CN116862904A - Minimum perceived difference-based display panel Mura defect global evaluation method - Google Patents

Abstract

The invention discloses a display panel Mura defect global evaluation method based on minimum perceived difference, which overcomes the defect of manually detecting the panel Mura defect, so that the final Mura defect quantitative evaluation is not influenced by artificial subjective identification factors, and the efficiency and the accuracy of the quality detection of a display panel are obviously improved.

Description

Minimum perceived difference-based display panel Mura defect global evaluation method

Technical Field

The invention belongs to the technical field of image processing, and particularly relates to a minimum perceived difference-based display panel Mura defect global evaluation method.

Background

Currently, the market of electronic devices such as mobile phones and computers has been in explosive growth, wherein electronic displays such as Liquid Crystal Displays (LCDs), plasma Displays (PDPs), organic Light Emitting Displays (OLEDs), etc. are increasingly popular with consumers as high-quality displays, which also puts forward higher and higher requirements on the productivity and quality of display panels, and most manufacturers currently evaluate the quality detection of display panels mainly by visual inspection personnel, and the quality detection results are greatly influenced by artificial subjective factors, and the evaluation lacks of fixed standards, and the error and working time are positively correlated, so that the complete display panel quality detection system needs to be developed urgently, and the accuracy and efficiency of detection are improved.

In general, one major factor in reducing the quality of a display image of a panel is non-uniformity, so-called "Mura", which is a common visual defect in displays such as LCDs, OLEDs, and is generally represented by low contrast, non-uniform brightness areas, blurred edges, and the like, which may cause visual discomfort to users. Meanwhile, the Mura defects are the most complex and difficult to detect and evaluate in visual defects, and the Mura defects comprise multiple types of punctiform Mura, linear Mura, massive Mura and the like.

Most electronic display manufacturers currently use limited samples for visual inspection, however, such detection methods inevitably introduce factors such as subjective identification, and the accuracy of the obtained detection results is low and the detection efficiency is low. With the development of technology, researchers began using machine vision instead of human eyes, and a high-precision camera was used in combination with a high-efficiency detection algorithm to evaluate a panel. And shooting an image of the display screen under a certain condition by using a high-precision camera, and then detecting and evaluating the Mura defect in the image by using a corresponding algorithm, so as to further give a quantization index of the Mura defect of the display panel.

However, how to accurately quantify the assessment of Mura defects during inspection has been an industry recognized challenge. Therefore, a method that can evaluate Mura defects with high accuracy is urgently needed. While various measurement methods for luminance Mura or chrominance Mura have been reported in various academic conferences, various methods are still in the laboratory development and testing stage, and no report has clearly provided a complete Mura automated quantitative assessment scheme useful for industrial production.

1. The existing display panel Mura defect evaluation method generally only evaluates the brightness uniformity and cannot consider the chromaticity Mura defect;

2. the existing evaluation methods generally need to detect and identify the shape and the existing position of the Mura through a complex image algorithm, then evaluate the Mura according to different Mura types, and mainly ignore relatively weak defects according to the local defect characteristics of a screen, so that global uniformity evaluation of the display panel on the whole is often impossible;

3. the existing method only evaluates the image shot by aiming at the angle of the camera facing the screen, but cannot consider the Mura defect of the screen from other angles, and the defect front view angle of some display panels is not easy to observe as that of the side view direction;

4. the existing method generally does not carry out special treatment on the edge area of the screen, and the problem of edge Mura defect missing detection is difficult to avoid;

5. the existing method is often only verified on simulated defect images or actual data with fewer Mura types, and the method is difficult to ensure to be applicable to various Mura defects.

Disclosure of Invention

The invention aims to overcome the existing defects and provide a display panel Mura defect global evaluation method based on minimum perceived difference so as to solve the problems in the background art.

In order to achieve the above purpose, the present invention provides the following technical solutions: a display panel Mura defect global evaluation method based on minimum perceived difference comprises the following steps:

firstly, the system acquires complete images of a display panel from multiple angles through a high-definition camera, positions a screen area for cutting, scaling and denoising and filtering, converts a preprocessed RGB color gamut image into a CIE-XYZ color gamut image, and further converts the CIE-XYZ color gamut image into a color matching space (wk, rg, by);

step two, calculating a test contrast diagram and a reference contrast diagram, generating a differential contrast diagram by the test contrast diagram and the reference contrast diagram, and introducing an edge mask to the Mura defect existing in the image part to process the image;

step three, performing Fourier transform on three components of a color space (wk, rg and by) respectively to obtain frequency components, calculating contrast sensitivity CSF functions describing a human visual system, filtering the three frequency components of wk/rg/by respectively by applying single-channel Contrast Sensitivity Functions (CSF) of the components, and performing inverse Fourier transform to the corresponding space domain;

step four, converting the space domain (wk, rg, by) image of the inverse Fourier transform to CIE-XYZ space, further converting to CIE-Lab space, and calculating a luminance component L and a chrominance component C under the CIE-Lab color gamut;

step five, respectively calculating a minimum perceived difference graph (JND) of a brightness component L and a chrominance component C, and calculating a global maximum JND value of the brightness component and the chrominance component and a ratio of a Mura defect area to a total screen area according to a formula;

and step six, calculating evaluation indexes LEV and CEV of the brightness component and the chrominance component according to the global maximum JND value and the Mura defect area occupation ratio, summarizing the evaluation results of the multi-angle shot images, and finally comprehensively evaluating the brightness and chrominance evaluation indexes to obtain a final Mura defect quantitative evaluation value MEV of the display panel.

Preferably, in the first step:

the multi-angle shooting display panel of the camera can generally take 3 angles, the camera faces to the center point of the screen, and the axis of the lens coincides with the central axis of the screen at a viewing angle 1; the camera is opposite to the center point of the screen and is positioned at the left side of the screen and at a viewing angle 2 forming an angle theta 1 with the central axis of the screen; the camera is opposite to the center point of the screen and is positioned at the right side of the screen and at a viewing angle 3 forming an angle theta 2 with the central axis of the screen;

the specific conversion formula for converting the RGB color gamut image into the CIE-XYZ color gamut image is as follows:

the invention firstly normalizes the RGB value to transform the range to 0-1, denoising filtering can be completed by combining Gaussian filtering and median filtering of a spatial domain, three channels of median filtering respectively carry out [5,5] convolution kernel size, and the image is expanded by copying the value of an edge region so as to avoid the condition that the value of four corners of the filtered image is 0; then, the XYZ values are obtained by using the formula and multiplied by 100, and the range of the XYZ values is converted into 0-100 so as to be used for subsequent calculation;

the CIE-XYZ gamut image is transformed into a color pair space according to the following conversion matrix:

wk＝0.279X+0.72Y-0.107Z

rg＝-0.449X+0.29Y-0.077Z。

by＝0.086X-0.59Y+0.501Z

preferably, in the second step:

differential contrast map DC (x, y) is derived from test contrast map C _Test (x, y) and reference contrast map C _Refence (x, y) generation, in particular as follows

Wherein C is _Test (x, y) is the original image converted into the color space (wk, rg, by), C _Refence (x, y) filtering the background image of the original image after filtering all mura defects by Gaussian filtering with a frequency low enough as shown in the following formula

X, y in the above formula, the rows and columns of the image;

r—visual resolution, unit: px/deg;

s-Gaussian scale, units: the deg, generally 2deg;

if mura defects exist in the edge part of the image, the conventional method is very easy to detect the mura defects, so that the image needs to be processed by using an edge mask, and the specific algorithm is as follows

Wherein M, N-image height and width, unit: pix;

b _gain the boundary gain can be generally 1;

b _scale boundary ratio, unit: degs, generally 0.5 degs are available;

and (3) carrying out dot multiplication on each pixel of the differential contrast map DC (x, y) and the edge mask function B (x, y) one by one, so that the accuracy of an algorithm on edge mura defect detection can be improved.

Preferably, in the third step:

the formula for calculating the contrast sensitivity function CSF of the human visual system is as follows:

wherein the method comprises the steps of

ω＝(u ² +v ² ) ^1/2

(u, v) is the spatial frequency coordinates in cycles per degree, the angle being the angle between the human eye to a point on the screen and the vertical to the screen;

when processing the wk component, the values in the contrast sensitivity function CSF can be taken as

γ(ω)＝0.003100ω ² -0.10680ω+1.396

m _p ＝5.5 σ _p ＝1/40 ε＝1

m _c ＝5.5 σ _c ＝1/12 b _ias ＝55

When processing the rg component, the individual values in the contrast sensitivity function CSF can be taken as

γ(ω)＝0.001531ω ² -0.06149ω+1.140

m _p ＝1.5 σ _p ＝1/70 ε＝0

When processing by components, the respective values in the contrast sensitivity function CSF can be taken as

γ(ω)＝0.001919ω ² -0.06427ω+1.090

m _p ＝1.5 σ _p ＝1/45 ε＝1/4

m _c ＝7.5 σ _c ＝1/20 b _ias ＝0

In the aboveIs the average value of the three components wk, rg and by of the color space;

according to the formula, the transfer function of the human visual system CSF model of three components of the color space wk, rg and by can be calculated; it should be noted that the unit of the spatial frequency ω involved in the above formula is a period/degree, that is, a formula given from the viewpoint of an observer, and the unit of the spectrogram obtained by fourier transform is a period/pixel, so that the two units are not identical, and it is necessary to perform conversion of the unit when calculating the spatial frequency ω of CSF using the following formula:

wherein H is _R To measure the number of pixels of the screen level in the image, V _D For visual vertical distance, W _L Is the width (mm) of the display screen;

in addition, the calculation involves calculationWhen the three components are negative, the average value of the wk, rg and by components is required to be subjected to modulus taking, so that the average value of the three components is always positive, and the data obtained by calculating the CSF model can be consistent with the experimental result;

then, frequency domain filtering is carried out on the three components (WK, RG, BY) BY using a transfer function obtained BY calculating the CSF model, and matrix point multiplication is carried out on a spectrogram of the three components (WK, RG, BY) and a corresponding CSF function, so that the frequency domain filtering can be completed; it should be noted that the CSF function values corresponding to the three components (WK, RG, BY) are larger to prevent the subsequent computation from being abnormal, the invention firstly performs synchronous normalization on the CSF function values of the three components before the frequency domain filtering, that is, finds the maximum value in the CSF function values of the three components, divides the maximum value BY the maximum value respectively, and ensures that the value of the CSF function of the three components is always within [0,1] to facilitate the subsequent computation; finally, the three component frequency domain maps filtered BY the CSF function are inverse fourier transformed into spatial domain components (WK, RG, BY).

Preferably, in the fourth step:

the conversion matrix for transforming the color space (wk, rg, by) into CIE-XYZ space is as follows:

X＝0.626555wk-1.867178rg-0.153156by

Y＝1.369855wk+0.934756rg+0.436229by

Z＝1.505651wk+1.421324rg+2.536021by

then, the image is further transformed from CIE-XYZ space to CIE-Lab space to obtain three components L, a and b;

wherein f (x) is a piecewise function

X in the above formula _n ,Y _n ,Z _n The values with highest occurrence frequency of X, Y and Z components in CIE-XYZ space are obtained by counting the rounded values of each component;

from a, b, the chrominance component C can be calculated according to the formula, i.e

Preferably, in the fifth step:

after obtaining the luminance component L and the chrominance component C, the minimum perceived difference maps (JND) of the luminance and chrominance components are then calculated, respectively, according to the Weber theorem

Wherein delta represents the luminance component L and the chrominance component C,is a gaussian window function, then β is the minkofsky distance coefficient, typically β has a value of 2.4;

from the above local JND graph, the global JND values for the luminance and chrominance components can be further generalized using minkofsk basis functions, which can be called Pooling (Pooling), so its expression is:

when gamma-infinity, a global maximum JND value of both luminance and chrominance components is obtained, i.e

JND _{Δ_max} ＝max[JND(x,y)]Δ∈{L,C}；

Next, the ratio of the Mura defect area of the luminance and chrominance JND map to the total screen area is calculated as follows

Wherein p (x, y) and q (x, y) are each

Preferably, in the sixth step:

the luminance evaluation index LEV is lev=jnd _{L_area1} +0.28JND _{L_max} +6.47JND _{L_area2}

Wherein JND _{L_area1} JND is the ratio of Mura defect area with brightness JND graph value more than or equal to 1 to total screen area _{L_max} JND is the global maximum of the luminance JND graph _{L_area2} The ratio of the Mura defect area with the brightness JND diagram value more than or equal to 2 to the total screen area is represented;

the chromaticity evaluation index CEV is cev=jnd _{C_area1} +0.16JND _{C_max} +3.19JND _{C_area2}

Wherein JND _{C_area1} JND is the ratio of Mura defect area with chroma JND picture value more than or equal to 1 to total screen area _{C_max} JND is the global maximum of the chroma JND graph _{C_area2} Representing the ratio of the Mura defect area with the chromaticity JND diagram value more than or equal to 2 to the total screen area;

wherein N is the total number of display panel images shot at multiple angles, LEV _i I epsilon {1,2,3, …, N } is a brightness evaluation index under multiple angles, CEV _i I epsilon {1,2,3, …, N } is a chromaticity evaluation index under multiple angles, gamma is the ratio between the two terms, where 0.6 is generally taken, p is a Minkofski coefficient, and 2 is generally taken;

the luminance and chromaticity evaluation indexes are comprehensively calculated according to the following formula to obtain a final display panel Mura defect quantization evaluation value MEV which is mev=2.353 LEV _total +3.573CEV _total ；

The display panel quality can be quantitatively evaluated according to the calculated Mura defect quantitative evaluation value MNEV.

Compared with the prior art, the invention provides a display panel Mura defect global evaluation method based on minimum perceived difference, which has the following beneficial effects:

1. compared with the existing display panel Mura defect evaluation method, the system architecture provided by the invention not only can evaluate the brightness uniformity, but also can simultaneously consider the chroma Mura defect, and the algorithm expansibility is better;

2. compared with the existing evaluation method, the method provided by the invention does not need to detect and identify the shape and the existing position of Mura before quantitative evaluation, so that the defect evaluation can be carried out on the panel more comprehensively on the whole;

3. compared with the existing method, the method provided by the invention has the advantages that images of the screen are shot by the camera from multiple angles for evaluation, and the multi-angle defect evaluation result is fused, so that the robustness of the algorithm is better;

4. compared with the existing method, the method provided by the invention considers the problem that Mura defects in the edge area of the screen are easy to leak detection, and provides a method for processing images by using an edge mask, so that the algorithm has higher intelligence;

5. compared with the existing method, the system provided by the invention has the defect that verification is only performed on simulated defect images or actual data with fewer Mura types, and verification is performed through 20 types of simulated Mura and collected actual Mura data, so that the method can be applied to various types of Mura defects.

The invention overcomes the defect of the Mura defect of the manual detection panel, so that the evaluation of the Mura defect is not influenced by the artificial subjective identification factors; the method can be applied to quantitative evaluation of the Mura defects of the brightness and the chromaticity of the display panel, can also take account of the situation that the camera shoots at different angles of the screen to carry out overall comprehensive evaluation, and also considers the influence of the leakage detection problem of the Mura defects of the edge area of the screen on the evaluation, thereby remarkably improving the accuracy of acquiring the Mura defects of the display panel, enabling more panel manufacturers to benefit from the technology and promoting the establishment of relevant standards of the quantitative evaluation of the Mura defects of the display panel.

Drawings

The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate and together with the embodiments of the invention and do not constitute a limitation to the invention, and in which:

FIG. 1 is a schematic diagram of Mura defect of a display panel according to the present invention;

FIG. 2 is a schematic diagram illustrating image acquisition of a multi-angle display panel according to the present invention;

FIG. 3 is a flow chart of a method for global evaluation of Mura defects of a display panel according to the present invention;

FIG. 4 is a schematic diagram showing the comparison of the correlation coefficients between the subjective evaluation of the defects and the evaluation result of the algorithm in the actual test method of the present invention;

FIG. 5 is a schematic diagram of correlation coefficients between subjective evaluation and algorithm evaluation results of 20 main Mura defects according to the method of the present invention;

fig. 6 is a schematic structural diagram of a Mura defect global evaluation system for a display panel according to the present invention.

Detailed Description

The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

Referring to fig. 1-6, the present invention provides a technical solution: a display panel Mura defect global evaluation method based on minimum perceived difference comprises the following steps:

firstly, the system acquires complete images of a display panel from multiple angles through a high-definition camera, positions a screen area for cutting, scaling and denoising and filtering, converts a preprocessed RGB color gamut image into a CIE-XYZ color gamut image, and further converts the CIE-XYZ color gamut image into a color matching space (wk, rg, by);

step two, calculating a test contrast diagram and a reference contrast diagram, generating a differential contrast diagram by the test contrast diagram and the reference contrast diagram, and introducing an edge mask to the Mura defect existing in the image part to process the image;

step three, performing Fourier transform on three components of a color space (wk, rg and by) respectively to obtain frequency components, calculating contrast sensitivity CSF functions describing a human visual system, filtering the three frequency components of wk/rg/by respectively by applying single-channel Contrast Sensitivity Functions (CSF) of the components, and performing inverse Fourier transform to the corresponding space domain;

step four, converting the space domain (wk, rg, by) image of the inverse Fourier transform to CIE-XYZ space, further converting to CIE-Lab space, and calculating a luminance component L and a chrominance component C under the CIE-Lab color gamut;

step five, respectively calculating a minimum perceived difference graph (JND) of a brightness component L and a chrominance component C, and calculating a global maximum JND value of the brightness component and the chrominance component and a ratio of a Mura defect area to a total screen area according to a formula;

and step six, calculating evaluation indexes LEV and CEV of the brightness component and the chrominance component according to the global maximum JND value and the Mura defect area occupation ratio, summarizing the evaluation results of the multi-angle shot images, and finally comprehensively evaluating the brightness and chrominance evaluation indexes to obtain a final Mura defect quantitative evaluation value MEV of the display panel.

In the present invention, preferably, in the first step:

the multi-angle shooting display panel of the camera can generally take 3 angles, the camera faces to the center point of the screen, and the axis of the lens coincides with the central axis of the screen at a viewing angle 1; the camera is opposite to the center point of the screen and is positioned at the left side of the screen and at a viewing angle 2 forming an angle theta 1 with the central axis of the screen; the camera is opposite to the center point of the screen and is positioned at the right side of the screen and at a viewing angle 3 forming an angle theta 2 with the central axis of the screen;

the specific conversion formula for converting the RGB color gamut image into the CIE-XYZ color gamut image is as follows:

the invention firstly normalizes the RGB value to transform the range to 0-1, denoising filtering can be completed by combining Gaussian filtering and median filtering of a spatial domain, three channels of median filtering respectively carry out [5,5] convolution kernel size, and the image is expanded by copying the value of an edge region so as to avoid the condition that the value of four corners of the filtered image is 0; then, the XYZ values are obtained by using the formula and multiplied by 100, and the range of the XYZ values is converted into 0-100 so as to be used for subsequent calculation;

the CIE-XYZ gamut image is transformed into a color pair space according to the following conversion matrix:

wk＝0.279X+0.72Y-0.107Z

rg＝-0.449X+0.29Y-0.077Z。

by＝0.086X-0.59Y+0.501Z

in the present invention, preferably, in the second step:

differential contrast map DC (x, y) is derived from test contrast map C _Test (x, y) and reference contrast map C _Refence (x, y) generation, in particular as follows

Wherein C is _Test (x, y) is the original image converted into the color space (wk, rg, by), C _Refence (x, y) filtering the background image of the original image after filtering all mura defects by Gaussian filtering with a frequency low enough as shown in the following formula

X, y in the above formula, the rows and columns of the image;

r—visual resolution, unit: px/deg;

s-Gaussian scale, units: the deg, generally 2deg;

if mura defects exist in the edge part of the image, the conventional method is very easy to detect the mura defects, so that the image needs to be processed by using an edge mask, and the specific algorithm is as follows

Wherein M, N-image height and width, unit: pix;

b _gain the boundary gain can be generally 1;

b _scale boundary ratio, unit: degs, generally 0.5 degs are available;

and (3) carrying out dot multiplication on each pixel of the differential contrast map DC (x, y) and the edge mask function B (x, y) one by one, so that the accuracy of an algorithm on edge mura defect detection can be improved.

In the present invention, preferably, in step three:

the formula for calculating the contrast sensitivity function CSF of the human visual system is as follows:

wherein the method comprises the steps of

ω＝(u ² +v ² ) ^1/2

(u, v) is the spatial frequency coordinates in cycles per degree, the angle being the angle between the human eye to a point on the screen and the vertical to the screen;

when processing the wk component, the values in the contrast sensitivity function CSF can be taken as

γ(ω)＝0.003100ω ² -0.10680ω+1.396

m _p ＝5.5 σ _p ＝1/40 ε＝1

m _c ＝5.5 σ _c ＝1/12 b _ias ＝55

When processing the rg component, the individual values in the contrast sensitivity function CSF can be taken as

γ(ω)＝0.001531ω ² -0.06149ω+1.140

m _p ＝1.5 σ _p ＝1/70 ε＝0

When processing by components, the respective values in the contrast sensitivity function CSF can be taken as

γ(ω)＝0.001919ω ² -0.06427ω+1.090

m _p ＝1.5 σ _p ＝1/45 ε＝1/4

m _c ＝7.5 σ _c ＝1/20 b _ias ＝0

In the aboveIs the average value of the three components wk, rg and by of the color space;

according to the formula, the transfer function of the human visual system CSF model of three components of the color space wk, rg and by can be calculated; it should be noted that the unit of the spatial frequency ω involved in the above formula is a period/degree, that is, a formula given from the viewpoint of an observer, and the unit of the spectrogram obtained by fourier transform is a period/pixel, so that the two units are not identical, and it is necessary to perform conversion of the unit when calculating the spatial frequency ω of CSF using the following formula:

wherein H is _R To measure the number of pixels of the screen level in the image, V _D For visual vertical distance, W _L Is the width (mm) of the display screen;

in addition, the calculation involves calculationWhen the three components are negative, the average value of the wk, rg and by components is required to be subjected to modulus taking, so that the average value of the three components is always positive, and the data obtained by calculating the CSF model can be consistent with the experimental result;

then, frequency domain filtering is carried out on the three components (WK, RG, BY) BY using a transfer function obtained BY calculating the CSF model, and matrix point multiplication is carried out on a spectrogram of the three components (WK, RG, BY) and a corresponding CSF function, so that the frequency domain filtering can be completed; it should be noted that the CSF function values corresponding to the three components (WK, RG, BY) are larger to prevent the subsequent computation from being abnormal, the invention firstly performs synchronous normalization on the CSF function values of the three components before the frequency domain filtering, that is, finds the maximum value in the CSF function values of the three components, divides the maximum value BY the maximum value respectively, and ensures that the value of the CSF function of the three components is always within [0,1] to facilitate the subsequent computation; finally, the three component frequency domain maps filtered BY the CSF function are inverse fourier transformed into spatial domain components (WK, RG, BY).

In the present invention, preferably, in the fourth step:

the conversion matrix for transforming the color space (wk, rg, by) into CIE-XYZ space is as follows:

X＝0.626555wk-1.867178rg-0.153156by

Y＝1.369855wk+0.934756rg+0.436229by

Z＝1.505651wk+1.421324rg+2.536021by

then, the image is further transformed from CIE-XYZ space to CIE-Lab space to obtain three components L, a and b;

L＝116f(Y/Y _n )-16

a＝500[f(X/X _n )-f(Y/Y _n )]

b＝200[f(Y/Y _n )-f(Z/Z _n )]

wherein f (x) is a piecewise function

X in the above formula _n ,Y _n ,Z _n The values with highest occurrence frequency of X, Y and Z components in CIE-XYZ space are obtained by counting the rounded values of each component;

from a, b, the chrominance component C can be calculated according to the formula, i.e

In the present invention, preferably, in the fifth step:

after obtaining the luminance component L and the chrominance component C, the minimum perceived difference maps (JND) of the luminance and chrominance components are then calculated, respectively, according to the Weber theorem

Wherein delta represents the luminance component L and the chrominance component C,is a gaussian window function, then β is the minkofsky distance coefficient, typically β has a value of 2.4;

from the above local JND graph, the global JND values for the luminance and chrominance components can be further generalized using minkofsk basis functions, which can be called Pooling (Pooling), so its expression is:

when gamma-infinity, a global maximum JND value of both luminance and chrominance components is obtained, i.e

JND _{Δ_max} ＝max[JND(x,y)]Δ∈{L,C}；

Next, the ratio of the Mura defect area of the luminance and chrominance JND map to the total screen area is calculated as follows

Wherein p (x, y) and q (x, y) are each

In the present invention, preferably, in step six:

the luminance evaluation index LEV is lev=jnd _{L_area1} +0.28JND _{L_max} +6.47JND _{L_area2}

Wherein JND _{L_area1} JND is the ratio of Mura defect area with brightness JND graph value more than or equal to 1 to total screen area _{L_max} JND is the global maximum of the luminance JND graph _{L_area2} The ratio of the Mura defect area with the brightness JND diagram value more than or equal to 2 to the total screen area is represented;

the chromaticity evaluation index CEV is cev=jnd _{C_area1} +0.16JND _{C_max} +3.19JND _{C_area2}

Wherein JND _{C_area1} JND is the ratio of Mura defect area with chroma JND picture value more than or equal to 1 to total screen area _{C_max} JND is the global maximum of the chroma JND graph _{C_area2} Representing the ratio of the Mura defect area with the chromaticity JND diagram value more than or equal to 2 to the total screen area;

wherein N is the total number of display panel images shot at multiple angles, LEV _i I epsilon {1,2,3, …, N } is a brightness evaluation index under multiple angles, CEV _i I epsilon {1,2,3, …, N } is a chromaticity evaluation index under multiple angles, gamma is the ratio between the two terms, where 0.6 is generally taken, p is a Minkofski coefficient, and 2 is generally taken;

the luminance and chromaticity evaluation indexes are comprehensively calculated according to the following formula to obtain a final display panel Mura defect quantization evaluation value MEV which is mev=2.353 LEV _total +3.573CEV _total ；

The display panel quality can be quantitatively evaluated according to the calculated Mura defect quantitative evaluation value MNEV.

The invention has realized corresponding method and system, and verify through the analog Mura of 20 Mura types and collection Mura image on the actual display panel, wherein the closer the correlation coefficient is 1, the higher the consistency between the result of the patent method and subjective judgment is, as shown in figure 4, the total average correlation coefficient between the evaluation result of the patent evaluation method on the actual collection display panel image and subjective visual evaluation result is 0.934;

fig. 5 shows the correlation coefficients of subjective evaluation and algorithm evaluation results of the method in the 20 main Mura defects, and it can be seen that 16 kinds of correlation coefficients in the 20 Mura types are more than 0.92, which proves that the method has feasibility.

Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.

Claims (7)

1. The display panel Mura defect global evaluation method based on the minimum perceived difference is characterized by comprising the following steps of:

firstly, the system acquires complete images of a display panel from multiple angles through a high-definition camera, positions a screen area for cutting, scaling and denoising and filtering, converts a preprocessed RGB color gamut image into a CIE-XYZ color gamut image, and further converts the CIE-XYZ color gamut image into a color matching space (wk, rg, by);

step two, calculating a test contrast diagram and a reference contrast diagram, generating a differential contrast diagram by the test contrast diagram and the reference contrast diagram, and introducing an edge mask to the Mura defect existing in the image part to process the image;

step three, performing Fourier transform on three components of a color space (wk, rg and by) respectively to obtain frequency components, calculating contrast sensitivity CSF functions describing a human visual system, filtering the three frequency components of wk/rg/by respectively by applying single-channel Contrast Sensitivity Functions (CSF) of the components, and performing inverse Fourier transform to the corresponding space domain;

step four, converting the space domain (wk, rg, by) image of the inverse Fourier transform to CIE-XYZ space, further converting to CIE-Lab space, and calculating a luminance component L and a chrominance component C under the CIE-Lab color gamut;

step five, respectively calculating a minimum perceived difference graph (JND) of a brightness component L and a chrominance component C, and calculating a global maximum JND value of the brightness component and the chrominance component and a ratio of a Mura defect area to a total screen area according to a formula;

and step six, calculating evaluation indexes LEV and CEV of the brightness component and the chrominance component according to the global maximum JND value and the Mura defect area occupation ratio, summarizing the evaluation results of the multi-angle shot images, and finally comprehensively evaluating the brightness and chrominance evaluation indexes to obtain a final Mura defect quantitative evaluation value MEV of the display panel.

2. The minimum perceived difference-based display panel Mura defect global assessment method of claim 1, wherein: in the first step:

the multi-angle shooting display panel of the camera can generally take 3 angles, the camera faces to the center point of the screen, and the axis of the lens coincides with the central axis of the screen at a viewing angle 1; the camera is opposite to the center point of the screen and is positioned at the left side of the screen and at a viewing angle 2 forming an angle theta 1 with the central axis of the screen; the camera is opposite to the center point of the screen and is positioned at the right side of the screen and at a viewing angle 3 forming an angle theta 2 with the central axis of the screen;

the specific conversion formula for converting the RGB color gamut image into the CIE-XYZ color gamut image is as follows:

the invention firstly normalizes the RGB value to transform the range to 0-1, denoising filtering can be completed by combining Gaussian filtering and median filtering of a spatial domain, three channels of median filtering respectively carry out [5,5] convolution kernel size, and the image is expanded by copying the value of an edge region so as to avoid the condition that the value of four corners of the filtered image is 0; then, the XYZ values are obtained by using the formula and multiplied by 100, and the range of the XYZ values is converted into 0-100 so as to be used for subsequent calculation;

the CIE-XYZ gamut image is transformed into a color pair space according to the following conversion matrix:

wk＝0.279X+0.72Y-0.107Z

rg＝-0.449X+0.29Y-0.077Z。

by＝0.086X-0.59Y+0.501Z

3. the minimum perceived difference-based display panel Mura defect global assessment method of claim 1, wherein: in the second step,:

differential contrast map DC (x, y) is derived from test contrast map C _Test (x, y) and reference contrast map C _Refence (x, y) generation, in particular as follows

Wherein C is _Test (x, y) is the original image converted into the color space (wk, rg, by), C _Refence (x, y) filtering the background image of the original image after filtering all mura defects by Gaussian filtering with a frequency low enough as shown in the following formula

X, y in the above formula, the rows and columns of the image;

r—visual resolution, unit: px/deg;

s-Gaussian scale, units: the deg, generally 2deg;

if mura defects exist in the edge part of the image, the conventional method is very easy to detect the mura defects, so that the image needs to be processed by using an edge mask, and the specific algorithm is as follows

Wherein M, N-image height and width, unit: pix;

b _gain the boundary gain can be generally 1;

b _scale boundary ratio, unit: degs, generally 0.5 degs are available;

and (3) carrying out dot multiplication on each pixel of the differential contrast map DC (x, y) and the edge mask function B (x, y) one by one, so that the accuracy of an algorithm on edge mura defect detection can be improved.

4. The minimum perceived difference-based display panel Mura defect global assessment method of claim 1, wherein: in the third step:

the formula for calculating the contrast sensitivity function CSF of the human visual system is as follows:

wherein the method comprises the steps of

ω＝(u ² +v ² ) ^1/2

(u, v) is the spatial frequency coordinates in cycles per degree, the angle being the angle between the human eye to a point on the screen and the vertical to the screen;

when processing the wk component, the values in the contrast sensitivity function CSF can be taken as

γ(ω)＝0.003100ω ² -0.10680ω+1.396

m _p ＝5.5 σ _p ＝1/40 ε＝1

m _c ＝5.5 σ _c ＝1/12 b _ias ＝55

When processing the rg component, the individual values in the contrast sensitivity function CSF can be taken as

γ(ω)＝0.001531ω ² -0.06149ω+1.140

m _p ＝1.5 σ _p ＝1/70 ε＝0

When processing by components, the respective values in the contrast sensitivity function CSF can be taken as

γ(ω)＝0.001919ω ² -0.06427ω+1.090

m _p ＝1.5 σ _p ＝1/45 ε＝1/4

m _c ＝7.5 σ _c ＝1/20 b _ias ＝0

In the aboveIs the average value of the three components wk, rg and by of the color space;

according to the formula, the transfer function of the human visual system CSF model of three components of the color space wk, rg and by can be calculated; it should be noted that the unit of the spatial frequency ω involved in the above formula is a period/degree, that is, a formula given from the viewpoint of an observer, and the unit of the spectrogram obtained by fourier transform is a period/pixel, so that the two units are not identical, and it is necessary to perform conversion of the unit when calculating the spatial frequency ω of CSF using the following formula:

wherein H is _R To measure the number of pixels of the screen level in the image, V _D For visual vertical distance, W _L Is the width (mm) of the display screen;

in addition, the calculation involves calculationWhen the three components are in a negative number, the average value of the wk, rg and by components is required to be subjected to modulo, so that the average value of the three components is always positive, and the number calculated by the CSF model is obtainedThe experimental result can be consistent with the experimental result;

then, frequency domain filtering is carried out on the three components (WK, RG, BY) BY using a transfer function obtained BY calculating the CSF model, and matrix point multiplication is carried out on a spectrogram of the three components (WK, RG, BY) and a corresponding CSF function, so that the frequency domain filtering can be completed; it should be noted that the CSF function values corresponding to the three components (WK, RG, BY) are larger to prevent the subsequent computation from being abnormal, the invention firstly performs synchronous normalization on the CSF function values of the three components before the frequency domain filtering, that is, finds the maximum value in the CSF function values of the three components, divides the maximum value BY the maximum value respectively, and ensures that the value of the CSF function of the three components is always within [0,1] to facilitate the subsequent computation; finally, the three component frequency domain maps filtered BY the CSF function are inverse fourier transformed into spatial domain components (WK, RG, BY).

5. The minimum perceived difference-based display panel Mura defect global assessment method of claim 1, wherein: in the fourth step:

the conversion matrix for transforming the color space (wk, rg, by) into CIE-XYZ space is as follows:

X＝0.626555wk-1.867178rg-0.153156by

Y＝1.369855wk+0.934756rg+0.436229by

Z＝1.505651wk+1.421324rg+2.536021by

then, the image is further transformed from CIE-XYZ space to CIE-Lab space to obtain three components L, a and b;

L＝116f(Y/Y _n )-16

a＝500[f(X/X _n )-f(Y/Y _n )]

b＝200[f(Y/Y _n )-f(Z/Z _n )]

wherein f (x) is a piecewise function

X in the above formula _n ,Y _n ,Z _n X, Y in CIE-XYZ spaceThe value with the highest frequency of occurrence of the Z component can be generally obtained by counting the value obtained by rounding each component;

from a, b, the chrominance component C can be calculated according to the formula, i.e

6. The minimum perceived difference-based display panel Mura defect global assessment method of claim 1, wherein: in the fifth step:

after obtaining the luminance component L and the chrominance component C, the minimum perceived difference maps (JND) of the luminance and chrominance components are then calculated, respectively, according to the Weber theorem

Wherein delta represents the luminance component L and the chrominance component C,is a gaussian window function, then β is the minkofsky distance coefficient, typically β has a value of 2.4;

from the above local JND graph, the global JND values for the luminance and chrominance components can be further generalized using minkofsk basis functions, which can be called Pooling (Pooling), so its expression is:

when gamma-infinity, a global maximum JND value of both luminance and chrominance components is obtained, i.e

JND _{Δ_max} ＝max[JND(x,y)] Δ∈{L,C}；

Next, the ratio of the Mura defect area of the luminance and chrominance JND map to the total screen area is calculated as follows

Wherein p (x, y) and q (x, y) are each

7. The minimum perceived difference-based display panel Mura defect global assessment method of claim 1, wherein: in the sixth step:

the luminance evaluation index LEV is lev=jnd _{L_area1} +0.28JND _{L_max} +6.47JND _{L_area2}

Wherein JND _{L_area1} JND is the ratio of Mura defect area with brightness JND graph value more than or equal to 1 to total screen area _{L_max} JND is the global maximum of the luminance JND graph _{L_area2} The ratio of the Mura defect area with the brightness JND diagram value more than or equal to 2 to the total screen area is represented;

the chromaticity evaluation index CEV is cev=jnd _{C_area1} +0.16JND _{C_max} +3.19JND _{C_area2}

Wherein JND _{C_area1} JND is the ratio of Mura defect area with chroma JND picture value more than or equal to 1 to total screen area _{C_max} JND is the global maximum of the chroma JND graph _{C_area2} Representing the ratio of the Mura defect area with the chromaticity JND diagram value more than or equal to 2 to the total screen area;

wherein N is the total number of display panel images shot at multiple angles, LEV _i I epsilon {1,2,3, …, N } is a brightness evaluation index under multiple angles, CEV _i I epsilon {1,2,3, …, N } is a chromaticity evaluation index under multiple angles, gamma is the ratio between the two terms, where 0.6 is generally taken, p is a Minkofski coefficient, and 2 is generally taken;

the luminance and chromaticity evaluation indexes are comprehensively calculated according to the following formula to obtain a final display panel Mura defect quantization evaluation value MEV which is mev=2.353 LEV _total +3.573CEV _total ；

Thus, the display panel quality can be quantitatively evaluated according to the calculated Mura defect quantitative evaluation value MEV.