Screen uniformity detection method based on linear array CCD vertical scanning imaging

Abstract

The invention belongs to the technical field of visual imaging detection, and particularly relates to a screen uniformity detection method based on linear array CCD vertical imaging scanning. According to the method, vertical imaging of a display screen pixel array is achieved through the linear CCD camera, 2D imaging is carried out through the data preprocessing and splicing algorithm, image distortion at different positions is effectively avoided, the movement complexity of the camera is reduced, the stability of a scanning path is improved, in addition, the prismatic cylindrical mirror is adopted for effectively restraining the influence of incident light, and the imaging precision is improved. The problem that a high-quality image is obtained under the interference of ambient light in the field of screen detection is fundamentally and effectively solved, the problems of large size and compatibility of existing industrial screen uniformity detection are solved, the calculation complexity is reduced, the real-time detection performance is improved, and an effective means is provided for 2D imaging and detection of a display screen.

Description

Technical Field

[0001] The invention belongs to the technical field of visual imaging detection, and in particular relates to a screen uniformity detection method based on linear array CCD vertical scanning imaging. Background Art

[0002] At present, high-precision uniformity detection of display screens has always been a technical problem that has troubled manufacturers. This is limited by the current camera visual imaging mechanism. Due to the different imaging angles at different positions and the structural characteristics of the display unit composed of backlight, liquid crystal, and Fresnel lens, the brightness obtained by imaging is difficult to directly reflect the display uniformity. In addition, in luminous imaging, factors such as glow and other imaging crosstalk and camera pixel consistency have further increased the difficulty of display screen uniformity detection. Therefore, developing new technologies that break through the defects of imaging principles has become an inevitable way to solve this technical problem. Summary of the invention

[0003] In view of the above problems, the present invention proposes a vertical scanning imaging and analysis technology based on a linear array CCD. The present invention installs a linear CCD camera at the end of a six-axis robot arm to realize vertical imaging of the pixel array of the display screen using the linear CCD camera, and adopts data preprocessing and stitching algorithms to perform 2D imaging. At the same time, the linear array camera data is analyzed and processed using an algorithm to finally complete the uniformity evaluation of the display screen. By vertically sweeping the scan, image distortion at different positions can be avoided, the movement complexity of the camera can be reduced, and the stability of the scanning path can be improved. Secondly, through the cooperation of the prismatic cylindrical mirror, the angle and intensity of the incident light can be effectively controlled, and the high-quality image can be obtained under the interference of ambient light. It also overcomes the large volume compatibility problem of the current industrial screen uniformity detection, reduces the computational complexity, and improves the real-time detection performance.

[0004] The technical solution of the present invention is:

[0005] The screen uniformity detection method based on linear array CCD vertical scanning imaging includes the following steps:

[0006] S1. Scan the display screen to be tested in a vertical plane to obtain scanning data. The vertical plane scanning method is to set the line CCD lens to be perpendicular to the surface of the display screen to be tested, and set a prismatic cylindrical mirror between the line CCD and the display screen to be tested, and control the line CCD to scan the display screen to be tested in a column-by-column manner, and scan the display screen to be tested from one side to the other side in sequence to obtain multiple columns of scanning data, wherein the width of each column is controlled by the horizontal displacement distance of the control line CCD, and there is an overlapping repeated scanning area between adjacent scanning columns;

[0007] S2. Perform image processing using the obtained multiple columns of scanning data to obtain a 2D image of the display screen to be tested. The specific method is as follows:

[0008] The multiple columns of scanned data obtained are defined as C 1 , C 2 ,……,C N , N is the number of columns, and a single column of pixels is denoted as C N =[p 1 ,p 2 ,...,p m ], m is the number of pixels in a single column, filtering and denoising are performed, and the processed data is defined as C 1 '、C' 2 ,……,C' N , calculate the repeated scanning area of ​​two adjacent columns, remove the redundant repeated data and get the final column data defined as C 1 ”、C' 2 ', ..., C" N , stitching the final column data according to the spatial position to obtain the final 2D image;

[0009] S3. Use the obtained 2D image to detect screen uniformity.

[0010] Furthermore, the calculation of the i-th pixel position of the filtering and denoising process is shown in formula (1):

[0011]

[0012] Among them, the parameter p i+j represents the i+jth pixel value in the column data, k represents the radius of the filter window, and G(j) represents the Gaussian weight function:

[0013]

[0014] The weight of the i-th position is determined by G(j), where j represents the pixel with the same pixel as the center pixel p. i , σ represents the standard deviation of the Gaussian function, which controls the smoothness of the filter.

[0015] Furthermore, the calculation method of the repeated scanning area is as shown in formula (3) and formula (4):

[0016] First, the original data c N The columns are denoised as follows:

[0017]

[0018] Then, the column data is smoothed by the repeated region fusion algorithm, and the data after eliminating the fusion region is aligned and spliced. The repeated region fusion algorithm is shown in the following formula (4).

[0019]

[0020] In formula 4:

[0021] C” N is the column pixel value after fusion, w N Represents the weight of the corresponding pixel value, C' N--1 (m) and C' N (m) represents two columns of pixel values ​​of the repeated area.

[0022] Finally, all processed column data are concatenated to obtain 2D image data.

[0023]

[0024] In formula (5), I final For stitching, we get the image matrix, W col is the width of each column, x start The insertion position of the i-th column data.

[0025] The beneficial effects of the present invention are:

[0026] (1) In the screen uniformity detection of the present invention, the line CCD camera is used to vertically scan and can capture tiny light spots or uneven areas on the screen. By vertically scanning, image distortion at different positions can be avoided, ensuring more accurate detection of unevenness such as brightness and color difference.

[0027] (2) Through the cooperation of prismatic cylindrical mirrors, the angle and intensity of the incident light can be effectively controlled, thus overcoming the interference of ambient light to obtain high-quality images and improving the accuracy of display screen uniformity detection.

[0028] (3) It overcomes the large volume and compatibility issues of current industrial screen uniformity detection, reduces computational complexity, and improves real-time detection performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 Line CCD camera vertical scanning imaging technology flow chart

[0030] Figure 2 It is a schematic diagram of the line scan imaging process;

[0031] Figure 3 Line scan imaging microscope block diagram

[0032] Figure 4 This is a 2D image processing flow chart;.

[0033] Figure 5 This is a schematic diagram of the structure of this example. DETAILED DESCRIPTION

[0034] The technical solution of the present invention is described in detail below in conjunction with the accompanying drawings and embodiments:

[0035] like Figure 1 As shown, the specific implementation of the present invention can be described as the following steps:

[0036] (1) The robot moves at a constant speed along the main axis;

[0037] (2) Line CCD scanning imaging;

[0038] (3) Image stitching;

[0039] (4) Scanning ends;

[0040] (5) Stitching is completed.

[0041] The specific implementation of step (1) is as follows:

[0042] The six-axis robot moves at a constant speed along the main axis of the screen, so that the line CCD camera is perpendicular to the screen at 90°, and the entire screen is scanned column by column. Each time a column of data is scanned, the robot arm moves one step to the right until the entire plane is scanned.

[0043] like Figure 2 As shown, the specific implementation of step (2) is:

[0044] (2.1) When the line CCD camera acquires the screen image reflected by the imaging cylindrical mirror, the screen pixel scanning image is obtained, including the column scanning data C 1 , column sweep data C 2 , ..., column sweep image C N , completing the sweep of the entire plane.

[0045] like Figure 3 As shown, the specific implementation method of step (3) is as follows:

[0046] (2.2) The imaging object passes through the cylindrical mirror to form a mirror image of the imaging object, where the surface element pixel point A of the imaging object corresponds to the point A' of the line CCD photosensitive pixel, and the point B of the imaging object corresponds to the point B' of the line CCD photosensitive pixel.

[0047] like Figure 4 As shown, the specific implementation of step (4) is:

[0048] (4.1) The column data C obtained by scanning the line CCD 1 , Line CCD scan column data C 2 , ..., line CCD scan column data C N , the denoising algorithm calculates the line sweep column data C 1 ', Line sweep column data C' 2 , ..., line sweep column data C' N , and get the processed column data.

[0049] (4.2) The scanned data obtained after processing (4.1) is subjected to repeated fusion region algorithm, and the column data C 1 'With C' 2 Calculate the repeated area and column data C' 2 With C 3 'Calculate repeated areas, column data C' N-1 With C' N Calculate the duplicate regions.

[0050] (4.3) The column data C after eliminating the fusion area 1 "、Column data C' 2 ', ..., column data C" N The final 2D image is obtained by stitching according to the spatial position.

[0051] Example

[0052] like Figure 5 As shown, the present invention provides a line CCD vertical scanning imaging structure, including two sets of image acquisition devices,

[0053] It comprises a six-axis robot arm 1, a camera fixture 2 is placed at the end of the robot arm, a CCD camera 3 is clamped under the fixture, a cylindrical mirror device 4 is placed under the camera 3, a display screen 5 and a six-axis robot arm carrier 6.

[0054] In this embodiment, the line CCD camera is fixed on the robot arm through a clamp and moves along with the robot arm.

[0055] In this embodiment, the device is first placed at a fixed position on the production line, and the conveyor belt is waited for to reach the designated area before taking pictures. First, the six-axis robotic arm moves to the initial point in the upper left corner of the screen to trigger the picture taking. Then the robotic arm translates a distance from left to right and triggers the picture taking again until the entire plane is covered. The obtained data is spliced ​​to obtain a two-dimensional image, and the uniformity of the screen area is detected.

[0056] The line CCD vertical scanning imaging image in this example is realized as follows:

[0057] S1, input start signal;

[0058] S2. According to the start signal, the CCD camera is controlled by the six-axis robot to move at a constant speed from left to right;

[0059] S3, the line CCD camera performs a plane sweep 2D imaging on the imaging object and obtains the line CCD column data C 1 , Line CCD column data C 2 , ..., line CCD column data C N ;

[0060] S301: the imaging object passes through the cylindrical mirror to form a mirror image of the imaging object, wherein the imaging object surface element point A corresponds to the photosensitive pixel point A' of the imaging object, and the imaging object point B corresponds to the photosensitive pixel point B' of the imaging object;

[0061] After S4 completes the scanning of the CCD camera through the host computer control line, it outputs the result to the host computer for processing, completing the 2D imaging of the vertical scanning of the imaging object;

[0062] S401, the column data C obtained by scanning the line CCD 1 , column data C obtained by line CCD scanning 2 , ..., column data C obtained by line CCD scanning N Obtain good pixels in sweep imaging through denoising algorithms;

[0063] S402: The column data C of the good pixels obtained in S401 1 ', column data of good pixels C' 2 , ..., good pixel column data C' N Column data C is obtained by repeated region detection algorithm 1 "、Column data C' 2 ', ..., column data C" N .

[0064] S403 , using a splicing algorithm to splice the column data that have undergone repeated region detection to obtain a final two-dimensional image.

[0065] The present invention uses a line CCD camera to scan vertically to generate a position sequence in different spaces, and then uses 2D image reconstruction to complete the 2D splicing reconstruction of the entire screen plane. Therefore, the image sequence obtained by the line CCD vertical scanning can effectively solve the influence of the imaging angle of the traditional camera and the structural characteristics of the display unit, and complete the detection of screen uniformity. In addition, the camera end of the present invention uses a line CCD plus a cylindrical mirror to achieve 2D imaging. At the same time, the line scanning detection speed is fast and is not limited to the size of the display screen. It is a low-cost display screen uniformity detection path. Applying it to the process of display screen uniformity detection can effectively improve the detection efficiency and detection accuracy, and has broad industrial application prospects.

Claims

1. Screen uniformity detection method based on linear array CCD vertical scanning imaging, characterized in that: The following steps are involved: S1. Scan the display screen to be tested in a vertical plane to obtain scanning data. The vertical plane scanning method is to set the line CCD lens to be perpendicular to the surface of the display screen to be tested, and set a prismatic cylindrical mirror between the line CCD and the display screen to be tested, and control the line CCD to scan the display screen to be tested in a column-by-column manner, and scan the display screen to be tested from one side to the other side in sequence to obtain multiple columns of scanning data, wherein the width of each column is controlled by the horizontal displacement distance of the control line CCD, and there is an overlapping repeated scanning area between adjacent scanning columns; S2. Perform image processing using the obtained multiple columns of scanning data to obtain a 2D image of the display screen to be tested. The specific method is as follows: The multiple columns of scanned data defined are C1, C2, ..., C N , N is the number of columns, and a single column of pixels is denoted as C N =[p1,p2,...,p m ], m is the number of pixels in a single column, and filtering and denoising are performed. The processed data are defined as C1', C'2, ..., C' N , calculate the repeated scanning area of ​​two adjacent columns, remove the redundant repeated data and get the final column data defined as C1”, C’2’, …, C” N , stitching the final column data according to the spatial position to obtain the final 2D image; S3. Use the obtained 2D image to detect screen uniformity.

2. The screen uniformity detection method based on linear array CCD vertical scanning imaging according to claim 1 is characterized in that: The calculation method of the filtering and denoising process for the i-th pixel position is: Among them, the parameter p i+j represents the i+jth pixel value in the column data, k represents the radius of the filter window, and G(j) represents the Gaussian weight function: The weight of the i-th position is determined by G(j), where j represents the pixel with the same pixel as the center pixel p. i , σ represents the standard deviation of the Gaussian function.

3. The screen uniformity detection method based on linear array CCD vertical scanning imaging according to claim 1 is characterized in that: The calculation method of the repeated scanning area is: First, the original data c N Column denoising: Then, the column data is smoothed by the repeated region fusion algorithm, and the data after eliminating the fusion region is aligned and spliced. The repeated region fusion algorithm is as follows: Among them, C N is the column pixel value after fusion, w N Represents the weight of the corresponding pixel value, C' N--1 (m) and C' N (m) represents two columns of pixel values ​​of the repeated area; Finally, all processed column data are concatenated to obtain 2D image data: Among them, I final For stitching, we get the image matrix, W col is the width of each column, x start The insertion position of the i-th column data.

Images