Segmentation method for image in LED screen seam region, LED screen bright and dark line correction method

Abstract

The present invention discloses a method for segmenting an image of a seam region of an LED screen and a method for correcting bright and dark lines of an LED screen. The segmentation method includes: performing a first binarization process on the seam region image to obtain a binarized seam region image; determining a plurality of linear regression sub-regions from the binarized seam region image; segmenting a target region image from the seam region image according to each linear regression sub-region, and each target region image respectively includes an image of edge beads located in the same row; performing a second binarization process on each target region image to obtain a binarized target region image; performing linear fitting on pixel points satisfying a gray value condition in each binarized target region image to obtain a central straight line, and the central straight line is used as a connection line of the center points of each edge bead in the same row. The center points of the beads are accurately found through statistical calculation methods, avoiding problems such as contour search failure that may occur when using the bead contour to find the center point, and at the same time solving the problem of bead adhesion.

Description

Technical Field

[0001] The present invention belongs to the technical field of LED screens, and specifically relates to an LED screen image segmentation method and an LED screen bright-dark line correction method. Background Art

[0002] An LED screen is assembled by splicing LED small modules. During the actual splicing process, manual errors in the splicing operation will cause uneven gaps at the splicing points, resulting in a bright-dark line effect on the screen. That is, due to errors in the lamp bead spacing at the splicing seam compared to other places, it appears brighter or darker visually. To eliminate the bright-dark line, it is necessary to change the brightness at the splicing seam so that the brightness at the splicing seam is visually consistent with that of other positions.

[0003] Before leaving the factory, all the lamp beads on a single LED module will be photographed at a very close distance using a professional camera, and then a segmentation algorithm will be used to very precisely segment each lamp bead in the picture to calculate the brightness within the segmented area. The algorithm compares the brightness and then adjusts the control parameters to make all the lamp beads have consistent brightness in the three colors of red, green, and blue. However, this cannot solve the problem of bright-dark lines at the splicing points caused by errors during manual splicing of modules or boxes because at this time the LED screen has been assembled, the whole screen is very large, and it is not suitable for the previous adjustment method. At the same time, LED screens are generally assembled at the use site. When the screen is very large or the installation height is very high, it will be difficult for manual use of a professional camera to take pictures of each small module, and the process of taking pictures with a professional camera is also relatively slow. A faster and simpler method for extracting LED module images needs to be considered.

[0004] For this reason, automatic correction methods have been proposed in the prior art. However, existing automatic correction methods require clear and separated single lamp beads to be photographed, and a professional camera is needed to collect the lamp bead images at the splicing seam. This requires people to maintain the shooting stability. When the screen size is very large, the collector needs to constantly change their position to collect pictures. Since the methods used for bright-dark line correction after collecting pictures all need to segment each lamp bead at the splicing seam to extract the center point of the lamp bead, when the shooting is unclear, that is, the lit lamp beads are adhered in the photo and cannot be segmented, the existing methods will lose their effectiveness. Some researchers have proposed using drones for shooting, but the pictures taken by drones are even less clear, and there will be jitter, making it more likely for the lamp beads in the pictures to adhere. The previous correction algorithms are not applicable to the pictures taken by drones. Summary of the Invention

[0005] The technical problem solved by the present invention is: how to segment the LED screen image taken by a drone to determine the center position of the lamp beads, so as to accurately correct the bright-dark lines of the screen subsequently.

[0006] The present application discloses a method for segmenting an image of a stitching area of an LED screen. The segmentation method includes:

[0007] Performing a first binarization process on the stitching area image to obtain a binarized stitching area image;

[0008] Determining a plurality of linear regression sub-regions from the binarized stitching area image;

[0009] Segmenting a target area image from the stitching area image according to each linear regression sub-region, and each target area image respectively includes images of edge lamp beads located in the same row;

[0010] Performing a second binarization process on each target area image to obtain a binarized target area image;

[0011] Performing linear fitting on the pixel points satisfying the gray value condition in each binarized target area image to obtain a central straight line, and the central straight line is used as the connection line of the central points of each edge lamp bead in the same row.

[0012] Preferably, the method for determining a plurality of linear regression sub-regions from the binarized stitching area image includes:

[0013] Performing a closing operation on the binarized stitching area image, and cropping both sides of the image after the closing operation along the vertical direction of the edge of the LED module stitching;

[0014] Performing polygon fitting on the cropped image to obtain four vertices, and sequentially connecting the four vertices to form a rectangular area;

[0015] Equally dividing the rectangular area along the parallel direction of the edge of the LED module stitching to obtain a plurality of linear regression sub-regions, where the number of the linear regression sub-regions is equal to the number of rows of edge lamp beads in the stitching area.

[0016] Preferably, the method for performing linear fitting on the pixel points satisfying the gray value condition in each binarized target area image to obtain a central straight line is:

[0017] Using the least squares linear regression algorithm to fit the coordinate values of each pixel point with a gray value greater than zero in the binarized target area image to obtain a central straight line.

[0018] Preferably, the gray value threshold of the first binarization process is less than the gray value threshold of the second binarization process.

[0019] The present application discloses a device for segmenting an image of a stitching area of an LED screen. The segmentation device includes:

[0020] The first binarization processing module is used to perform the first binarization processing on the seam area image to obtain the binarized seam area image;

[0021] The linear regression sub-region determination module is used to determine a plurality of linear regression sub-regions from the binarized seam area image;

[0022] The target region image segmentation module is used to segment the target region image from the seam area image according to each linear regression sub-region, and each target region image respectively contains the images of the edge beads located in the same row;

[0023] The second binarization processing module is used to perform the second binarization processing on each target region image to obtain the binarized target region image;

[0024] The linear fitting module is used to perform linear fitting on the pixel points that meet the gray value condition in each binarized target region image to obtain the central straight line, and the central straight line is used as the connection line of the center points of each edge bead in the same row.

[0025] The present application also discloses a computer-readable storage medium, and the computer-readable storage medium stores a segmentation program for the LED screen seam area image. When the segmentation program for the LED screen seam area image is executed by a processor, the above-mentioned segmentation method for the LED screen seam area image is implemented.

[0026] The present application also discloses a computer device, and the computer device includes a computer-readable storage medium, a processor, and a segmentation program for the LED screen seam area image stored in the computer-readable storage medium. When the segmentation program for the LED screen seam area image is executed by the processor, the above-mentioned segmentation method for the LED screen seam area image is implemented.

[0027] The present application also discloses a method for correcting the bright and dark lines of an LED screen, and the method for correcting the bright and dark lines of the LED screen includes:

[0028] Obtain the seam area image of the LED screen to be corrected, and the seam area image includes at least two rows of edge bead images of each LED module in two spliced LED modules, and the arrangement direction of each row of edge beads is parallel to the LED module seam edge;

[0029] Adopt the above-mentioned segmentation method to perform segmentation processing on the seam area image to obtain the central straight line corresponding to each row of edge beads;

[0030] Calculate the distance between the beads in each bead unit to be corrected according to the function expressions of each central straight line, and each bead unit to be corrected includes the beads in each row of edge beads along the vertical direction of the LED module seam edge and in the same row;

[0031] Determine the correction coefficient of the to-be-corrected lamp bead unit according to the distance between each lamp bead;

[0032] Adjust the brightness of each to-be-corrected lamp bead unit according to each of the correction coefficients.

[0033] Preferably, the method for determining the correction coefficient of the to-be-corrected lamp bead unit according to the distance between each lamp bead is to calculate the correction coefficient k according to the following formula i :

[0034]

[0035] where dis i represents the distance between two lamp beads on both sides of the splicing position in the i-th to-be-corrected lamp bead unit, and truedis i represents the average value of the distances between two adjacent lamp beads in the same LED module in the i-th to-be-corrected lamp bead unit.

[0036] Preferably, the method for determining the correction coefficient of the to-be-corrected lamp bead unit according to the distance between each lamp bead is to calculate the correction coefficient K according to the following formula:

[0037]

[0038] where dis i represents the distance between two lamp beads on both sides of the splicing position in the i-th to-be-corrected lamp bead unit, truedis i represents the average value of the distances between two adjacent lamp beads in the same LED module in the i-th to-be-corrected lamp bead unit, and N represents the total number of to-be-corrected lamp bead units.

[0039] This application also discloses another method for correcting the bright-dark line of an LED screen. The method for correcting the bright-dark line of the LED screen includes:

[0040] Obtain the image of the splicing area of the to-be-corrected LED screen. The image of the splicing area includes at least two rows of edge lamp bead images of each LED module in two spliced LED modules, and the arrangement direction of each row of edge lamp beads is parallel to the splicing edge of the LED module;

[0041] Perform segmentation processing on the image of the splicing area by using the above segmentation method, and the corresponding central straight line of each row of edge lamp beads;

[0042] Divide the image of the splicing area into a splicing area and a non-splicing area according to each of the central straight lines;

[0043] Calculate the grayscale mean value of the splicing area and the grayscale mean value of the non-splicing area respectively, and use the ratio of the grayscale mean value of the non-splicing area to the grayscale mean value of the splicing area as the correction coefficient;

[0044] Adjust the brightness of each lamp bead in the seam area according to the correction coefficient.

[0045] A method for segmenting an image of a seam area of an LED screen and a method for correcting bright and dark lines of an LED screen disclosed by the present invention have the following technical effects:

[0046] The method for segmenting an image of a seam area of an LED screen accurately finds the center points of lamp beads through statistical calculation, avoiding problems such as possible failure in contour finding that may occur when using the lamp bead contour to find the center point, and at the same time solving the problem of lamp bead adhesion at the seam of the LED module, laying a foundation for subsequent bright and dark line correction. Further, according to the segmentation result, the correction coefficient can be calculated based on the distance or grayscale difference, enabling precise adjustment of the brightness of each lamp bead, or overall brightness adjustment of the lamp beads in the seam area, effectively reducing the influence of bright and dark lines. Description of the Drawings

[0047] Figure 1 It is a flowchart of the LED screen image segmentation method according to Embodiment 1 of the present invention;

[0048] Figure 2 It is a lighting diagram of a single LED module according to Embodiment 1 of the present invention;

[0049] Figure 3 It is a partial image of the LED screen according to Embodiment 1 of the present invention;

[0050] Figure 4 It is a schematic diagram of the seam area image according to Embodiment 1 of the present invention;

[0051] Figure 5 It is a schematic diagram of the linear regression sub-region according to Embodiment 1 of the present invention;

[0052] Figure 6 It is an image of the seam area of an LED screen composed of small-pitch lamp beads according to Embodiment 1 of the present invention;

[0053] Figure 7 It is a schematic diagram after binarizing the seam area image according to Embodiment 1 of the present invention;

[0054] Figure 8 It is a binary image of the target area according to Embodiment 1 of the present invention;

[0055] Figure 9 It is a segmentation schematic diagram of the seam area image according to Embodiment 1 of the present invention;

[0056] Figure 10 Flow chart of the LED screen bright and dark line correction method according to Embodiment 2 of the present invention;

[0057] Figure 11 Schematic diagram of the bead unit to be corrected according to Embodiment 2 of the present invention;

[0058] Figure 12 Flow chart of the LED screen bright and dark line correction method according to Embodiment 3 of the present invention;

[0059] Figure 13 Schematic diagram of the division of the splicing area and the non - splicing area according to Embodiment 3 of the present invention;

[0060] Figure 14 Principle block diagram of the LED screen seam area image segmentation device according to Embodiment 4 of the present invention;

[0061] Figure 15 Schematic diagram of the computer device according to Embodiment 6 of the present invention. Detailed implementation manners

[0062] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0063] Before describing the various embodiments of the present application in detail, first briefly describe the technical concept of the present application: In the prior art, when using the automatic correction method to adjust the brightness of the beads in the seam area, it is first necessary to obtain a clear image of the beads in the lit state. However, when using a hand - held camera for shooting, due to factors such as jitter, or when using a drone for shooting, due to reasons such as low resolution, the bead images are adhered and cannot be accurately identified. For this reason, the present application provides a method for segmenting the LED screen seam area image. Based on the characteristics that the beads in the same row of the LED module are precisely on the same straight line and the bead brightness decreases from the center to the outside, combined with the image binarization processing method, the bead images in the same row are divided into the same sub - region, and the center points of each bead and the nearby pixels are extracted as the objects for linear fitting. After linear fitting, the center line is obtained, and the center line is used as the connection line of the center points of the edge beads in the same row. In this way, each row of beads can be accurately distinguished, thus providing the basic conditions for subsequent accurate brightness correction.

[0064] Specifically, as Figure 1 shown, the segmentation method of the LED screen seam area image in Embodiment 1 of the present invention includes the following steps:

[0065] Step S10: Perform the first binarization processing on the seam area image to obtain the seam area binarized image;

[0066] Step S20: Determine a number of linear regression sub-regions from the binary image of the seam region;

[0067] Step S30: Segment the target region image from the seam region image according to each linear regression sub-region, and each target region image respectively contains the images of the edge light beads located in the same row;

[0068] Step S40: Perform a second binary processing on each target region image to obtain a binary image of the target region;

[0069] Step S50: Perform a linear fitting on the pixel points that meet the gray value condition in each binary image of the target region to obtain a center line, and the center line is used as the connection line of the center points of each edge light bead in the same row.

[0070] Before describing each step of the first embodiment in detail, first describe the process of obtaining the seam region image of the LED screen. After splicing each LED module to form an LED screen, set the lighting mode of each LED module, keep at least one row of edge light beads on each side of the LED module lit, and turn off other light beads. As Figure 2 shown, the black circles represent lit light beads, and the blank circles represent extinguished light beads. Exemplarily, in the first embodiment, two rows of edge light beads on each side are lit, and other light beads are turned off. In order to facilitate distinguishing each seam region, the two light beads on both sides of each row of edge light beads are turned off. The partial image of the LED screen obtained after splicing is as Figure 3 shown, and the dotted lines in the figure represent the splicing points of adjacent LED modules.

[0071] A single seam region is composed of the edge light beads of two adjacent LED modules. Then, use the existing algorithm to extract the seam region in the photo through contour recognition and simple image processing operations. There are two types of extracted regions: horizontal seams and vertical seams. For unified processing, the algorithm regards the segmented seam picture as a gray matrix, and each pixel of the picture constitutes each element of the matrix. Transpose the matrix formed by the vertical seam so that the light beads in the same column within the seam become located in the same row, and complete the preprocessing operation of the seam picture to obtain the seam region image.

[0072] Exemplarily, as Figure 4As shown, the seam area contains four rows of edge beads. The identifiers of the four rows of beads are defined as the four rows from top to bottom: L1, L2, L3, and L4. The edge beads in the first row L1 and the second row L2 are located within the same LED module. The third row L3 and the fourth row L4 from top to bottom are located in another LED module. The distance between adjacent beads within the same module is fixed. The beads within the same module are assembled by machine with extremely small distance errors. However, the two rows of beads in the second row L2 and the third row L3 are located between different modules, and the splicing error between the modules is relatively large, which will produce a bright and dark line effect visually when lit, that is, the positions where the two rows of beads in the second row L2 and the third row L3 are located are brighter or darker than other parts of the screen. Here, taking the seam area containing four rows of edge beads as an example, the segmentation method will be described in detail.

[0073] First, set the grayscale threshold for the first binarization process to the maximum grayscale value of the seam area image minus eighty. Set the pixel points in the seam area image with grayscale values less than this grayscale threshold to zero grayscale uniformly, and set the pixel points in the seam area image with grayscale values greater than or equal to this grayscale threshold to a certain grayscale greater than zero, thereby forming a binarized image of the seam area.

[0074] Furthermore, the method for determining several linear regression sub-regions from the binarized image of the seam area includes the following steps:

[0075] Step S201: Perform a closing operation on the binarized image of the seam area, and crop the two side edges of the image after the closing operation along the vertical direction of the LED module seam edge. The kernel size of the closing operation is set to twice the width of the beads in the binarized image of the seam area, and the bead width is obtained by dividing the length of the binarized image of the seam area by the number of beads in a single row. After the closing operation, there is no gap between the beads. However, since the beads are circular, the number of vertices obtained directly using the polygon fitting algorithm is greater than four. Crop the image after the closing operation by two-thirds in the middle, that is, crop off the edges on both the left and right sides by one-sixth.

[0076] Step S202: Perform polygon fitting on the cropped image to obtain four vertices, and connect the four vertices in sequence to form a rectangular area. As Figure 5 shown, cropping and then performing polygon fitting using the polygon fitting algorithm can obtain four vertices. The four vertices are represented clockwise from the upper left corner as: a1, a2, a3, a4. Connect the four vertices to form a rectangular area.

[0077] Step S203: Divide the rectangular area equally along the parallel direction of the edge of the LED module seam to obtain a number of linear regression sub-areas, where the number of the linear regression sub-areas is equal to the number of rows of the edge beads in the seam area. Specifically, divide the connecting line segment of the two left vertices a1 and a4 into four equal parts, that is, three new breakpoints l1, l2, and l3 are formed between a1 and a4; similarly, divide the connecting line segment of the two right vertices a2 and a3 into four equal parts, that is, three new breakpoints r1, r2, and r3 are formed between a2 and a3, and four linear regression sub-areas A1, A2, A3, and A4 are formed by connecting the left and right breakpoints in the corresponding order.

[0078] Then, use the linear regression sub-areas to segment the target area image from the seam area image. Each target area image contains the images of the edge beads in the same row respectively, so that the images of the edge beads in the same row can be extracted, which is convenient for subsequent linear fitting.

[0079] Furthermore, based on the characteristics of LED assembly, the center errors of the beads in the same column or the same row in the same module are extremely small and are on the same straight line. Therefore, the linear regression algorithm can be used to fit this straight line. Figure 6 It is an image of the seam area of an LED screen composed of small-pitch beads. It can be seen that the portable camera carried by the drone cannot obtain clear single-bead images. Since the center of the bead is often the brightest place, that is, the place with the highest gray value in the seam area image, as Figure 7 shown, after binarizing the seam area image, the bead adhesion can be eliminated. Similarly, in this Embodiment 1, the target area images are secondarily binarized according to the gray threshold to eliminate the influence of adhesion, and only the pixel highlights near the center of the bead are retained. Set the gray threshold of the second binarization process to the maximum gray value of the target area image minus five, that is, uniformly set the gray values of the pixel points in the target area image that are less than the gray threshold to zero, and uniformly set the gray values of the pixel points in the target area image that are greater than or equal to the gray threshold to a gray value greater than zero, forming a target area binarized image as Figure 8 shown. Then, use the least squares linear regression algorithm to fit the coordinate values of each pixel point with a gray value greater than zero in the target area binarized image to obtain the center straight line. Perform linear fitting on each target area binarized image to obtain four center straight lines, and mark the four center images in the seam area image, as Figure 9 shown, and the segmentation of the edge beads in each row can be realized.

[0080] Among them, the gray-scale threshold of the first binarization process is smaller than that of the second binarization process. The size of the gray-scale threshold determines the number of pixel points retained after the binarization process. The larger the gray-scale threshold, the more pixel points are retained, and the smaller the gray-scale threshold, the fewer pixel points are retained. The first binarization process is to obtain a linear regression sub-region. To make the linear regression sub-region cover the bead images in the same row as much as possible, a smaller gray-scale threshold needs to be set at this time to retain more pixel points to obtain a more accurate linear regression sub-region. The second binarization process is for linear fitting, and the center line obtained by fitting needs to pass through the center points of each bead. Therefore, by setting a larger gray-scale threshold and retaining fewer pixel points, the center line can be obtained.

[0081] The method for segmenting the image of the LED screen seam area provided in the first embodiment accurately finds the center points of the beads through statistical calculation, avoiding problems such as possible failure in contour finding when using the bead contour to find the center points, and at the same time solving the problem of bead adhesion at the seams of the LED modules, laying a foundation for subsequent bright-dark line correction.

[0082] As Figure 10 shown, the second embodiment discloses a method for correcting the bright-dark line of an LED screen. The method for correcting the bright-dark line of an LED screen includes the following steps:

[0083] Step S601: Obtain the image of the seam area of the LED screen to be corrected. The image of the seam area includes at least two rows of edge bead images of each LED module in the two spliced LED modules, and the arrangement direction of each row of edge beads is parallel to the edge of the LED module seam.

[0084] Step S602: Use the segmentation method of the first embodiment to perform segmentation processing on the image of the seam area to obtain the center line corresponding to each row of edge beads.

[0085] Step S603: Calculate the distance between the beads in each bead unit to be corrected according to the function expressions of the center lines. Each bead unit to be corrected includes the beads in each row of edge beads along the vertical direction of the LED module seam edge and in the same row;

[0086] Step S604: Determine the correction coefficient of the bead unit to be corrected according to the distance between the beads;

[0087] Step S605: Adjust the brightness of each bead unit to be corrected according to each correction coefficient.

[0088] Among them, the method for obtaining the image of the seam area in step S601 can refer to the description of Embodiment 1 and will not be elaborated here. In step S603, a rectangular coordinate system is established in the image of the seam area marked with the central straight lines. The length direction of the image of the seam area represents the x-axis (abscissa), and the width direction represents the y-axis (ordinate). The functional expressions of the four central straight lines are respectively represented as f1, f2, f3, and f4. By sampling the four straight lines at a fixed interval x value, the corresponding y values can be obtained, and the difference in the corresponding y values can represent the difference in the center distances of the lamp beads.

[0089] First, it is necessary to determine the abscissas of the left and right edge points of the image of the seam area. The left and right edge points respectively represent the first and last lamp beads in a row of edge lamp beads. Let the total number of single-row edge lamp beads be N, and let the abscissa of the left edge point be x min , and the abscissa of the right edge point be x max . Then the interval value Dis is:

[0090] Dis = (x max - x min ) / N

[0091] Sample the central straight lines. Let the abscissa of each sampling point be x i . Then

[0092] x i = x min + i·Dis, i = 1, 2, 3,... N

[0093] i represents the number of sampling points. The abscissa of the i-th sampling point represents the abscissa value near the center of the i-th lamp bead from left to right. Further, according to the functional expression of the central straight line, the ordinate of the i-th sampling point can be determined.

[0094] The distance between two adjacent rows of lamp beads can be equivalent to the difference in the ordinates of two sampling points. The abscissa of the i-th sampling point corresponds to four ordinate values on different straight lines. Let the differences in the ordinates of two adjacent sampling points be successively:

[0095] dis1 i = f2(x i ) - f1(x i ), i = 1, 2, 3,..., N

[0096] dis2 i = f3(x i ) - f2(x i ), i = 1, 2, 3,..., N

[0097] dis3 i = f4(x i ) - f3(x i ), i = 1, 2, 3,..., N

[0098] Among them, dis1 i represents the vertical coordinate difference between the second row of lamp beads and the first row of lamp beads, and dis2 i represents the vertical coordinate difference between the third row of lamp beads and the second row of lamp beads, and dis3 i represents the vertical coordinate difference between the fourth row of lamp beads and the third row of lamp beads. The three differences respectively represent the distances between adjacent two lamp beads.

[0099] As Figure 11 shown, each lamp bead unit to be corrected includes four edge lamp beads. Since the second row of lamp beads and the first row of lamp beads are in the same LED module, and the fourth row of lamp beads and the third row of lamp beads are in the same LED module, so the dis1 i value and the dis3 i value are very close. Therefore, the average value of dis1 i and dis3 i can be used as the correct distance truedis i , and truedis i represents the average value of the distances between adjacent two lamp beads in the same LED module in the i-th lamp bead unit to be corrected. Dis2 i represents the seam distance, that is, dis2 i is the distance dis i between two lamp beads on both sides of the splicing position in the i-th lamp bead unit to be corrected. The brightness of the lamp bead perceived by the naked eye is mainly related to the area size on the left and right sides of the lamp bead and the luminous intensity of the lamp bead. As Figure 11 shown, the middle dotted line represents the splicing position, and the outer dotted line represents the range of lamp beads participating in the calculation. The area size on the left and right of the seam lamp bead is related to both the seam distance and the correct distance. The distances on both sides of the non-seam lamp bead are both correct distances. Regarding the light intensity influence range of the lamp bead as a rectangle formed by the gaps on its left and right sides, the width of the rectangle is certain, and the length is related to the lamp bead distance. Let the naked eye sensitivity at the splicing position and the non-splicing position of the i-th sampling point be: P 1i , P 2i . Then the ratio of the two is:

[0100]

[0101] It is necessary to adjust the naked eye sensitivity at the splicing position to be the same as that at the non-splicing position. Therefore, the correction coefficient is the inverse of the above formula. Let the correction coefficient of the i-th sampling point (i.e., the i-th lamp bead) be k i , then:

[0102]

[0103] Since some LED modules cannot adjust the brightness of each light bead individually and can only adjust the overall brightness of a row of light beads simultaneously, in another embodiment, the overall correction coefficient K can also be calculated:

[0104]

[0105] In this way, on the one hand, the brightness of the light bead unit to be corrected can be adjusted according to the correction coefficient k of each light bead unit to be corrected i and the brightness of the light bead units to be corrected can also be adjusted according to the overall correction coefficient K, that is, all the light beads in the splicing area adopt the same correction coefficient K.

[0106] For example Figure 12 As shown, further, Embodiment 3 of the present invention discloses another method for correcting the bright and dark lines of an LED screen. The method for correcting the bright and dark lines of an LED screen includes the following steps:

[0107] Step S701: Obtain the image of the splicing area of the LED screen to be corrected. The image of the splicing area includes at least two rows of edge light bead images of each LED module in the two spliced LED modules, and the arrangement direction of each row of edge light beads is parallel to the splicing edge of the LED module.

[0108] Step S702: Perform segmentation processing on the image of the splicing area by using the segmentation method of Embodiment 1, and the central straight line corresponding to each row of edge light beads.

[0109] Step S703: Divide the image of the splicing area into a splicing area and a non-splicing area according to each central straight line.

[0110] Step S704: Calculate the gray mean value of the splicing area and the gray mean value of the non-splicing area respectively, and use the ratio of the gray mean value of the non-splicing area to the gray mean value of the splicing area as the correction coefficient.

[0111] Step S705: Adjust the brightness of each light bead in the splicing area according to the correction coefficient.

[0112] In addition to obtaining the correction coefficient according to the distance, the correction coefficient can also be obtained through the difference between the gray mean value of the non-splicing area and the gray value of the splicing area. Exemplarily, the splicing area is Figure 13 shown as R1 in Figure 13 and the non-splicing area is shown as R2 in min The functional expressions of the four central straight lines are respectively represented as f1, f2, f3, f4. The splicing area R1 is the midline of the two central straight lines f1 and f2, the midline of the two central straight lines f3 and f4, the left edge x = x maxThe enclosed area, and the average gray value gray1 of all pixel points within this area R1 is statistically calculated. The non - splicing area R2 is enclosed by the mid - lines of f1, f1, f2 and the left edge x = x min and the right edge x = x max The enclosed area, and the average gray value gray2 of all pixel points within this area R2 is statistically calculated. Therefore, the correction coefficient K is:

[0113]

[0114] According to the above - mentioned correction coefficient, the brightness of each lamp bead within the seam area can be adjusted. In another embodiment, the seam area can also be divided into several sub - regions of lamp beads to be corrected. Then, in the above - mentioned manner, each sub - region of lamp beads to be corrected is divided into a splicing sub - region and a non - splicing sub - region. The average gray value of all pixel points within the splicing sub - region and the average gray value of all pixel points within the non - splicing sub - region are calculated, and the ratio of the two average gray values is used as the correction coefficient of this sub - region of lamp beads to be corrected. In this way, the brightness of each lamp bead can be adjusted more precisely.

[0115] The correction method in the third embodiment of the present invention calculates the correction coefficient based on the gray - scale difference, more intuitively solves the problem of bright - dark lines. At the same time, compared with the method of calculating the correction coefficient based on the distance difference, its advantage is that it can be iteratively corrected. As the iteration progresses, the result will be better and better.

[0116] As Figure 14 shown, the fourth embodiment of the present invention also discloses a segmentation device for an LED screen seam - area image. The segmentation device includes a first binary - processing module 100, a linear - regression sub - region determination module 200, a target - region image segmentation module 300, a second binary - processing module 400, and a linear - fitting module 500. The first binary - processing module 100 is used to perform the first binary - processing on the seam - area image to obtain a binary image of the seam area; the linear - regression sub - region determination module 200 is used to determine several linear - regression sub - regions from the binary image of the seam area; the target - region image segmentation module 300 is used to segment the target - region image from the seam - area image according to each linear - regression sub - region, and each target - region image respectively contains the images of the edge lamp beads in the same row; the second binary - processing module 400 is used to perform the second binary - processing on each of the target - region images to obtain a binary image of the target region; the linear - fitting module 500 is used to perform linear fitting on the pixel points that meet the gray - value condition in each binary image of the target region to obtain a central straight line, and the central straight line is used as the connection line of the center points of each edge lamp bead in the same row.

[0117] Further, the linear regression sub-region determination module 200 is further configured to: perform a closing operation on the binary image of the seam region, and crop both sides of the image after the closing operation along the vertical direction of the edge of the LED module seam; perform polygon fitting on the cropped image to obtain four vertices, and sequentially connect the four vertices to form a rectangular region; and equally divide the rectangular region along the parallel direction of the edge of the LED module seam to obtain a plurality of linear regression sub-regions, where the number of linear regression sub-regions is equal to the number of rows of the edge beads in the seam region.

[0118] Further, the linear fitting module 500 is further configured to use the least squares linear regression algorithm to fit the coordinate values of each pixel point with a gray value greater than zero in the binary image of the target region to obtain a central straight line.

[0119] Among them, the more detailed working processes of the first binary processing module 100, the linear regression sub-region determination module 200, the target region image segmentation module 300, the second binary processing module 400, and the linear fitting module 500 can refer to the description of the relevant steps in Embodiment 1, and will not be elaborated here.

[0120] Embodiment 5 also discloses a computer-readable storage medium, which stores a segmentation program for an LED screen seam region image. When the segmentation program for the LED screen seam region image is executed by a processor, the above-mentioned method for segmenting an LED screen seam region image is implemented.

[0121] Embodiment 6 also discloses a computer device. At the hardware level, as Figure 15 shown, the computer device includes a processor 12, an internal bus 13, a network interface 14, and a computer-readable storage medium 11. The processor 12 reads the corresponding computer program from the computer-readable storage medium and then runs, forming a request processing device at the logical level. Of course, in addition to the software implementation method, one or more embodiments of this specification do not exclude other implementation methods, such as logical devices or a combination of software and hardware, etc. That is to say, the execution subject of the following processing flow is not limited to each logical unit, but can also be hardware or a logical device. The computer-readable storage medium 11 stores a segmentation program for an LED screen seam region image. When the segmentation program for the LED screen seam region image is executed by the processor, the above-mentioned method for segmenting an LED screen seam region image is implemented.

[0122] A computer-readable storage medium includes permanent and non-permanent, removable and non-removable media that can implement information storage by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer-readable storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette tapes, disk storage, quantum memory, graphene-based storage media or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device.

[0123] The specific embodiments of the present invention have been described in detail above. Although some embodiments have been shown and described, those skilled in the art should understand that these embodiments can be modified and perfected without departing from the principles and spirit of the present invention as defined by the claims and their equivalents, and these modifications and improvements should also be within the protection scope of the present invention.

Claims

1. A method for segmenting an image in the seam area of an LED screen, characterized in that, The splitting method includes: Performing a first binarization process on the image of the seam area to obtain a binarized image of the seam area; Determining a number of linear regression sub-regions from the binarized image of the seam area; Segmenting a target region image from the image of the seam area according to each of the linear regression sub-regions, and each of the target region images respectively includes the images of the edge light beads located in the same row; Performing a second binarization process on each of the target region images to obtain a binarized image of the target region; Performing linear fitting on the pixel points that meet the gray value condition in each of the binarized images of the target region to obtain a central straight line, and the central straight line is used as the connection line of the center points of each edge light bead in the same row; Among them, the method for determining a number of linear regression sub-regions from the binarized image of the seam area includes: Performing a closing operation on the binarized image of the seam area, and cropping both sides of the image after the closing operation along the vertical direction of the edge of the LED module seam; Performing polygon fitting on the cropped image to obtain four vertices, and sequentially connecting the four vertices to form a rectangular region; Evenly dividing the rectangular region along the parallel direction of the edge of the LED module seam to obtain a number of linear regression sub-regions, where the number of the linear regression sub-regions is equal to the number of rows of the edge light beads in the seam area.

2. The method for segmenting an image of an LED screen seam region according to claim 1, wherein, The method for performing linear fitting on the pixel points that meet the gray value condition in each of the binarized images of the target region to obtain a central straight line is: Using the least squares linear regression algorithm to fit the coordinate values of each pixel point with a gray value greater than zero in the binarized image of the target region to obtain a central straight line.

3. The method for segmenting an image of an LED screen seam area according to claim 1, characterized in that, The gray value threshold of the first binarization process is less than the gray value threshold of the second binarization process.

4. An image segmentation device for the seam area of an LED screen, characterized in that, The splitting device includes: A first binarization processing module, configured to perform a first binarization process on the image of the seam area to obtain a binarized image of the seam area; A linear regression sub-region determination module, configured to determine a number of linear regression sub-regions from the binarized image of the seam area; A target region image segmentation module, configured to segment a target region image from the image of the seam area according to each of the linear regression sub-regions, and each of the target region images respectively includes the images of the edge light beads located in the same row; A second binarization processing module, configured to perform a second binarization process on each of the target region images to obtain a binarized image of the target region; A linear fitting module, configured to perform linear fitting on the pixel points that meet the gray value condition in each of the binarized images of the target region to obtain a central straight line, and the central straight line is used as the connection line of the center points of each edge light bead in the same row; Among them, the linear regression sub-region determination module is further configured to: perform a closing operation on the binary image of the seam region, and crop the two side edges of the image after the closing operation along the vertical direction of the LED module seam edge; perform polygon fitting on the cropped image to obtain four vertices, and sequentially connect the four vertices to form a rectangular region; equally divide the rectangular region along the parallel direction of the LED module seam edge to obtain a plurality of linear regression sub-regions, where the number of the linear regression sub-regions is equal to the number of rows of the edge beads in the seam region.

5. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a segmentation program for the LED screen seam region image, and when the segmentation program for the LED screen seam region image is executed by a processor, it implements the LED screen seam region image segmentation method according to any one of claims 1 to 3.

6. A computer device, characterized in that, The computer device includes a computer-readable storage medium, a processor, and a segmentation program for the LED screen seam region image stored in the computer-readable storage medium. When the segmentation program for the LED screen seam region image is executed by the processor, it implements the LED screen seam region image segmentation method according to any one of claims 1 to 3.

7. A method for correcting bright and dark lines of an LED screen, characterized in that, The LED screen bright-dark line correction method includes: Obtain a seam region image of the LED screen to be corrected, where the seam region image includes at least two rows of edge bead images of each of the two spliced LED modules, and the arrangement direction of each row of edge beads is parallel to the LED module seam edge; Perform segmentation processing on the seam region image by using the segmentation method according to any one of claims 1 to 3 to obtain the central straight line corresponding to each row of edge beads; Calculate the distance between the beads in each bead unit to be corrected according to the function expressions of the central straight lines, and each bead unit to be corrected includes the beads in each row of edge beads along the vertical direction of the LED module seam edge and in the same row; Determine the correction coefficient of the bead unit to be corrected according to the distance between the beads; Adjust the brightness of each bead unit to be corrected according to each correction coefficient.

8. The method for correcting bright and dark lines of an LED screen according to claim 7, characterized in that, The method for determining the correction coefficient of the to-be-corrected lamp bead unit according to the distance between each lamp bead is to calculate the correction coefficient according to the following formula : ， Among them, represents the distance between two light-emitting diodes on both sides of the splicing position in the th light-emitting diode unit to be corrected, and represents the average value of the distances between two adjacent light-emitting diodes in the same LED module in the th light-emitting diode unit to be corrected.

9. A method for correcting bright and dark lines of an LED screen, characterized in that, The LED screen bright-dark line correction method includes: Obtain a seam region image of the LED screen to be corrected, where the seam region image includes at least two rows of edge bead images of each of the two spliced LED modules, and the arrangement direction of each row of edge beads is parallel to the LED module seam edge; Perform segmentation processing on the seam region image by using the segmentation method according to any one of claims 1 to 3, and the central straight line corresponding to each row of edge beads; Divide the seam region image into a splicing region and a non-splicing region according to each central straight line; Calculate the gray scale mean value of the splicing region and the gray scale mean value of the non-splicing region respectively, and use the ratio of the gray scale mean value of the non-splicing region to the gray scale mean value of the splicing region as the correction coefficient; Adjust the brightness of each bead in the seam region according to the correction coefficient.

Images