Dynamic image partition identification and parameterization transmission method and device

Abstract

The embodiment of the invention provides a dynamic image partition recognition and parameterization transmission method and device, and the method comprises the steps: enabling an image to generate an edge image through Canny edge detection at an FPGA end, carrying out the dynamic binarization of the edge image, and outputting a binary image; carrying out bright region division and combination on the binary image through scanning labeling and a parent domain combination algorithm; the method comprises the following steps: integrating a plurality of bright area graphic parameters, extracting the integrated bright area graphic parameters, and transmitting the bright area graphic parameters through an improved GVSP protocol, in which a Parameter packet for transmitting the bright area graphic parameters is added.

Description

Technical Field

[0001] This document relates to the field of industrial vision inspection technology, and in particular to a method and apparatus for dynamic image partitioning recognition and parameterized transmission. Background Technology

[0002] In existing technologies, machine vision inspection plays an increasingly important role in strip and sheet production, enabling real-time measurement of strip and sheet width and detection of defects such as edge cracks and holes. However, existing inspection systems face the following problems and challenges: 1. The existing implementation uses an FPGA to control an image sensor to acquire images, which are then transmitted to a PC via a network cable for image processing by PC software. However, with the increasing speed of production lines and the growing requirements for detection accuracy, the amount of image data generated per second by the detection system is also increasing. The bandwidth of the network cable and the processing speed of the PC software have gradually become the bottlenecks of the system.

[0003] 2. Existing software algorithms determine the edge and defect location of the conveyor belt by analyzing edge points. However, under the condition of high-speed conveyor belt movement and blurring caused by vibration, the edge points generated by the Canny algorithm are prone to discontinuity or even complete disappearance.

[0004] 3. If the recognition function could be implemented on the FPGA, it would save a significant amount of transmission bandwidth and PC computing performance. However, the parallel structure of the FPGA cannot directly use PC algorithms, and the existing GigE Vision protocol only supports image transmission. Summary of the Invention

[0005] The purpose of this invention is to provide a method and apparatus for dynamic image partition recognition and parameterized transmission, which aims to solve the above-mentioned problems in the prior art.

[0006] This invention provides a method for dynamic image partitioning recognition and parameterized transmission, comprising: On the FPGA side, Canny edge detection is used to generate an edge map from the image, and the edge map is dynamically binarized and output as a binary image. The binary image is divided and merged into bright areas using a scan annotation and parent domain merging algorithm. The integrated bright area graphic parameters are extracted and transmitted through the improved GVSP protocol, wherein the improved GVSP protocol adds a Parameter packet for transmitting the bright area graphic parameters.

[0007] This invention provides a dynamic image partitioning recognition and parameterized transmission device, comprising: On the FPGA side, Canny edge detection is used to generate an edge map from the image, and the edge map is dynamically binarized and output as a binary image. The binary image is divided and merged into bright areas using a scan annotation and parent domain merging algorithm. The integrated bright area graphic parameters are extracted and transmitted through the improved GVSP protocol, wherein the improved GVSP protocol adds a Parameter packet for transmitting the bright area graphic parameters.

[0008] This invention also provides an electronic device, including: a memory, an FPGA, and a computer program stored in the memory and executable on the FPGA. When the computer program is executed by the FPGA, it implements the steps of the above-described dynamic image partitioning recognition and parameterized transmission method.

[0009] This invention also provides a computer-readable storage medium storing an information transmission implementation program, which, when executed by an FPGA, implements the steps of the above-described dynamic image partition recognition and parameterized transmission method.

[0010] By employing embodiments of the present invention, through dynamic binarization, region division and merging, and parameterized protocol extension, accurate identification and efficient transmission of edge holes on the board strip are achieved, significantly reducing data processing complexity and network bandwidth requirements. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in one or more embodiments of this specification or in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a flowchart of the dynamic image partition recognition and parameterized transmission method according to an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the detailed processing of the dynamic image partitioning recognition and parameterized transmission method according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the original image before pixel division in an embodiment of the present invention; Figure 4 This is a schematic diagram of the Canny edge map according to an embodiment of the present invention; Figure 5 This is a schematic diagram illustrating the criteria for classifying bright spots, dark spots, and gray spots in an embodiment of the present invention. Figure 6This is a schematic diagram of pixel division according to an embodiment of the present invention; Figure 7 This is a schematic diagram of the optimized erosion algorithm according to an embodiment of the present invention; Figure 8 This is a schematic diagram of the optimized dilation algorithm according to an embodiment of the present invention; Figure 9 This is a schematic diagram showing the state of edge points, bright spots, dark spots, and gray spots after corrosion and expansion in an embodiment of the present invention. Figure 10 This is a schematic diagram after binarization according to an embodiment of the present invention; Figure 11 This is a schematic diagram of the bright area marking results according to an embodiment of the present invention; Figure 12 This is a schematic diagram of example 1 of the bright area marking in an embodiment of the present invention; Figure 13 This is a schematic diagram of example 2 of the bright area marking in an embodiment of the present invention; Figure 14 This is a schematic diagram of example 3 of the bright area marking in an embodiment of the present invention; Figure 15 This is a schematic diagram of the improved GVSP protocol according to an embodiment of the present invention; Figure 16 This is a schematic diagram of the dynamic image partition recognition and parameterized transmission device according to an embodiment of the present invention; Figure 17 This is a schematic diagram of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0013] To enable those skilled in the art to better understand the technical solutions in one or more embodiments of this specification, the technical solutions in one or more embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this specification, and not all of the embodiments. Based on one or more embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of this document.

[0014] Method Implementation Examples According to embodiments of the present invention, a dynamic image partitioning recognition and parameterized transmission method is provided. Through dynamic binarization, region partitioning and merging, and parameterized protocol extension, the method achieves accurate recognition and efficient transmission of board edge holes, significantly reducing data processing complexity and network bandwidth requirements. Figure 1 This is a flowchart of the dynamic image partitioning recognition and parameterized transmission method according to an embodiment of the present invention, such as... Figure 1As shown, the dynamic image partitioning recognition and parameterized transmission method according to an embodiment of the present invention specifically includes: Step S101: On the FPGA side, an edge map is generated from the image using Canny edge detection, and the edge map is dynamically binarized and output as a binary image; specifically including: On the FPGA side, Canny edge detection is used to generate an edge map from the image, and the edge map is dynamically binarized: the pixels of the edge map are divided into bright spots, dark spots, gray spots, and edge points. Erosion and dilation are performed on the bright spots, dark spots, and gray spots to optimize the continuity of bright areas, thereby achieving dynamic binarization of the image and outputting a binary image. Specifically, dividing the pixels of the edge map into bright spots, dark spots, gray spots, and edge points, and performing erosion and dilation on the bright spots, dark spots, and gray spots to optimize the continuity of bright areas includes: Without changing the edge points in the pixels of the edge map, points below the Canny threshold are classified as dark points, points above 255-Canny threshold are classified as bright points, and the rest are classified as gray points. Perform the first predetermined number of erosion operations: if there is a gray point or edge point among the four points adjacent to a bright spot or dark spot, then turn the bright spot or dark spot into a gray point; Perform a second predetermined number of dilation operations: If a gray point has a bright or dark point among its four adjacent points, then the gray point is transformed into a bright or dark point through the dilation operation; if a gray point has both bright and dark points among its four adjacent points, then the gray point is transformed into a larger number of bright or dark points. Through erosion and dilation operations, gray spots are eliminated on one side of the edge, thus optimizing the continuity of the bright area.

[0015] Step S102 involves dividing and merging bright areas in the binary image using a scan annotation and parent region merging algorithm; specifically including: Bright areas are marked, and the same bright areas are assigned the same pixel value to divide different bright areas: During FPGA scanning, if there are no adjacent valid pixels and the current pixel is bright, it is marked as a new bright area and a new pixel value is assigned in sequence. If it is adjacent to multiple valid pixels, the multiple adjacent valid pixels are marked as a bright area and the pixel value of the bright area is assigned to the smallest pixel value among the valid pixels. Bright areas belonging to the same region are merged into a single region using a parent domain merging algorithm.

[0016] Step S103: Extract the integrated bright area graphic parameters and transmit them via the improved GVSP protocol. The improved GVSP protocol includes a Parameter packet for transmitting the bright area graphic parameters. The improved GVSP protocol format is: Parameter packet → Leader packet → Payload packet → Trailer packet. Specifically, extracting the integrated bright area graphic parameters includes: When the FPGA scans and marks bright areas, the area and boundary coordinates of the corresponding bright areas are calculated in real time. Based on the relationship between adjacent bright areas, the parameters of adjacent bright areas are integrated to obtain the graphic parameters of the bright areas. The graphic parameters of the bright areas specifically include: camera number, bright area area, and bright area boundary coordinates.

[0017] The technical solution of this invention achieves efficient identification and parameterized transmission of holes at the edges of metal strips through dynamic binarization, region merging algorithms, and GVSP protocol extensions. Core innovations include: multi-threshold classification and morphological optimization, region merging mechanism, and parameter data packet design. This system improves detection accuracy while reducing bandwidth consumption and is suitable for industrial inspection scenarios such as metal strips and precision manufacturing.

[0018] The technical solutions of the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0019] This invention proposes an algorithm for edge and hole recognition at the FPGA end. The input is the edge map output by the Canny algorithm; the output is the edge of the strip and the location information of the holes. Specifically, the processing steps are as follows: Based on the edge map generated by the Canny algorithm, the entire image is binarized. After binarization, bright spots are used as the analysis targets. Bright areas are divided according to their adjacent continuity and numbered, with the number used as the pixel value. The image is traversed to determine which bright areas are actually adjacent and merged. The image is traversed again, and based on the number of each bright spot, the extreme values ​​(e.g., extreme left value, extreme top value, etc.) of each bright area are calculated. Based on the bright area numbers, adjacency relationships, and extreme values ​​from the previous two steps, the actual number of bright areas and their location and morphological parameters are calculated. The entire algorithm's processing flow is described in [link to algorithm description]. Figure 2 As shown. The parameters are sent to the PC via network protocol.

[0020] 2. Dynamic Binarization and Morphological Optimization: Traditional binarization uses a single threshold to classify each pixel as a bright or dark spot. This leads to a problem where image regions with pixel values ​​close to the threshold exhibit jagged edges. Furthermore, it fails to utilize the Canny algorithm to find image edges. Therefore, this invention optimizes the binarization algorithm, specifically through the following steps: Pixel value division: based on Figure 3For example, after performing edge extraction using the Canny algorithm, the following results are obtained: Figure 4 The traditional Canny algorithm outputs an image like this, where all pixels are divided into edge points (...). Figure 4 Highlights in the middle) and other points ( Figure 4 (The black dots in the image). The technical solution of this invention classifies other points into three types: bright spots, dark spots, and gray spots. The classification is based on the Canny threshold: points below the Canny threshold are dark spots, points above 255 - the Canny threshold are bright spots, and the rest are gray spots. For example... Figure 5 As shown. The segmented image is as follows. Figure 6 As shown.

[0021] Erosion-Dilation Optimization: To achieve binarization, gray points on at least one side of the edge must be eliminated, requiring morphological operations of erosion and dilation. Traditional erosion and dilation operations only allow two pixel values: bright and dark. However, the image in this embodiment has four values: bright, dark, gray, and edge. Therefore, this embodiment improves the erosion and dilation operations: First, neither erosion nor dilation affects edge points. Erosion operation: Defined as follows: if a bright or dark point has a gray or edge point among its four adjacent points, then that bright or dark point is converted to a gray point. Its effect is to enlarge the gray point area and shrink the bright and dark point areas. For example... Figure 7 As shown, the definition of the dilation operation is that if a gray point has a bright or dark spot among its four adjacent points, then the gray point is transformed into a bright or dark spot. If both bright and dark spots exist among adjacent points, the one with more bright spots is transformed. The effect is to reduce the size of the gray point area and increase the size of the bright and dark spot areas. For example... Figure 8 As shown, due to occasional instances of scattered dark spots appearing on the other side of the edge during testing, this is likely an algorithmic issue with Canny when dealing with slanted edges. To eliminate these erroneous dark spots, erosion is performed twice, followed by dilation. Erosion and dilation consume resources; the number of erosions and dilations should be determined by considering both image quality and FPGA resources. Ultimately, this ensures that one side of the edge is free of gray spots. Figure 9 As shown, the right side of the edge is entirely composed of dark spots, with no gray spots; at this point, the gray and bright spots can be merged to achieve binarization. For example... Figure 10 As shown.

[0022] Bright area annotation: For the edge map given by the Canny algorithm (such as...) Figure 4 Existing algorithms analyze images based on edge points; however, when the image is blurry, edge points are often discontinuous, which affects the analysis results. Therefore, this invention no longer uses edge points as the analysis target; instead, it performs binarization processing on the image with the assistance of edge points, and then analyzes the bright areas within it. Figure 10As shown, there are two bright areas: the large area on the left is the blank area outside the strip, and the small circle on the right is a hole in the strip. Therefore, after obtaining the binarized image, the first step is to determine which bright spots belong to the same bright area based on their adjacency. In the technical solution of this embodiment, the bright areas are numbered, and this number is written into the pixel value of each bright spot in the bright area. Figure 10 As shown, the large bright area on the left is numbered 1, with each pixel value being 1; the bright area around the circular hole on the right is numbered 2, with each pixel value being 2; the edge points and dark points of the remaining non-bright areas are numbered 255. The resulting image is shown below. Figure 11 As shown.

[0023] contrast Figure 10 and Figure 11 As you can see, the bright and dark areas appear to be "flipped." This is because the bright areas are numbered sequentially starting from 1, which is displayed as black on the pixel. Specifically, when the FPGA scans the image, within the 3x3 grid window of the current pixel, if the current point is a bright spot and adjacent points are also bright spots, then the numbering of the adjacent bright spots is inherited. For example... Figure 11 As shown, the current bright spot inherits the number 1 from the bright spot to its left. If there are no adjacent bright spots, the current bright spot generates a new number. Figure 12 As shown, the current bright spot has no adjacent bright spots, so a new number 2 was generated. (Comparison) Figure 12 and Figure 13 It can be seen that, Figure 12 The bottom right half of the 3x3 grid is invalid data, therefore... Figure 13 The lower right half of the grid is omitted, leaving only half a grid. This is because when the image is transmitted in the FPGA, it is scanned from left to right and from top to bottom. Therefore, the lower right half of the grid contains pixels that have not yet been scanned and naturally cannot be numbered. It can be ignored in the bright area marking operation to save resources. If the current bright spot has multiple adjacent bright spots, the one with the smallest number is selected, such as... Figure 14 As shown. Figure 14 This situation involves adjacent bright areas where pixels have two different pixel numbers. This is usually caused by the specific shape of the bright areas. Therefore, a further processing step is needed to determine the adjacency relationship between the bright areas.

[0024] Determine the adjacency relationships of bright areas: Rescan the annotated image and record the adjacent bright areas for subsequent parameter merging. During scanning and annotation, the area and boundary coordinates (11 parameters including top / bottom y-values ​​and left / right edge x-values) of the corresponding bright areas can be calculated in real time. Based on the adjacency relationships of bright areas, parameters of adjacent bright areas are integrated (such as area accumulation and boundary extremum selection). Based on these parameters, the shape of the bright areas can be determined, such as whether they represent edges or holes, and their location.

[0025] Protocol Extension: Existing GVSP protocols can only transmit images. To transmit extracted parameters, the GVSP protocol needs modification. For example... Figure 15 As shown, a new Parameter packet has been added: In addition to the existing Leader, Payload, and Trailer packets, a new Parameter packet type has been added. Each packet transmits parameters for a bright area, including fields such as camera number, area, and boundary coordinates. Each image frame contains several Parameter packets (the number depends on the number of bright areas in the image), as well as the original Leader, Payload, and Trailer packets. Due to these modifications, it is no longer based on the universal GVSP protocol, so the PC software must be adapted.

[0026] The beneficial effects of the embodiments of the present invention are as follows: 1. This invention enables image analysis and recognition on an FPGA, reducing the burden on network transmission and PC software algorithms.

[0027] 2. Due to the parallel characteristics of FPGA, the embodiments of the present invention are performed in real time, and image processing is carried out simultaneously with image acquisition, without the need for additional image caching or processing time.

[0028] 3. The embodiments of the present invention have optimized the case where the edge points output by Canny are discontinuous. Instead of targeting edge points, they target bright areas, which can ensure that the edges and holes of the board strip can still be identified even when the image is out of focus or jittery.

[0029] 4. Compatible with existing GVSP protocol frameworks, facilitating debugging.

[0030] Device Example 1 According to embodiments of the present invention, a dynamic image partitioning recognition and parameterized transmission device is provided. Figure 16 This is a schematic diagram of the dynamic image partition recognition and parameterized transmission device according to an embodiment of the present invention, as shown below. Figure 16 As shown, the dynamic image partitioning recognition and parameterized transmission device according to an embodiment of the present invention specifically includes: The binarization module 160 is used to generate an edge map from an image using Canny edge detection on the FPGA side, perform dynamic binarization on the edge map, and output a binary image. Specifically, it is used to: generate an edge map from an image using Canny edge detection on the FPGA side, and perform dynamic binarization on the edge map: divide the pixels of the edge map into bright spots, dark spots, gray spots, and edge points; perform erosion and dilation on bright spots, dark spots, and gray spots to optimize the continuity of bright areas, realize dynamic binarization of the image, and output a binary image; specifically, keep the edge points in the edge map unchanged, and according to a preset Canny threshold, divide the pixels below the Canny threshold... Points above the 255-Canny threshold are classified as dark points, points above the 255-Canny threshold are classified as bright points, and the rest are classified as gray points. A first predetermined number of erosion operations are performed: if there are gray points or edge points among the four adjacent points of a bright or dark point, then the bright or dark point is turned into a gray point. A second predetermined number of dilation operations are performed: if there are bright or dark points among the four adjacent points of a gray point, then the gray point is turned into a bright or dark point through dilation. If there are both bright and dark points among the four adjacent points of a gray point, then the gray point is turned into a larger number of bright or dark points. Through erosion and dilation operations, the edge side is finally free of gray points, thus optimizing the continuity of the bright area. The segmentation module 162 is used to segment and merge bright areas in the Canny algorithm edge map through scanning annotation and parent domain merging algorithm; it is used to: annotate bright areas, assign the same pixel value to the same bright areas, and segment different bright areas: during FPGA scanning, if there are no adjacent valid pixels and the current pixel is bright, it is marked as a new bright area and assigned a new pixel value in sequence; if it is adjacent to multiple valid pixels, the multiple adjacent valid pixels are marked as a bright area and the pixel value of the bright area is assigned to the smallest pixel value among the valid pixels; and merge bright areas belonging to the same region into the same region through the parent domain merging algorithm. The transmission module 164 is used to extract and integrate the bright area graphic parameters, and transmit the bright area graphic parameters through an improved GVSP protocol. Specifically, the improved GVSP protocol adds a Parameter packet for transmitting the bright area graphic parameters. Specifically, it is used to: when the FPGA is scanning and marking bright areas, to calculate the area and boundary coordinates of the corresponding bright areas in real time, and to integrate the parameters of adjacent bright areas according to their proximity to obtain the bright area graphic parameters. The bright area graphic parameters specifically include: camera number, bright area area, and bright area boundary coordinates. The improved GVSP protocol format is: Parameter packet → Leader packet → Payload packet → Trailer packet.

[0031] The embodiments of the present invention are device embodiments corresponding to the above method embodiments. The specific operation of each module can be understood with reference to the description of the method embodiments, and will not be repeated here.

[0032] Device Example 2 This invention provides an electronic device, such as... Figure 17 As shown, it includes: a memory 170, an FPGA 172, and a computer program stored in the memory 170 and executable on the FPGA 172, wherein the computer program, when executed by the FPGA 172, implements the steps as described in the method embodiment.

[0033] Device Example 3 This invention provides a computer-readable storage medium storing an information transmission implementation program, which, when executed by an FPGA 172, implements the steps described in the method embodiment.

[0034] The computer-readable storage media described in this embodiment include, but are not limited to, ROM, RAM, disk, or optical disk.

[0035] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for dynamic image partitioning recognition and parameterized transmission, characterized in that, include: On the FPGA side, Canny edge detection is used to generate an edge map from the image, and the edge map is dynamically binarized and output as a binary image. The binary image is divided and merged into bright areas using a scan annotation and parent domain merging algorithm. The integrated bright area graphic parameters are extracted and transmitted through the improved GVSP protocol, wherein the improved GVSP protocol adds a Parameter packet for transmitting the bright area graphic parameters.

2. The method according to claim 1, characterized in that, On the FPGA side, Canny edge detection is used to generate an edge map from the image, and the edge map is dynamically binarized and output as a binary image. Specifically, this includes: On the FPGA side, Canny edge detection is used to generate an edge map from the image, and the edge map is dynamically binarized: the pixels of the edge map are divided into bright spots, dark spots, gray spots and edge points, and erosion and dilation are performed on bright spots, dark spots and gray spots to optimize the continuity of bright areas, realize the dynamic binarization of the image, and output a binary map.

3. The method according to claim 2, characterized in that, The pixels in the edge map are divided into bright spots, dark spots, gray spots, and edge spots. Erosion and dilation are performed on bright spots, dark spots, and gray spots to optimize the continuity of bright areas. Specifically, this includes: Without changing the edge points in the pixels of the edge map, points below the Canny threshold are classified as dark points, points above 255-Canny threshold are classified as bright points, and the rest are classified as gray points. Perform the first predetermined number of erosion operations: if there is a gray point or edge point among the four points adjacent to a bright spot or dark spot, then turn the bright spot or dark spot into a gray point; Perform a second predetermined number of dilation operations: If a gray point has a bright or dark point among its four adjacent points, then the gray point is transformed into a bright or dark point through the dilation operation; if a gray point has both bright and dark points among its four adjacent points, then the gray point is transformed into a larger number of bright or dark points. Through erosion and dilation operations, gray spots are eliminated on one side of the edge, thus optimizing the continuity of the bright area.

4. The method according to claim 1, characterized in that, The process of dividing and merging bright areas in the binary image using a scan annotation and parent domain merging algorithm specifically includes: Bright areas are marked, and the same bright areas are assigned the same pixel value to divide different bright areas: During FPGA scanning, if there are no adjacent valid pixels and the current pixel is bright, it is marked as a new bright area and a new pixel value is assigned in sequence. If it is adjacent to multiple valid pixels, the multiple adjacent valid pixels are marked as a bright area and the pixel value of the bright area is assigned to the smallest pixel value among the valid pixels. Bright areas belonging to the same region are merged into a single region using a parent domain merging algorithm.

5. The method according to claim 1, characterized in that, The extracted and integrated bright area graphic parameters specifically include: When the FPGA scans and marks bright areas, the area and boundary coordinates of the corresponding bright areas are calculated in real time. Based on the relationship between adjacent bright areas, the parameters of adjacent bright areas are integrated to obtain the graphic parameters of the bright areas. The graphic parameters of the bright areas specifically include: camera number, bright area area, and bright area boundary coordinates.

6. The method according to claim 1, characterized in that, The improved GVSP protocol format is: Parameter packet → Leader packet → Payload packet → Trailer packet.

7. A dynamic image partitioning recognition and parameterized transmission device, characterized in that, include: The binarization module is used to generate an edge map from an image using Canny edge detection on the FPGA side, dynamically binarize the edge map, and output the binary map. The partitioning module is used to partition and merge bright areas in the binary image using a scan annotation and parent domain merging algorithm. The transmission module is used to extract the integrated bright area graphic parameters and transmit the bright area graphic parameters through the improved GVSP protocol, wherein the improved GVSP protocol adds a Parameter packet for transmitting the bright area graphic parameters.

8. The apparatus according to claim 7, characterized in that, The binarization module is specifically used for: generating an edge map from an image using Canny edge detection on the FPGA side, and dynamically binarizing the edge map: dividing the pixels of the edge map into bright spots, dark spots, gray spots, and edge points; performing erosion and dilation on bright spots, dark spots, and gray spots to optimize the continuity of bright areas, thereby achieving dynamic binarization of the image and outputting a binary image; specifically: keeping the edge points in the pixels of the edge map unchanged, and according to a pre-set Canny threshold, dividing points below the Canny threshold into dark spots, dividing points above 255-Canny threshold into bright spots, and the rest into gray spots; performing a first predetermined number of erosion operations: if there is a gray spot or edge point among the four adjacent points of a bright spot or dark spot, then the bright spot or dark spot is changed to a gray spot; Perform a second predetermined number of dilation operations: if there are bright or dark points among the four adjacent points of a gray point, then the gray point is transformed into a bright or dark point through the dilation operation; if there are both bright and dark points among the four adjacent points of a gray point, then the gray point is transformed into a larger number of bright or dark points. Through the erosion and dilation operations, the gray point is eventually made to no longer have gray points on one side of the edge, thus optimizing the continuity of the bright area. The segmentation module is specifically used for: marking bright areas, assigning the same pixel value to the same bright areas, and dividing different bright areas: during FPGA scanning, if there are no adjacent valid pixels and the current pixel is bright, it is marked as a new bright area and assigned a new pixel value in sequence; if it is adjacent to multiple valid pixels, the multiple adjacent valid pixels are marked as a bright area, and the pixel value of the bright area is assigned to the smallest pixel value among the valid pixels; and merging bright areas belonging to the same region into the same region through a parent domain merging algorithm. The transmission module is specifically used to: when the FPGA is scanning and marking bright areas, to calculate the area and boundary coordinates of the corresponding bright areas in real time, and to integrate the parameters of adjacent bright areas according to the relationship between the bright areas to obtain the bright area graphic parameters. The bright area graphic parameters specifically include: camera number, bright area area and bright area boundary coordinates. The improved GVSP protocol format is: Parameter packet → Leader packet → Payload packet → Trailer packet.

9. An electronic device, characterized in that, include: The memory, the FPGA, and a computer program stored on the memory and executable on the FPGA, wherein the computer program, when executed by the FPGA, implements the steps of the dynamic image partition recognition and parameterized transmission method as described in any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores an implementation program for information transmission, which, when executed by the FPGA, implements the steps of the dynamic image partition recognition and parameterized transmission method as described in any one of claims 1 to 6.

Images