction function parameters on each cell element of YUV color space, i, j, k are cell element CE _YUVCall number, i=1,2,3 ...., N _R, j=1,2,3 ...., N _G, k=1,2,3 ...., N _B, N _RBe the interval number that forms owing to the cell element division on the R axle, N _GBe the interval number that forms owing to the cell element division on the G axle, N _BBe the interval number that forms owing to the cell element division on the B axle.

7. method as claimed in claim 6 is characterized in that, the mathematical optimization problem is set among the described step e comprises with the step of finding the solution:

In the accumulation of two vector errors described in the distortion RGB color space not be:

$\mathrm{MSE}={\&Integral;}_{c_{\mathrm{rRGB}}\&Element;{\left[\mathrm{0,1}\right]}^3}{(c_{\mathrm{cYUV}}-{\mathrm{Tc}}_{\mathrm{rRGB}})}^T(c_{\mathrm{cYUV}}-{\mathrm{Tc}}_{\mathrm{rRGB}}){\mathrm{dc}}_{\mathrm{rRGB}};$

Wherein: MSE represents mean square error;

In conjunction with c _CYUV=P (i, j, k) T (K (i, j, k) c _RRGB+ B (i, j, k))+Q (i, j k), draw:

$\mathrm{MSE}=$

${\&Integral;}_{c_{\mathrm{cRGB}}\&Element;{\left[\mathrm{0,1}\right]}^3}{(P(i,j,k)T(K(i,j,k)c_{\mathrm{rRGB}}+B(i,j,k))+Q(i,j,k)-{\mathrm{Tc}}_{\mathrm{rRGB}})}^T$

$(P(i,j,k)T(K(i,j,k)c_{\mathrm{rRGB}}+B(i,j,k))+Q(i,j,k)-{\mathrm{Tc}}_{\mathrm{rRGB}}){\mathrm{dc}}_{\mathrm{rRGB}}$

Thereby for cell element CE _YUV(i, j, k) i=1,2,3 ...., N _R, j=1,2,3 ...., N _G, k=1,2,3 ...., N _B, obtain following mathematical optimization problem:

$\{P_{\mathrm{Opt}}(i,j,k),Q_{\mathrm{Opt}}(i,j,k)\}=$

$\underset{P(i,j,k),Q(i,j,k)}{\arg \min }{\&Integral;}_{c_{\mathrm{cRGB}}\&Element;{\left[\mathrm{0,1}\right]}^3}{(P(i,j,k)T(K(i,j,k)c_{\mathrm{rRGB}}+B(i,j,k))+Q(i,j,k)-{\mathrm{Tc}}_{\mathrm{rRGB}})}^T$

$(P(i,j,k)T(K(i,j,k)c_{\mathrm{rRGB}}+B(i,j,k))+Q(i,j,k)-{\mathrm{Tc}}_{\mathrm{rRGB}}){\mathrm{dc}}_{\mathrm{rRGB}}$

Wherein: P _Opt(i, j, k), Q _Opt(i, j k) are the globally optimal solution of described mathematical optimization problem;

The described globally optimal solution P of described mathematical optimization problem _Opt(i, j, k), Q _Opt(i, j k) are exactly cell element CE in the YUV color space _YUV(i, j, linear Gamma correction function parameters k).

8. method as claimed in claim 1 or 2 is characterized in that, described step f comprises:

The cell element of each sampled value of yuv video signal of determining video input apparatus output under in the YUV color space;

Each sampled value affiliated cell element in the YUV color space according to described vision signal is determined the Gamma correction function that each sampled value is corresponding respectively;

According to determined Gamma correction function each sampled value of described vision signal is carried out Gamma correction.

9. video input apparatus gamma characteristic means for correcting in the video communication is characterized in that described device comprises: memory module, linearization block, cell element module, mathematical optimization module, find the solution module and Gamma correction module;

Memory module: be used for receiving and storing the positional information of positional information, each cell element of YUV color space of each cell element of RGB color space, gamma link that video input apparatus comprised in the gamma characteristic function of RGB color space and corresponding respectively linear Gamma correction function parameters representation and the linear Gamma correction function parameters of each cell element in the YUV color space;

Linearization block: be used for calculating the parameter that the linearisation of gamma characteristic function on each cell element of RGB color space of the gamma link that video input apparatus comprised represented according to the memory module canned data, promptly obtain the linear gamma characteristic function of each cell element, and output result of calculation is stored to memory module;

The cell element module: the positional information according to each cell element of RGB color space of storing in the memory module is mapped to the YUV color space one to one with each cell element in the RGB color space, obtain the positional information of each cell element of YUV color space, and transfer to memory module and store;

The mathematical optimization module: utilize the linear gamma characteristic function stored in the memory module and linear Gamma correction function to the vector sum of proofreading and correct the YUV color space not distortion RGB color vector carry out related, make itself and distortion RGB color vector not be positioned at same signal space, utilize the vector sum after the described association based on average generalized distance minimum criteria between vector the mathematical optimization problem not to be set between the distortion RGB color vector;

Find the solution module: find the solution described mathematical optimization problem, determining the linear Gamma correction function parameters on each cell element in the YUV color space, and transfer to memory module and store;

The Gamma correction module: the positional information of each cell element in linear Gamma correction function parameters representation, linear Gamma correction function parameters and the YUV color space that is used for storing according to memory module, the yuv video signal of video input apparatus output is pursued the Gamma correction of cell element.

10. means for correcting as claimed in claim 9 is characterized in that, the RGB color space that described RGB color space is normalization; Described YUV color space is a normalization RGB color space by the YUV color space that described color notation conversion space generated.

Images