CN117058118B - A method for evaluating the image degradation effect caused by flat shielding glass screens - Google Patents

Abstract

The present invention discloses a method for evaluating the image deterioration effect caused by a planar shielding glass screen, comprising the following steps: S1. Given the shielding effectiveness and the test frequency or frequency band requirements, the light transmittance is calculated and the line width and line period of the planar shielding glass screen are determined; S2. Given the relevant planar shielding glass screen parameters and camera parameters, the screen point spread function is obtained; S3. The point spread function is multiplied by the calculated light transmittance, the light intensity distribution is attenuated, and the diffraction degradation model is obtained; S4. The energy threshold is set to simplify the point spread function; S5. A fuzzy image of diffraction degradation is obtained; S6. The difference between the original clear color plate image and the fuzzy image is calculated according to the PSNR and SSIM functions, so as to realize the image degradation rapid calculation and evaluation of the image deterioration effect caused by the planar shielding glass screen. The present invention can quickly evaluate the image deterioration degree caused by the screen in advance according to the designed screen parameters and camera parameters during the design process.

Description

A method for evaluating the image degradation effect caused by flat shielding glass screens

Technical Field

The invention relates to electromagnetic shielding and optical design, and in particular to a method for evaluating the image deterioration effect caused by a planar shielding glass mesh.

Background technique

For the detection of sensitive phenomena of test objects such as cars and airplanes in darkroom test rooms, it is usually necessary to use cameras to observe some visible sensitive phenomena, such as the rotation of the pointer on the dashboard, the flashing of lights, etc. For test rooms where the camera is located in a strong electromagnetic environment, it is usually necessary to install a shielding net or coating on the camera to achieve electromagnetic shielding and thus protect the camera. Due to its grid shape, the planar shielding net usually causes grid diffraction problems in the image, which leads to blurred images. Therefore, it is necessary to provide an evaluation method for evaluating the image deterioration caused by screen diffraction according to the design parameters during the design stage of the shielding net camera.

In optical diffraction, the point spread function is an important physical quantity for evaluating the distribution and intensity of optical diffraction, but it is also the source of image blur. However, there is currently a lack of research on the diffraction blur of plane-shielded glass screen cameras based on optical diffraction theory, making it difficult to provide an evaluation method for image blur caused by screen diffraction.

Summary of the invention

The purpose of the present invention is to overcome the shortcomings of the prior art and provide a method for evaluating the image degradation effect caused by a flat shielding glass screen, which can quickly evaluate the degree of image degradation caused by the screen in advance according to the designed screen parameters and camera parameters during the design process.

The object of the present invention is achieved through the following technical solution: A method for evaluating the image degradation effect caused by a flat shielding glass screen, comprising the following steps:

S1. Given the shielding effectiveness and test frequency or frequency band requirements, calculate the light transmittance and determine the line width and line period of the planar shielding glass mesh;

S2. Given the relevant plane shielding glass screen parameters and camera parameters, obtain the screen point spread function;

S3. multiplying the point spread function by the calculated light transmittance, performing attenuation of the light intensity distribution, and obtaining a diffraction degradation model;

S4. setting an energy threshold, simplifying the point spread function, and obtaining a blur kernel after simplification;

S5. performing blur kernel calculation on different channels of the color plate image according to different spectral response frequencies according to the blur kernel representation and convolving the image with the channel image to obtain a diffraction-degraded blur map;

S6. The difference value between the original clear color plate image and the blurred image is calculated according to the PSNR and SSIM functions, so as to realize the rapid calculation and evaluation of the image degradation effect caused by the planar shielding glass mesh.

Furthermore, in step S1, through the Ulrich’s model, the transmittance T of the plane shielding glass mesh at vertical incidence is approximately:

Where T(0,0) is the transmittance of the screen under zero frequency and zero wave vector conditions, that is, the transmittance ratio of the incident electromagnetic wave, which is between 0 and 1, where g is the line period of the screen; 2a is the line width of the screen; the thickness of the screen is t;

The electromagnetic shielding effectiveness is:

SE = -10log(T)

Ulrich's model is applicable when Where λ is the wavelength of the incident wave, _ng , _n0 are the refractive index of the substrate and the refractive index of the medium on both sides of the screen respectively;

Given the shielding effectiveness SE and the test frequency f or the test wavelength λ, the line period g and line width a are estimated to determine whether the planar shielding glass mesh involved will undergo optical diffraction;

Furthermore, the screen parameters and camera parameters described in step S2 include:

The length and width of the light-transmitting part of the shielding screen are b _x and b _y ;

Shielding wire mesh grating constants d _x , _dy ;

The mesh number of the shielding screen in the x-direction and the y-direction is N ₁ , N ₂ ;

Camera focal length f;

Camera target size s;

Camera resolution r = (m, n);

The spectral center wavelengths of the responses of the three channels of BRG are λ _B , λ _G , and λ _R .

Furthermore, the step S2 comprises:

Generally, the camera target surface size is measured in inches, which indicates the diagonal length of the rectangular target surface. In the industry, 1 inch is 16 mm, and the aspect ratio of a general sensor is 4:3.

When the target surface size s ₁ (in) is given, the diagonal length of the sensor is calculated based on the target surface size:

_s2 (mm) = 16*s1 (in)

Then, the actual physical length of the target surface is calculated based on the aspect ratio of 4:3 and the Pythagorean theorem:

s _w (mm) = 12.8 * s ₁ (in), s _h (mm) = 9.6 * s ₁ (in)

In the process of blur kernel calculation, the coordinates are generally symmetric about 0, so the actual physical length is taken as half, which represents the semi-axis length of the target surface in length and width:

According to the paraxial approximation principle, the _diffraction angles _of the light in two directions are estimated as follows:

Estimate the point spread function of the planar shielded glass mesh:

in

Where _I0 represents the light intensity generated by a single slit at the center of the pattern; x, y represent the physical coordinates of the pixel point on the target surface [ _-xl , _xl ] and [ _-yl , _yl ] axes respectively; _δx , _δy represent the optical path difference of light in the x and y directions, and _α1 , _α2 , _β1 , and _β2 are intermediate variables.

Furthermore, in step S3, due to the problem of the shielding mesh, the transmittance of the light decreases, that is, the image becomes dim. The transmittance of the center level is derived by combining the plane shielding glass mesh point spread function with Ulrich's transmittance model:

Combined with the point spread function, the diffraction degradation model of the planar shielding glass mesh can be derived as follows:

Where _fc (x,y) represents the degraded image on image channel c, _hc (x,y;λ) represents the point spread function on image channel c, _gc (x,y;λ) is the potential undegraded spectral image on image channel c, and (x,y) is the physical coordinate of the CCD or CMOS target surface of the image sensor.

Furthermore, in the simplified process described in step S4, assuming that the light intensity I ₀ generated by the center of the pattern is 1, then for any channel c in RGB, the point spread function on the x-axis and y-axis is expressed as:

For target coordinates whose x and y intervals are [-x _l ,x _l ] and [-y _l ,y _l ] respectively, x and y in the above formula must satisfy:

-x _l ≤x ≤x _l ,-y _l ≤y ≤y _l

Set thresh as the simplified blur kernel threshold, and its value is between 0 and 1. Then there is a truncated valid area [-x _s ,x _s ], [-y _s ,y _s ] that satisfies:

_{0≤xs≤xl , 0≤ys≤yl ​}

Make the ratio of the energy sum of the elements in the effective area to the total energy greater than the given threshold, that is:

This is an inequality related to the wavelength λ. When the spectral response wavelength λ of a given channel c is given, the rectangle represented by [ _-xs , _xs ] × [ _-ys , _ys ] is the blur kernel of the channel, denoted as _kc1 (x, y; λ);

When c=1, 2, or 3, blur kernels of the three RGB channels are obtained, where c=1 represents the R channel, c=2 represents the G channel, and c=3 represents the B channel.

Furthermore, the step S6 comprises:

Let the point spread function h _c (x, y; λ) in the diffraction degradation model be equal to the simplified blur kernel k _c1 (x, y; λ), and the generative model of the blurred image is obtained:

Apply the above model to the three channel images g _c (x, y; λ) of the clear image g (x, y) in turn to obtain the degraded image f _c (x, y) of each channel and concatenate them to obtain the final blurred degraded image f (x, y);

For the image evaluation of the blurred image degradation effect, the PSNR and SSIM functions are used to evaluate the degree of degradation between the original undegraded clear color plate image g(x, y) and the blurred image f(x, y) degraded by the blur kernel:

Where m, n are the width and height of the image, bitdepth is the image depth, μ _x , μ _y , σ _x , σ _y , σ _xy are the mean, variance, and covariance of image x, y; c ₁ , c ₂ are constants.

The beneficial effect of the present invention is that the present invention can quickly evaluate the degree of image deterioration caused by the screen in advance according to the designed screen parameters and camera parameters during the design process, thereby providing reference and guidance for the design of related parameters, reducing the uncertainty and time and resource costs from design to processing.

BRIEF DESCRIPTION OF THE DRAWINGS

Fig. 1 is a flow chart of the present invention;

FIG2 is an optical system diagram of the present invention corresponding to imaging of a plane-shielded glass screen camera;

FIG3 is a schematic diagram of image degradation;

FIG. 4 is a schematic diagram of a color palette image.

Detailed ways

The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings, but the protection scope of the present invention is not limited to the following.

Starting from the perspective of optical diffraction and camera imaging principles, the present invention realizes the estimation of light transmittance brought by plane shielding glass mesh, derivation of point spread function of plane shielding glass mesh, establishment of diffraction degradation model, simplification of blur kernel and evaluation of degraded image through relevant theoretical deduction and modeling.

As shown in Figure 2, the screen of a plane-shielded glass screen camera usually takes the form of a light window. Generally, the screen is placed in front of the lens, and the distance between the two is basically close during actual installation; behind the lens is a photosensitive sensor. For any object outside the camera, the reflected light propagates as a spherical wave. Since the relationship between the distance of the object and the focal length usually satisfies Fraunhofer diffraction, when the light reaches the camera, its wave surface can be equivalent to a plane wave. Therefore, diffraction and interference will occur when the light passes through the shielding net, resulting in blurring during imaging.

For the above optical imaging system, further, as shown in FIG1 , firstly, given the shielding effectiveness SE and the test frequency f or wavelength λ requirements, assuming that the shielding effectiveness we need is not less than 60 dB, the test frequency is 2 GHz, according to the Ulrich model:

It can be calculated that the light transmittance is less than or equal to 1e-6, and its wavelength is 0.15m, which meets the applicable conditions of the Ulrich model:

Further substituting the parameters into Ulrich’s model, we can obtain:

The ratio of a and g can be further limited, and the line period and line width can be estimated to be on the order of micrometers, for example At this time, the line period g is approximately 64μm and the line width a is 6.4μm.

Then we can measure the corresponding screen parameters and camera parameters, mainly including:

(1) Length and width of the light-transmitting part of the shielding screen: b _x , _by ;

(2) Shielding wire mesh grating constants d _x , _dy ;

(3) The mesh count of the shielding screen in the x-direction and the y-direction N ₁ , N ₂ ;

(4) Camera focal length f;

(5) Camera target surface size s;

(6) Camera resolution r = (m, n);

(7) The spectral center wavelengths of the responses of the three channels of BRG are λ _B , λ _G , and λ _R .

Based on the above measurement parameters, first calculate the actual physical length (s _w , s _h ) according to the camera target surface size. Assume that the imaging target surface size of a 1080P camera is 1/3” inch, that is The length of its diagonal is:

s(mm)＝16/3≈6mm

Then the actual physical length of the corresponding target surface is:

s _w (mm) = 4.8 _{mm. sh} (mm) = 3.6 (mm)

That is, its corresponding physical size is 4.8mm×3.6mm, and its corresponding semi-axis lengths in length and width are:

x _l =2.4 mm, y _l =1.8 mm

For a 1080P camera, further calculations show that its pixel size is 2.5μm×3.3μm. Based on the semi-axis length, the image sensor coordinate axes with x-axis and y-axis ranges of (-2.4mm.2,4mm) and (-1.8mm.1.8mm) can be established, and the two axes are divided into 1920 points and 1080 points respectively.

According to the paraxial approximation principle, the diffraction angles of light in two directions, θ _x and θ _y , can be estimated as follows:

Then, the point spread function of the planar shielding glass mesh is estimated based on the known measurement parameters:

in

The corresponding blur kernel size is 1920 × 1080. Combined with the previously calculated transmittance, according to the linearity and spatial invariance of the imaging system, the current diffraction degradation model, that is, the original degradation method, can be obtained:

Where _fc (x,y) represents the degraded image on channel c, _hc (x,y;λ) represents the point spread function on channel c, _gc (x,y;λ) is the potential undegraded color plate image on channel c, as shown in Figure 4, (x,y) is the physical coordinate of the image sensor CCD or CMOS target surface calculated previously.

In order to reduce the amount of calculation, we can set the energy threshold and simplify the point spread function. For the BGR three-channel image, the three channels need to be convoluted and blurred separately, as shown in Figure 3. Since the plane shielding glass mesh has little effect on the spectral response function, the measured spectral response center wavelengths of the three channels λ _B , λ _G , λ _R are assumed to be 450nm, 540nm, and 660nm. Taking the blue channel as an example, the energy threshold thresh is set to 0.9, which needs to meet the following conditions:

At this time, the simplified blur kernel size is only 14×14. At this time, the simplified blur kernel replaces the h _c (x, y; λ) function in the original degradation formula and performs convolution degradation with the blue channel image to obtain the blurred image of this channel. The same operation is performed on the G channel and the R channel, and finally the three channels are merged to obtain the final blurred image.

Finally, the difference between the original clear color plate image and the blurred image is calculated based on the PSNR and SSIM functions:

Where m and n are the image width and height measured previously, bitdepth is the image depth, μ _x , μ _y , σ _x , σ _y , σ _xy are the mean, variance and covariance of the clear image and the blurred image; c ₁ and c ₂ are constants.

PSNR describes the distortion between the image to be evaluated and the reference image. The larger the value, the smaller the distortion. SSIM describes the structural similarity of the two images. The larger the value, the more similar the structures are. Through the values of the two evaluation indicators, the image degradation caused by the flat shielding glass mesh can be quickly calculated and evaluated.

The methods disclosed and proposed by the present invention can be implemented by those skilled in the art by appropriately changing the conditions, requirements and parameters by referring to the contents of this article. Although the algorithm of the present invention has been described by preferred embodiments, relevant technical personnel can obviously modify or re-combine the methods and technical routes described herein without departing from the content, spirit and scope of the present invention to achieve the final preparation technology. It should be particularly noted that all similar substitutions and modifications are obvious to those skilled in the art, and they are all considered to be included in the spirit, scope and content of the present invention.

Claims (2)

1. A method for evaluating an image deterioration effect caused by a planar screen glass wire, characterized by: the method comprises the following steps:

s1, given shielding effectiveness and test frequency or frequency band requirements, calculating light transmittance and determining line width and line period of a planar shielding glass screen;

s2, giving related plane shielding glass silk screen parameters and camera parameters, and obtaining a silk screen point diffusion function;

the screen parameters and the camera parameters described in step S2 include:

length and width b of light-transmitting part of shielding silk screen _x ,b _y ；

Grating constant d of shielding silk screen _x ,d _y ；

Mesh number N of shielding silk screen in x direction and y direction ₁ ,N ₂ ；

A camera focal length f;

camera target surface size s;

camera resolution r= (m, n);

spectral center wavelength lambda of response of three channels of BRG _B ,λ _G ,λ _R ；

The step S2 includes:

the size of the target surface of the camera is expressed as the length of the corner line of the rectangular target surface in inches, the length of 1 inch is 16mm, and the aspect ratio of the sensor is set to be 4:3;

when the target surface size s is given ₁ Then, the diagonal line length s of the sensor is calculated according to the target surface size ₂ ：

s ₂ ＝16*s ₁

Then according to the aspect ratio s _w ：s _h =4:3, setCalculating the actual physical length of the target surface:

s _w ＝12.8*s ₁ ,s _h ＝9.6*s ₁

in the fuzzy core representation, the coordinates are symmetrical about 0, so the actual physical length is halved, i.e. the half-axis length in length and width of the target surface is represented:

estimating diffraction angle theta of light in two directions according to paraxial approximation principle _x ,θ _y The method comprises the following steps of:

estimating the diffusion function of the plane shielding glass fiber mesh points:

wherein the method comprises the steps of

Wherein I is ₀ Representing the intensity of light produced by a single slit at the center of the pattern; x and y respectively represent the pixel point at the target surface [ -x ] _l ,x _l ]，[-y _l ,y _l ]Physical coordinates on the axis; delta _x ,δ _y Representing the optical path difference of light in the x, y directions, alpha ₁ 、α ₂ 、β ₁ 、β ₂ Is an intermediate variable;

s3, multiplying the point spread function by the calculated light transmittance, and carrying out attenuation on light intensity distribution to obtain a diffraction degradation model;

in the step S3, due to the problem of the shielding net, the transmittance of the light is reduced, that is, the image becomes dim, and the transmittance of the central stage is derived by combining the planar shielding glass fiber lattice point diffusion function with the transmittance model of Ulrich:

the diffraction degradation model of the planar shielding glass mesh can be derived by combining the point spread functions as follows:

wherein f _c (x, y) represents a degraded image on the image channel c, h _c (x,y; λ) represents the point spread function, g, on the image channel c _c (x, y; λ) the physical coordinates of the image sensor CCD or CMOS target surface;

s4, setting an energy threshold, simplifying a point spread function, and obtaining a fuzzy core after simplification;

in the simplification process described in step S4, the intensity I of the light generated in the center of the pattern is set ₀ In units 1, then for any channel c in RGB, the point spread function is represented on the x-axis and y-axis as:

for x, y intervals are [ -x respectively _l ,x _l ]，[-y _l ,y _l ]X, y in the above formula are required to satisfy:

-x _l ≤x≤x _l ,-y _l ≤y≤y _l

setting thresh as a simplified fuzzy core threshold, wherein the value is between 0 and 1, and truncated effective area [ -x ] exists _s ,x _s ]，[-y _s ,y _s ]The method meets the following conditions:

0≤x _s ≤x _l ,0≤y _s ≤y _l

so that the ratio of the sum of the energies of the elements within the active area to the total energy is greater than a given threshold, namely:

this is an inequality with respect to the wavelength λ, when the spectral response wavelength λ, [ -x ] of a given channel c _s ,x _s ]×[-y _s ,y _s ]The rectangle represented is then the fuzzy core of the channel, noted ask _c (x,y；λ)；

When c=1, 2,3, a fuzzy core of three channels of RGB is obtained, wherein c=1 represents an R channel, c=2 represents a G channel, and c=3 represents a B channel;

s5, performing fuzzy kernel calculation on different channels of the color plate graph according to different spectral response frequencies according to fuzzy kernel representation, and convolving the fuzzy kernel calculation with the channel image to obtain a diffraction degradation fuzzy graph;

s6, calculating a difference value between an original clear color plate image and a blurred image according to PSNR and SSIM functions, and realizing rapid calculation and evaluation of image degradation of an image degradation effect caused by a plane shielding glass silk screen;

the step S6 includes:

let the point spread function h in the diffraction degradation model _c (x, y; lambda) is equal to the reduced fuzzy kernel k _c (x, y; λ) to obtain a model of the generation of the blurred image:

r, G, B three-channel image g for clear image g (x, y) _c (x, y; lambda) acquiring a degraded image f of each channel by sequentially applying the above model _c (x, y) and stitching to obtain a final blurred degraded image f (x, y);

image evaluation for blur map deterioration effect the degree of deterioration between the original undegraded clear color plate image g (x, y) and the blurred image f (x, y) degraded with the blur kernel was evaluated using two functions of PSNR and SSIM:

where m, n are the width and height of the image, bitdepth is the depth of the image, μ _x ,μ _y ,σ _x ,σ _y ,σ _xy Is the mean, variance and covariance of the images x and y; c ₁ ,c ₂ Is constant.

2. A method for evaluating an image deterioration effect caused by a planar screen glass wire according to claim 1, wherein: the step S1 includes:

s101, giving shielding effectiveness SE, and calculating light transmittance T according to the shielding effectiveness, wherein the calculation mode is as follows:

SE＝-10log(T)

s102, through a Ulrich's model, the transmittance T of a vertical incidence plane shielding glass screen is approximately as follows:

wherein T (0, 0) is the transmittance of the silk screen under the conditions of zero frequency and zero wave vector, namely the transmittance ratio of the incident electromagnetic wave is between 0 and 1, and g is the line period of the silk screen; 2a is the line width of the silk screen; the thickness of the silk screen is t;

the applicable conditions of the Ulrich's model are as followsWhere λ is the wavelength of the incident wave, n _g ,n ₀ Respectively the refractive index of the substrate and the refractive index of the medium at two sides of the silk screen;

giving wavelength lambda, and simultaneously giving the ratio of a to g, and calculating a and g through a Ulrich' S model based on the light transmittance calculated in the step S101;

where g is also called grating constant, and its line period size in x, y direction is expressed as grating constant d _x ,d _y The method comprises the steps of carrying out a first treatment on the surface of the a represents the radius of the wire, which satisfies the following relation with the grating constant:

2a＝d _x -b _x ＝d _y -b _y

wherein b _x ,b _y The length and width of the light-transmitting part in the x and y directions of the grating are respectively shown.