Shield tool image recognition system based on FPGA morphological operator

Abstract

The present application relates to the technical field of image analysis, and provides a shield tool image recognition system based on FPGA morphological operators. The shield tool image is acquired through an image acquisition module, and the shield tool image is segmented into shield image blocks to pre-process the shield tool image in parallel; the morphological processing module constructs a morphological operation array to process the shield tool image to extract the contour information, edge gradient information and topological structure information of the shield tool, and fuses the shield image blocks in combination with the topological structure information, which helps to deeply analyze the shape and structural characteristics of the shield tool and provides key data support for subsequent identification; the image recognition module determines the tool type, wear degree and wear type of the shield tool according to the contour information, edge gradient information and topological structure information of the shield tool, thereby improving the accuracy and reliability of shield tool status monitoring; the result output module displays the tool type, wear degree and wear type of the shield tool, and generates early warning information.

Description

Technical Field

[0001] The present application relates to the field of image analysis technology, and in particular to a shield tool image recognition system based on FPGA morphological operators. Background Art

[0002] During shield construction, shield cutters are critical components, and their condition directly impacts construction efficiency, quality, and safety. However, numerous challenges currently exist in the field of shield cutter image recognition and condition monitoring. Traditional shield cutter image recognition technology faces significant challenges in accuracy. Many existing methods struggle to accurately extract feature information about shield cutters, leading to misjudgments when determining cutter type, wear severity, and wear type (eccentric wear, chipping, and uniform wear). For example, these methods rely solely on single image features. Furthermore, conventional image acquisition and preprocessing methods fail to fully utilize hardware resources, resulting in slow processing speeds and an inability to implement parallel computing. The excessive processing time required to meet the real-time requirements of shield construction is insufficient. For example, during shield construction, the condition of shield cutters requires real-time monitoring. If image recognition processing is too slow, problems such as cutter wear cannot be detected promptly, impacting construction progress and safety. Summary of the Invention

[0003] In response to the deficiencies of the existing technology, this application provides a shield tool image recognition system based on FPGA morphological operators.

[0004] The present application provides a shield tool image recognition system based on FPGA morphological operators, which includes: an image acquisition module, a morphological processing module, an image recognition module and a result output module;

[0005] The image acquisition module is used to acquire the shield tool image and divide the shield tool image into shield image blocks according to the computing resources of the FPGA to pre-process the shield tool image in parallel;

[0006] The morphological processing module is used to adjust the structural elements according to the edge direction and texture characteristics of the shield image block, construct a morphological operation array, process the shield tool image through the structural elements, so as to extract the contour information and edge gradient information of the shield tool, determine whether the pixel point of the shield image block is an edge pixel based on the contour information and edge gradient information of the shield tool, obtain the unrefined edge pixels, detect the neighborhood connectivity of the edge pixels in the shield image block to refine and extract the topological structure information, and fuse the shield image block in combination with the topological structure information;

[0007] The image recognition module is used to determine the tool type and wear degree of the shield tool based on the profile information, edge gradient information and topological structure information of the shield tool;

[0008] The result output module is used to display the tool type and wear degree of the shield tool and generate early warning information.

[0009] As an optional implementation, the shield tool image processing strategy includes:

[0010] Adaptively adjust the shape and size of the structural element according to the edge direction and texture characteristics of the shield image block;

[0011] Constructing a morphological operation array to perform morphological operations through structural elements to obtain the contour information of the shield tool and extract the edge gradient information of the shield tool;

[0012] Refine and extract the topological structure information of the shield tool based on its contour information and edge gradient information;

[0013] The shield image blocks are fused according to the block index of the shield image blocks and combined with the topological structure information of the shield tool.

[0014] As an optional implementation, the logic for adaptively adjusting the shape and size of the structural element includes:

[0015] Determine the texture feature of each shield image block based on the local contrast of the shield image block, and determine the edge direction of each shield image block based on the gradient direction clustering of the shield image block;

[0016] Adjust the size of the structural element according to the texture characteristics of the shield image block;

[0017] The shape of the structural element is adjusted according to the edge direction of the shield image block.

[0018] As an optional implementation, the execution logic of the morphological operation includes:

[0019] For each shield image block, according to the shape and size of the structural element, the upper left corner pixel of each shield image block is used as the sliding starting point, and the shield image blocks are traversed one by one;

[0020] In the process of traversing the shield image block, the erosion operation array, the expansion operation array, the opening operation array and the closing operation array are sequentially executed to obtain the contour information of the shield tool;

[0021] The edge gradient information of the shield tool is extracted based on the contour information of the shield tool.

[0022] As an optional implementation, the refinement extraction method of the topology information includes:

[0023] Based on the contour information and edge gradient information of the shield tool, it is determined whether the pixel point of each shield image block is an edge pixel;

[0024] Taking the edge pixel as the center, count the number of edge pixels in the neighborhood of the edge pixel to determine whether the edge pixel is thinned, and continue iterating until the thinning termination condition is met;

[0025] Obtain unrefined edge pixels and detect the neighborhood connectivity of edge pixels in each shield image block to determine the topological structure information of the shield tool;

[0026] Organize and mark the topological structure information of shield tools.

[0027] As an optional implementation manner, the detection logic of the neighborhood connection status includes:

[0028] Traverse and count the unrefined edge pixels, and determine the information of bifurcation points and intersection points through threshold judgment;

[0029] The search is for edge pixels with a neighborhood connectivity of 2 and a stable gradient direction as the starting point of the straight line segment. The gradient direction is used as the initial direction of the straight line segment. The unrefined edge pixels in the next neighborhood along the gradient direction are sequentially checked, as well as the deviation angle between the edge pixel and the initial direction of the straight line segment, to determine the end point of the straight line segment and obtain the straight line segment information.

[0030] The edge pixel with a search neighborhood connection number of 2 and an unstable gradient direction is used as the starting point of the curve segment. The gradient direction is used as the initial direction of the curve segment. The unrefined edge pixels in the next neighborhood and the curvature between adjacent unrefined edge pixels are checked along the gradient direction to determine the end point of the curve segment and obtain the curve segment information.

[0031] As an optional implementation, the judgment logic of the tool type and wear degree of the shield tool includes:

[0032] Build a standard shield tool model library;

[0033] Perform similarity calculation and comparison on the profile information of the shield tool and the profile information in the standard shield tool model library to obtain a candidate list of shield tools;

[0034] Compare the edge gradient information of the shield tool with the edge gradient information in the standard shield tool model library to narrow down the candidate list of shield tools;

[0035] Deeply match the topological structure information of the shield tool with the topological structure information in the standard shield tool model library to determine the tool type of the shield tool;

[0036] The degree of shield tool wear is determined based on the profile change and edge gradient change.

[0037] As an optional implementation, the shield tool image is acquired through a camera, and the preprocessing logic of the shield tool image includes:

[0038] Segment the shield tool image based on the FPGA's computing resources to obtain shield image blocks;

[0039] Establish a block index for each shield image block;

[0040] The denoising and enhancement operations of each shield image block are performed in parallel in the FPGA.

[0041] As an optional implementation manner, the judgment sub-logic of the wear degree of the shield tool includes:

[0042] Extract the contour area from the contour information, and calculate the contour deformation variable of the shield tool's contour area and the contour area in the standard shield tool model library through Euclidean distance;

[0043] Extract the gradient intensity from the edge gradient information and calculate the gradient attenuation between the gradient intensity of the shield tool and the gradient intensity in the standard shield tool model library;

[0044] Configure the contour deformation threshold and the gradient attenuation threshold, compare the contour deformation amount with the contour deformation threshold to obtain the contour deformation degree, and compare the gradient attenuation amount with the gradient attenuation threshold to obtain the gradient attenuation degree;

[0045] The degree of shield tool wear is comprehensively judged based on the degree of profile deformation and gradient attenuation.

[0046] As an optional implementation manner, the judgment sub-logic of the wear type of the shield tool includes:

[0047] Mirror the shield cutter along its central axis and calculate the symmetry of the profile before and after flipping.

[0048] Extract the gradient direction from the edge gradient information, count the distribution of the gradient direction on the shield tool contour, and calculate the mean value of the gradient direction of different shield image blocks to determine the gradient direction change;

[0049] The wear type of shield tool is comprehensively judged based on the profile symmetry and gradient direction change. The wear types include eccentric wear, chipping and uniform wear.

[0050] Compared with the existing technology, the beneficial effects of the present application are: obtaining the shield tool image through the image acquisition module, and dividing the shield tool image into shield image blocks to pre-process the shield tool image in parallel, ensuring that a comprehensive and clear shield tool image is obtained, and providing a high-quality image basis for subsequent processing, greatly improving the accuracy of image information, and reducing recognition errors caused by image quality problems.

[0051] A morphological operation array is constructed through the morphological processing module to process the shield tool image in order to extract the contour information, edge gradient information and topological structure information of the shield tool, and fuse the shield image blocks with the topological structure information. For the complex shape of the shield tool image, the contour information of the shield tool can be efficiently extracted, and the edge gradient information and topological structure information can be accurately obtained, which is helpful for in-depth analysis of the shape and structural characteristics of the shield tool, and provides key data support for subsequent identification.

[0052] The image recognition module determines the tool type, wear degree and wear type of the shield tool based on the contour information, edge gradient information and topological structure information of the shield tool. Based on the various key information extracted by the morphological processing module, the tool type, wear degree and wear type of the shield tool can be accurately determined, thereby improving the accuracy and reliability of shield tool status monitoring.

[0053] The result output module displays the tool type, wear degree and wear type of the shield cutter, and generates early warning information. The tool type, wear degree and wear type of the identified shield cutter are displayed intuitively, making it convenient for staff to quickly understand the tool status and issue early warnings to replace the shield cutter. This helps to reasonably arrange construction scheduling, carry out equipment maintenance in advance, avoid construction accidents and delays caused by excessive tool wear, and improve the safety and efficiency of shield construction.

[0054] Combining FPGA technology with morphological operators and applying it to shield tool image recognition scenarios, the FPGA's reconfigurable logic resources and parallel processing capabilities are utilized to improve the performance and efficiency of the image recognition system, making it innovative and leading in the shield construction industry. At the same time, the image recognition system can meet the actual needs of shield tool status monitoring and identification during shield construction. By obtaining the tool type, wear degree and wear type information of the shield tool in real time and accurately, and generating early warnings in a timely manner, it ensures the health and safety of the equipment during shield machine operation, providing strong support for the safe and efficient conduct of shield construction, and has high practical application value and promotion significance. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be derived from these drawings without inventive work. Among them:

[0056] Figure 1 This is a system structure diagram of a shield tool image recognition system based on FPGA morphological operators provided in an embodiment of the present application;

[0057] Figure 2 A diagram of a shield tool image processing strategy for a shield tool image recognition system based on an FPGA morphological operator provided in an embodiment of the present application;

[0058] Figure 3 This is a morphological operation execution logic diagram of the shield tool image recognition system based on FPGA morphological operators provided in an embodiment of the present application;

[0059] Figure 4 A diagram of a refined extraction method for topological structure information of a shield tool image recognition system based on FPGA morphological operators provided in an embodiment of the present application. DETAILED DESCRIPTION

[0060] In order to make the objectives, technical solutions and advantages of the embodiments of the present application more obvious and easy to understand, the technical solutions in the embodiments of the present application are clearly and completely described below in conjunction with the drawings in the specification. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments.

[0061] Example 1

[0062] like Figure 1 As shown, it is a system structure diagram of the shield tool image recognition system based on FPGA morphological operator provided in an embodiment of the present application. The system includes an image acquisition module, a morphological processing module, an image recognition module and a result output module.

[0063] The image acquisition module is used to acquire and preprocess shield tool images.

[0064] Specifically, the shield tool image is obtained through the camera, and the preprocessing logic of the shield tool image includes:

[0065] Segment the shield tool image based on the FPGA's computing resources to obtain shield image blocks;

[0066] Establish a block index for each shield image block;

[0067] The denoising and enhancement operations of each shield image block are performed in parallel in the FPGA.

[0068] The shield tool image is acquired through multiple cameras, including industrial cameras and 3D cameras. The industrial camera is used to identify the tool type, wear degree and wear type of the shield tool. The 3D camera is used to obtain the size information of the shield tool. The shield tool is photographed from different angles. For example, cameras are set at a certain angle interval (such as 45 degrees) around the cutter head to ensure that all sides and key parts of the shield tool are covered. All cameras are ensured to shoot at the same time to ensure that the acquired shield tool image reflects the status of the shield tool at the same moment. For each acquired shield tool image, scale-invariant feature transformation or addition is used. A fast and robust feature algorithm is used to extract feature points from shield tool images. These feature points are scale-invariant and rotation-invariant, and can be accurately matched in shield tool images from different perspectives. A feature point-based matching algorithm is used to find matching feature point pairs between different shield tool images. By matching the feature points, the spatial position relationship between different shield tool images is determined. Based on the spatial position relationship obtained by feature point matching, a weighted average fusion algorithm is used to splice multiple shield tool images into a complete shield tool image. During the fusion process, the pixel values of the overlapping areas are reasonably weighted to ensure a natural and seamless transition between the spliced images.

[0069] The spliced complete shield tool image is evenly divided according to a fixed size to obtain shield image blocks of the same size. For example, the size of each shield image block is n×n pixels. This is to adapt to the processing characteristics of the FPGA operation unit to ensure that each shield image block can be independently and efficiently processed simultaneously in different operation units of the FPGA, thereby improving the overall processing efficiency. A unique block index identifier is constructed for each shield image block to effectively manage and transmit data for the shield image block in the subsequent parallel processing process. The block index information includes the position coordinates of the shield image block in the original shield tool image, ensuring that the processed shield image block can be accurately restored to the complete shield tool image. Through this shield image block division and index establishment method, the parallel processing infrastructure of the shield tool image is established, enabling the FPGA to efficiently process each shield image block independently, improving the overall processing speed, and at the same time ensuring that the processed shield image block can be accurately restored to the complete shield tool image, providing a good foundation for subsequent morphological processing.

[0070] The type and intensity of the noise in the shield image block are determined by calculating the variance of the pixel grayscale values in each shield image block. This is based on the characteristic differences in the pixel grayscale value distribution of different noise types. According to the type and intensity of the noise in the shield image block, a suitable denoising algorithm is selected and the denoising operation is performed. For example, for shield image blocks dominated by Gaussian noise, a parallel Gaussian filtering algorithm is used to perform filtering calculations simultaneously in each shield image block. For shield image blocks with more salt and pepper noise, a parallel median filtering algorithm is used for denoising. The denoising operations of different shield image blocks are performed simultaneously in different computing units of the FPGA, which can greatly improve the denoising efficiency. After denoising, the originally blurred shield tool outline of the shield tool image becomes clear, and the detailed information in the shield tool image is retained, which improves the quality of the shield tool image and provides more accurate data for subsequent morphological processing.

[0071] After performing the denoising operation, in order to simplify the complexity of shield image block processing, grayscale conversion is performed in parallel on each shield image block, converting the color shield image block into a grayscale shield image block. The grayscale conversion process of each shield image block is independent of each other, making full use of the parallel computing resources of the FPGA, improving the conversion speed, and saving time for the entire preprocessing process; for the grayscale shield image blocks, the contrast of each shield image block is enhanced in parallel through histogram equalization in the FPGA. Through this parallel processing method, the detailed information of the shield image block can be better preserved, making the edge and texture of the shield tool more prominent, while also improving the processing speed while ensuring the processing effect, meeting the real-time requirements of the image recognition system.

[0072] The morphological processing module is used to construct a morphological operation array to process shield tool images and extract the contour information, edge gradient information and topological structure information of the shield tool.

[0073] Specifically, if Figure 2 As shown in Figure 2, the processing strategies for shield tool images include:

[0074] Adaptively adjust the shape and size of the structural element according to the edge direction and texture characteristics of the shield image block;

[0075] Constructing a morphological operation array to perform morphological operations through structural elements to obtain the contour information of the shield tool and extract the edge gradient information of the shield tool;

[0076] Refine and extract the topological structure information of the shield tool based on its contour information and edge gradient information;

[0077] The shield image blocks are fused according to the block index of the shield image blocks and combined with the topological structure information of the shield tool.

[0078] Furthermore, the logic for adaptively adjusting the shape and size of the structural element includes:

[0079] Determine the texture feature of each shield image block based on the local contrast of the shield image block, and determine the edge direction of each shield image block based on the gradient direction clustering of the shield image block;

[0080] Adjust the size of the structural element according to the texture characteristics of the shield image block;

[0081] The shape of the structural element is adjusted according to the edge direction of the shield image block.

[0082] Based on the characteristic that standard deviation can reflect the degree of data discreteness, the texture features of each shield image block are obtained by calculating the local standard deviation of the shield image block. If the local standard deviation of the shield image block is greater than or equal to the preset standard deviation threshold, it means that the texture of the shield image block is rough and there are many details. At this time, the size of the structural element needs to be reduced to better capture the details; on the contrary, when the local standard deviation of the shield image block is less than the preset standard deviation threshold, it means that the texture of the shield image block is smooth. At this time, the size of the structural element needs to be increased to improve the processing efficiency without losing important information. This improves the accuracy and efficiency of the morphological operation in processing the texture features of the shield image block as a whole.

[0083] The gradient direction can reflect the direction of pixel intensity change in the shield image block, so the edge direction of each shield image block is obtained by calculating the gradient direction of the shield image block, and the gradient direction of the shield image block is clustered, where the main direction after clustering is the edge direction of the shield image block. According to the edge direction, a structural element of a suitable shape is selected so that the morphological operation can better fit the edge characteristics of the shield image block. When the edge direction of the shield image block is a single direction (horizontal, vertical or a certain slope direction), for example, the edge direction of the shield image block is vertical, a structural element with a vertical shape is selected, such as a vertical line. When using this When morphological operations are performed on vertical linear structuring elements, the edges can be better highlighted along the vertical direction, making the edges in the vertical direction clearer and more continuous. When the edge direction of the shield image block is multi-directional, such as twill, structuring elements in the shape of ellipses and diamonds are used to adapt to the multi-directional edge direction, ensuring that the edges in all directions can be effectively processed, avoiding the omission of edge information due to mismatch of the structural element shape. This provides a more reliable basis for the subsequent accurate extraction of the shield tool's contour information, edge gradient information and topological structure information, and improves the accuracy of the entire image recognition system in extracting shield tool image features.

[0084] Furthermore, the execution logic of morphological operations is as follows: Figure 3 As shown, specifically including:

[0085] For each shield image block, according to the shape and size of the structural element, the upper left corner pixel of each shield image block is used as the sliding starting point, and the shield image blocks are traversed one by one;

[0086] In the process of traversing the shield image block, the erosion operation array, the expansion operation array, the opening operation array and the closing operation array are sequentially executed to obtain the contour information of the shield tool;

[0087] The edge gradient information of the shield tool is extracted based on the contour information of the shield tool.

[0088] For each shield image block, the starting position of the structuring element in the shield image block is determined based on the structuring element generated by the morphological operator, with the upper left corner pixel of each shield image block as the starting point. In order to ensure that the structuring element can act on every pixel area of the shield image block, thereby comprehensively performing morphological operations on the shield tool image, the structuring element is traversed row by row and column by column on each shield image block according to its shape and size. When the structuring element covers a certain pixel area of the shield image block, the grayscale value of the pixel at the corresponding position in the structuring element is compared with the grayscale values of all pixels in the pixel area. For example, when the structuring element is a 3×3 rectangle, the grayscale value of the central pixel point in the rectangular area is compared with the grayscale values of the 9 pixels in the rectangular area, and then the pixels of the shield image block are updated. Through this traversal method, the morphological operation array can be applied to the shield image block, thereby realizing the overall processing of the shield tool image.

[0089] Different morphological operation arrays update the pixels of the shield image block inconsistently. For example, the erosion operation array assigns the minimum pixel grayscale value in the rectangular area to the pixel point of the shield image block corresponding to the central pixel point in the rectangular area, thereby completing an erosion operation. When the structural operation slides over all positions on the shield image block, the erosion operation of the shield image block is completed, and so on, the erosion operation of all shield image blocks is completed. Its essence is to search for the shield tool (tool head) in the shield tool image to preliminarily outline the main body contour of the shield tool; and the expansion operation array assigns the maximum pixel grayscale value in the rectangular area to the pixel point of the shield image block corresponding to the central pixel point in the rectangular area, thereby completing an expansion operation. When the structural operation slides over all positions on the shield image block, the expansion operation of the shield image block is completed, and so on, the expansion operation of all shield image blocks is completed. In the shield tool image, some holes will appear after the erosion operation, and the expansion operation can fill the erosion. The voids that appear after the shield tool are enhanced, and the contour information of the shield tool is enhanced. Its essence is to expand the pixel values around the shield tool (cutter head) toward the center, thereby filling the voids and strengthening the contour; the opening operation array is to first perform an erosion operation on the shield image block, remove the small noise points and protrusions in the shield image block, and preliminarily outline the main contour of the shield tool. On the basis of the shield image block after the erosion operation, an expansion operation is performed to restore the eroded part in the corrosion process, and at the same time remove isolated noise and burrs, making the contour of the shield tool smoother. The steps are the same as the above-mentioned erosion operation array and expansion operation array; the closing operation array is to first perform an expansion operation on the shield image block. According to the steps of the expansion operation array, the edge part of the shield image block is expanded to fill the small voids. The erosion operation is performed on the shield image block after the expansion operation to shrink the part that is over-expanded in the expansion process, connect the broken edges, and further improve the contour details of the shield tool, thereby obtaining continuous and complete shield tool contour information.

[0090] First, through the corrosion operation, the small noise points and protrusions in the shield image block are removed, and the main outline of the shield tool is preliminarily outlined. Then the dilation operation is performed to fill the voids caused by corrosion, further strengthen the contour information of the shield tool, and make the contour more continuous and complete. These two basic morphological operations lay the foundation for obtaining contour information; on the basis of completing the corrosion operation and dilation operation, the opening operation and closing operation are performed. The opening operation removes isolated noise and burrs in the shield image block and makes the contour smoother. The closing operation helps to connect the broken edges and further improve the contour details of the shield tool. Through the combination of this series of morphological operation arrays, more accurate shield tool contour information is gradually extracted. The synergistic effect of these morphological operation arrays makes the contour of the shield tool clearer, more accurate and complete in the image, providing a reliable foundation for subsequent image recognition.

[0091] After completing the execution of the above morphological operation array, the computing resources of FPGA are used to adopt the edge detection algorithm such as the Sobel operator to calculate the gradient intensity and gradient direction of each pixel point of the shield tool contour. For example, for a pixel point on the contour of the shield tool, the horizontal and vertical gradient components of the pixel point are calculated. Assume that the horizontal gradient component G of the pixel point is x =20, the vertical gradient component of the pixel is G y =30, then the gradient strength is , the calculated gradient intensity of the pixel is about 36.06, and the gradient direction is ,in represents a two-parameter inverse tangent function. By considering the sign signals of the two parameters in the brackets, the quadrant in which the angle is located can be determined more accurately, so that the gradient direction of the pixel point is approximately 56.31 degrees. By calculating the gradient intensity and gradient direction of each pixel point in the entire shield tool image, the edge gradient information of the shield tool is obtained. Since the shield image block has been preliminarily processed by morphological operations to remove noise and interference, the calculated edge gradient information is more accurate and prominent, which can better reflect the edge characteristics of the shield tool, and is helpful for subsequent judgment of the tool type and wear degree analysis of the shield tool, thereby improving the accuracy and reliability of the entire image recognition system.

[0092] Furthermore, if Figure 4 As shown in the figure, the refined extraction methods of topological structure information include:

[0093] Based on the contour information and edge gradient information of the shield tool, it is determined whether the pixel point of each shield image block is an edge pixel;

[0094] Taking the edge pixel as the center, count the number of edge pixels in the neighborhood of the edge pixel to determine whether the edge pixel is thinned, and continue iterating until the thinning termination condition is met;

[0095] Obtain unrefined edge pixels and detect the neighborhood connectivity of edge pixels in each shield image block to determine the topological structure information of the shield tool;

[0096] Organize and mark the topological structure information of shield tools.

[0097] The contour information of the shield tool directly reflects the outer boundary of the shield tool, while the edge gradient information of the shield tool reflects the areas where the pixel intensity changes significantly. These areas usually correspond to the contour edge of the shield tool. When the contour information and edge gradient information of the shield tool are extracted, each pixel point of the shield tool contour is an edge pixel. Starting from the upper left corner of the overall shield tool image, each pixel point in the shield image block is scanned one by one to determine whether the pixel point is an edge pixel. If it is not an edge pixel, the pixel point is skipped and the next pixel point is scanned. For the edge pixel, the situation of its neighboring pixels is further analyzed.

[0098] Taking the edge pixel as the center, consider the situation of the 8 neighboring pixels of the edge pixel to determine whether the edge pixel is refined, that is, deleted. This is based on the basic characteristics of the topological structure. Non-critical edge pixels often have fewer and specifically distributed neighboring edge pixels. For example, among the 8 neighboring pixels of the edge pixel, only two adjacent pixels are edge pixels, and these two adjacent pixels are not diagonally adjacent. At the same time, the edge pixel is not the endpoint pixel of the shield image block, then the edge pixel can be refined. The endpoint pixel can be determined by judging that the shield image block has only one edge pixel. In the judgment process, care should be taken to avoid mistakenly deleting intersections. It is necessary to check whether the edge pixel meets the characteristics of the intersection. These intersections are crucial to maintaining the integrity of the topological structure. If they meet the requirements, the edge pixel is retained.

[0099] According to the above-mentioned refinement process of edge pixels, all pixels in the shield tool image are scanned and processed once, and an iteration is completed to check whether the refinement termination condition is met. If not, the process starts again from the upper left corner of the entire shield tool image and repeats the process of scanning and refining each pixel in the shield image block one by one, and the next iteration is performed until the refinement termination condition is met, so that unimportant edge pixels can be gradually removed. The refinement termination condition is that the iteration is stopped when the number of changes in edge pixels in the shield tool image is less than a certain threshold (such as 5 pixels) after the maximum number of iterations (such as 100 times) or several consecutive iterations (such as 3 times). After multiple iterations, the remaining unrefined edge pixels are the key pixels that can represent the topological structure of the shield tool.

[0100] The topological structure information of the shield tool is determined by detecting the neighborhood connectivity of each shield image block based on the last remaining unrefined edge pixels. The topological structure information includes intersections, bifurcation points and topological structure features, and the topological structure features include straight line segments and curve segments. For example, the number of pixel neighborhood connections at intersections and bifurcation points is large and presents a specific distribution pattern, while straight line segments and curve segments can be judged by the direction and continuity of adjacent edge pixels. These topological structure information fully reflects the shape and structural characteristics of the shield tool, providing key data support for subsequent image recognition.

[0101] The extracted topological structure information is sorted, and the location and attributes of branch points, intersections and topological structure features are classified and stored. Each extracted topological structure feature is marked, such as assigning a unique number to each branch point and intersection, and corresponding encoding is also performed for different shape features, such as L for straight line segments and C for curved segments. The topological structure information is stored in an array to facilitate subsequent image recognition to compare and determine the tool type and wear degree of the shield tool. It can be quickly and accurately called and compared, thereby improving the accuracy and efficiency of determining the tool type and wear degree of the shield tool.

[0102] Furthermore, the detection logic of the neighborhood connectivity includes:

[0103] Traverse and count the unrefined edge pixels, and determine the information of bifurcation points and intersection points through threshold judgment;

[0104] The search is for edge pixels with a neighborhood connectivity of 2 and a stable gradient direction as the starting point of the straight line segment. The gradient direction is used as the initial direction of the straight line segment. The unrefined edge pixels in the next neighborhood along the gradient direction are sequentially checked, as well as the deviation angle between the edge pixel and the initial direction of the straight line segment, to determine the end point of the straight line segment and obtain the straight line segment information.

[0105] The edge pixel with a search neighborhood connection number of 2 and an unstable gradient direction is used as the starting point of the curve segment. The gradient direction is used as the initial direction of the curve segment. The unrefined edge pixels in the next neighborhood and the curvature between adjacent unrefined edge pixels are checked along the gradient direction to determine the end point of the curve segment and obtain the curve segment information.

[0106] For each unrefined edge pixel in each shield image block, traverse from the upper left corner of the shield image block. For each unrefined edge pixel traversed, check each pixel in its 8 neighborhoods to determine whether each pixel in the 8 neighborhoods is also an unrefined edge pixel and count the number of unrefined edge pixels for each pixel in the 8 neighborhoods. If the number of unrefined edge pixels for each pixel in the 8 neighborhoods is greater than or equal to 3, then the unrefined edge pixel is determined to be a bifurcation point or an intersection point. Continuously traverse to obtain all intersections and bifurcation points and store them in a bifurcation point list, where the bifurcation point list records the pixel coordinates of each intersection or bifurcation point. The angular relationship and positional relationship between these edge pixels can be used to distinguish whether the edge pixel is a bifurcation point or an intersection point. For example, a pattern where three or more edge pixels extend from the edge pixel in different directions is a bifurcation point, and a pattern where two edges intersect each other is an intersection point. This can effectively identify key nodes in the shield tool topology.

[0107] All unrefined edge pixels in the shield image block are traversed again. Since unrefined edge pixels will be located at one end of a straight line segment, edge pixels with a neighborhood connectivity of 2 and a stable gradient direction are searched as the starting points of the straight line segment. The stability of the gradient direction can be determined by calculating the angle between the line connecting the two unrefined edge pixels in the neighborhood and the center pixel point. If the angle is within a certain threshold range, such as less than 15 degrees, the gradient direction is considered stable and these edge pixels are used as the starting points of the straight line segment.

[0108] For the starting point of each straight line segment found, the gradient direction is used as the initial direction of the straight line segment. Along the initial direction of the straight line segment, the step size of one pixel is moved each time. Based on the linear characteristics of the straight line segment, the range of the straight line segment is determined by tracking the directional consistency of the edge pixels. The unrefined edge pixels in the next neighborhood are checked in turn, and the deviation angle between the newly checked unrefined edge pixels and the initial direction of the current straight line segment is calculated. When the deviation angle is greater than a certain threshold, such as 15 degrees, it is considered that a straight line segment is found, and the newly checked unrefined edge pixel is the end point of the straight line segment, so that the straight line segment information in the shield tool topology structure can be accurately obtained and stored in the straight line segment list.

[0109] All unrefined edge pixels in the shield image block are traversed again, and edge pixels with a neighborhood connectivity of 2 and unstable gradient directions are searched as the starting points of the curve segment. The instability of the gradient direction can be determined by calculating the angle between the line connecting the two unrefined edge pixels in the neighborhood and the center pixel point. If the angle is greater than 30 degrees, the gradient direction is considered unstable and these edge pixels are used as the starting points of the curve segment.

[0110] For each starting point of the curve segment found, the gradient direction is used as the initial direction of the curve segment. Along the initial direction of the curve segment, based on the curvature characteristics of the curve segment, the range of the curve segment is determined by tracking the curvature change of the edge pixels. The unrefined edge pixels in the next neighborhood and the curvature between adjacent unrefined edge pixels are checked in turn. The curvature can be estimated by the angle change between adjacent unrefined edge pixels. For example, the angle change between adjacent unrefined edge pixels is , the straight-line distance between adjacent unrefined edge pixels is , then the curvature between adjacent unthinned edge pixels is , thereby continuously calculating the curvature between adjacent unrefined edge pixels. When the reciprocal of the curvature is greater than the set minimum curvature radius, the reciprocal of the curvature is the curvature radius (at this time ), which means that the curvature of the curve in this section is relatively small and meets the set curve segment end condition, then it is considered that a curve segment is found, and the last unrefined edge pixel is the end point of the curve segment, thereby accurately obtaining the curve segment information in the shield tool topology structure and storing the curve segment information in the curve segment list.

[0111] Through the above-mentioned detection logic of neighborhood connection, key features such as bifurcation points, intersection points, straight line segments and curve segments in the shield tool topology structure can be accurately extracted. These key features completely describe the shape and structural characteristics of the shield tool, providing accurate data support for subsequent image recognition. The extracted bifurcation points, intersection points, straight line segment information and curve segment information are stored in corresponding lists respectively, making these topological structure information more orderly and easy to manage. In the subsequent image recognition process, these topological structure information can be quickly called and compared, greatly improving the accuracy and efficiency of judging the tool type and wear degree of the shield tool.

[0112] After completing the adaptive adjustment of structural elements in morphological operations, the execution of morphological operations, and the extraction of shield tool topological structure information, the block indexes of all shield image blocks are preliminarily arranged according to their positions in the original shield tool image based on the block index of each shield image block to form a preliminary image stitching framework. Based on the preliminary stitching, topological structure alignment is performed using the topological structure information.

[0113] For example, it is found that there is a straight line segment on the right edge of shield image block A, and there is also a straight line segment on the left edge of shield image block B, and these two straight line segments should be continuous in the topological structure. By fine-tuning the position of shield image block B, these two straight line segments are accurately connected; similar matching and alignment operations are also performed for bifurcation points and intersections. If there is a bifurcation point in the lower right corner of shield image block A, and there is also a corresponding bifurcation point in the upper left corner of shield image block C, ensure that the two bifurcation points are in the same position after splicing, and the surrounding topological structure can be naturally connected.

[0114] After completing the position matching and topological structure alignment, the overlapping areas of adjacent shield image blocks are fused. For example, for the overlapping area of shield image blocks A and B, the topological structure information is used to determine how the pixels in the overlapping area should be fused. If the topological structure in the overlapping area shows that there should be a continuous curve segment, then during fusion, the pixels in the overlapping area are smoothly transitioned so that the curve segment looks natural and continuous after splicing, without obvious splicing marks. The image quality is improved, which helps to observe and analyze the characteristics of the shield tool more clearly. The pixel values of the overlapping area can be calculated using methods such as weighted averaging to obtain the fused pixel values.

[0115] By combining the block index and topological structure information for fusion, the integrity of the shield tool image is ensured, so that the various parts of the shield tool are accurately spliced together. Topological features such as bifurcation points, intersections, straight segments and curved segments can be naturally connected and presented, without obvious fractures or dislocations caused by block processing. This provides complete image data for subsequent image recognition to determine the tool type and degree of wear of the shield tool.

[0116] The image recognition module is used to determine the tool type, wear degree and wear type of the shield tool based on the contour information, edge gradient information and topological structure information of the shield tool.

[0117] Specifically, the judgment logic of the shield tool type, wear degree and wear type includes:

[0118] Build a standard shield tool model library;

[0119] Perform similarity calculation and comparison on the profile information of the shield tool and the profile information in the standard shield tool model library to obtain a candidate list of shield tools;

[0120] Compare the edge gradient information of the shield tool with the edge gradient information in the standard shield tool model library to narrow down the candidate list of shield tools;

[0121] Deeply match the topological structure information of the shield tool with the topological structure information in the standard shield tool model library to determine the tool type of the shield tool;

[0122] The wear degree and wear type of shield tool are comprehensively judged based on the contour information and edge gradient information.

[0123] The standard shield tool model library is the basis for subsequent comparison and judgment. It contains the contour information, edge gradient information and topological structure information of various types of shield tools in normal conditions. By comparing with the standard shield tool model library, the type, wear degree and wear type of the tool to be tested can be accurately judged.

[0124] The contour information, edge gradient information and topological structure information of shield tools of various tool types are collected to build a standard shield tool model library. This information represents the typical characteristics of shield tools of different tool types and is the basis for subsequent comparison and judgment. By establishing a standard shield tool model library, the actual collected shield tool image features can be compared with it to identify the tool type, wear degree and wear type of the shield tool. The tool types include hobs, scrapers and cutters. For example, through multiple images at different angles and lighting conditions, the contour information, edge gradient information and topological structure information of shield tools of different tool types are extracted after processing. Among them, the contour of the hob is relatively rounded, and the edge gradient is relatively strong in some parts. The cutter has a unique bifurcation structure, and its topological structure has obvious branch point characteristics. The edge of the scraper is relatively straight, and the straight line segment accounts for a large proportion in the topological structure. This information is sorted and stored in the standard shield tool model library.

[0125] Contour information is the basic shape feature of shield cutters. Shield cutters of different tool types usually have different contour shapes. By calculating the similarity between the contour information of the shield cutter and the contour information in the standard shield cutter model library, it is possible to preliminarily screen out tool types with similar contours to the current shield cutter and form a candidate list of shield cutters. Commonly used similarity calculation methods such as Euclidean distance and cosine similarity can quantify the degree of difference between the two contours. For example, if the calculation results show that the similarity between the contour information of the shield cutter and the contour information of the scraper and roller is greater than a preset threshold, the scraper and roller will be included in the candidate list of shield cutters.

[0126] Shield cutters of different tool types also have differences in edge characteristics. Even shield cutters with similar profiles may have different edge gradient information due to different designs and manufacturing processes. The edge gradient information of the shield cutter is compared with the edge gradient information in the standard shield cutter model library to further narrow the candidate list of shield cutters and exclude some tool types that have only similar profiles but mismatched edge gradient information. For example, the edge gradient information of the shield cutter is compared with the edge gradient information of the scraper and hob in the standard shield cutter model library. It is found that the distribution and intensity of the edge gradient of the shield cutter in some key parts are more consistent with the hob, but there are significant differences with the scraper. Therefore, the scraper is excluded from the candidate list, leaving only the hob.

[0127] Topological structure information is a representation of the internal structure and shape of the shield cutter. Shield cutters of different cutter types often have unique patterns in topological structure, which has a high degree of recognition. By deeply matching the topological structure information of the shield cutter with the topological structure information in the standard shield cutter model library, the tool type of the shield cutter can be accurately determined; for example, the topological structure information of the shield cutter is further deeply matched with the topological structure information of the hob, and the branch points, intersection points, straight line segment information and curve segment information are carefully checked. It is found that the topological structures of the two are completely consistent, thereby determining that the tool type of the shield cutter is a hob; by constructing a standard shield cutter model library and comparing the contour information, edge gradient information and topological structure information in turn, the tool type of the shield cutter can be accurately determined. This step-by-step screening and precise matching method greatly improves the accuracy of tool type judgment and reduces the possibility of misjudgment.

[0128] Contour information and edge gradient information can reflect the wear of the tool. Comprehensive consideration of these two types of information can more accurately determine the degree of wear and the type of wear. During the actual use of shield tools, the contour and edge gradient of the shield tool will change. Wear will cause contour deformation and edge blunting, requiring the replacement of the shield tool. These changes will be reflected in the contour information and edge gradient information. By comparing the contour information and edge gradient information of the current shield tool with the contour information and edge gradient information under normal conditions in the standard shield tool model library, the degree of wear of the shield tool can be determined. For example, the degree of contour deformation and the weakening of edge gradient strength can be used as the basis for judging the degree of wear.

[0129] Furthermore, the judgment sub-logic of the wear degree of the shield tool includes:

[0130] Extract the contour area from the contour information, and calculate the contour deformation variable of the shield tool's contour area and the contour area in the standard shield tool model library through Euclidean distance;

[0131] Extract the gradient intensity from the edge gradient information and calculate the gradient attenuation between the gradient intensity of the shield tool and the gradient intensity in the standard shield tool model library;

[0132] Configure the contour deformation threshold and the gradient attenuation threshold, compare the contour deformation amount with the contour deformation threshold to obtain the contour deformation degree, and compare the gradient attenuation amount with the gradient attenuation threshold to obtain the gradient attenuation degree;

[0133] The degree of shield tool wear is comprehensively judged based on the degree of profile deformation and gradient attenuation.

[0134] The change in contour area is an important manifestation of shield tool wear. The degree of shield tool wear can be quantified by calculating the contour deformation variable. For the extracted contour, the pixel counting method is used to calculate the number of pixels it contains, and the number of pixels is multiplied by the actual area corresponding to a single pixel. The actual area corresponding to a single pixel can be obtained by calibrating the resolution and actual size of the image, thereby obtaining the actual area of the contour. The contour area of a standard shield tool of the same type as the tool to be detected is found in the standard shield tool model library. Suppose the contour area of the tool to be detected is A1, and the contour area of the standard shield tool is A0, then the contour deformation variable is |A1-A0|; thereby, the change in the contour area of the shield tool is quantified, making the degree of wear between different shield tools comparable, facilitating the subsequent setting of thresholds for judgment, and providing a specific value for subsequent comparison with the contour deformation threshold, which is the basis for judging the degree of contour deformation.

[0135] The attenuation of gradient intensity is also an important manifestation of shield tool wear. The degree of shield tool wear can be further quantified by calculating the gradient attenuation. The gradient intensity G1 of the shield tool is obtained from the above steps, and the gradient intensity G0 of the corresponding standard shield tool is obtained from the standard shield tool model library. The gradient attenuation is calculated as |G1-G0|. The wear of the shield tool surface is quantified from a microscopic perspective. Combined with the change in contour area, it can more comprehensively reflect the degree of shield tool wear and provide a specific value for subsequent comparison with the gradient attenuation threshold to determine the degree of gradient attenuation.

[0136] By setting the profile deformation threshold and gradient attenuation threshold, the profile deformation and gradient attenuation can be converted into specific deformation and attenuation degrees, facilitating a comprehensive assessment of the degree of wear. A large number of shield tool samples with varying degrees of wear are collected, and the profile deformation and gradient attenuation are calculated for each of these samples. The standard deviation is used to determine the ranges of profile deformation and gradient attenuation corresponding to different degrees of wear (mild, moderate, and severe), thereby determining the profile deformation threshold and gradient attenuation threshold. The profile deformation threshold includes a profile deformation base value (mild deformation threshold) and a profile deformation extreme value (moderate deformation threshold). The gradient attenuation threshold includes a gradient attenuation base value (mild attenuation threshold) and a gradient attenuation extreme value (moderate attenuation threshold). The profile deformation is compared with the profile deformation threshold, and the gradient attenuation is compared with the gradient attenuation threshold to obtain the quantified profile deformation degrees (mild, moderate, and severe) and gradient attenuation degrees (mild, moderate, and severe), respectively. The quantified wear indicators are converted into intuitive degree grades, allowing operators to quickly understand the wear condition of the shield tool and providing clear degree grade information for comprehensive assessment of the wear degree.

[0137] Comprehensive consideration of the degree of contour deformation and the degree of gradient attenuation can more comprehensively and accurately judge the degree of wear of the worn tool; based on actual experience and experimental data, the following comprehensive judgment rules are formulated: if the degree of contour deformation and the degree of gradient attenuation are both mild, the degree of wear is mild; if one is moderate and the other is mild, the degree of wear is moderate; if one is severe, or both are moderate, the degree of wear is severe; thereby improving the accuracy and reliability of wear degree judgment, avoiding the limitations of single indicator judgment, and providing accurate wear degree information for the result output module to display and generate early warning information.

[0138] Furthermore, the judgment sub-logic of the wear type of the shield tool includes:

[0139] Mirror the shield cutter along its central axis and calculate the symmetry of the profile before and after flipping.

[0140] Extract the gradient direction from the edge gradient information, count the distribution of the gradient direction on the shield tool contour, and calculate the mean value of the gradient direction of different shield image blocks to determine the gradient direction change;

[0141] The wear type of shield tool is comprehensively judged based on the profile symmetry and gradient direction change. The wear types include eccentric wear, chipping and uniform wear.

[0142] Different wear types will cause the symmetry of the tool profile to change to varying degrees. Calculating the profile symmetry can provide an important basis for determining the wear type. The center axis of the shield tool profile is found, the profile is mirrored along the center axis, and then the similarity of the profiles before and after the flip is calculated, such as the overlapping area ratio, as a measure of profile symmetry. This quantifies the symmetry of the shield tool profile, provides a specific numerical basis for subsequent wear type determination, and provides specific numerical information for determining the wear type based on profile symmetry.

[0143] Different wear types will lead to different distributions and changes in the gradient direction of the tool edge. This change can be reflected by counting the distribution number of gradient directions and calculating the mean, providing a basis for judging the wear type; extract the edge gradient direction information of the tool to be detected, divide the gradient direction into several intervals, such as 0-180 degrees into 10 intervals, count the number of pixels in the gradient direction in each interval, and calculate the mean of the gradient direction in each shield image block at the same time, compare the difference in the mean values of the gradient directions of adjacent shield image blocks, count the number of shield image blocks with a difference greater than the difference threshold (such as 15°), and determine the change in the gradient direction; thereby quantifying the distribution and change of the gradient direction, providing a specific numerical basis for the subsequent judgment of the wear type, and providing specific numerical information for judging the wear type according to the change in the gradient direction.

[0144] Comprehensive consideration of contour symmetry and gradient direction changes can more accurately judge the wear type of shield tools; according to the combination of contour symmetry and gradient direction changes, judgment rules are formulated. If the contour symmetry is poor (obtained by comparing the contour symmetry with the symmetry threshold), it is eccentric wear; if the contour symmetry is not bad, the gradient direction change is judged; if the gradient direction changes greatly (obtained by comparing the number of shield image blocks with statistical differences greater than the difference threshold with the number threshold), it is chipping; if the gradient direction changes uniformly, it is uniform wear; thereby improving the accuracy of wear type judgment and providing accurate wear type information to the result output module for displaying and generating early warning information.

[0145] At the same time, in actual applications, if it is identified that the contour direction of the hob does not change for a long time, it can also be judged as eccentric wear of the tool. For example, after the hob is stuck, only one side is worn, and the other side does not change. Then we need to first judge according to the judgment sub-logic of the wear type of the above-mentioned shield tool. If the preliminary judgment result cannot directly obtain the wear type judgment result of eccentric wear, chipping and uniform wear, then it is necessary to introduce the change time of the contour direction for secondary judgment. For example, set a detection time period (such as 5 minutes). When it is detected that the hob continues to rotate within the detection time period, but the contour direction does not change, it is judged that the final wear type of the hob is eccentric wear, thereby improving the accuracy and reliability of the wear type judgment and ensuring the safe progress of shield construction.

[0146] The result output module is used to display the tool type, wear degree and wear type of the shield tool and generate early warning information.

[0147] After the tool type, wear degree and wear type of the shield tool are determined in the image recognition module, corresponding warning information needs to be generated according to the wear degree. Different wear degrees are pre-set to correspond to different warning levels. For example, light wear corresponds to a yellow warning, moderate wear corresponds to an orange warning, and heavy wear corresponds to a red warning. This setting is to allow operators to intuitively understand the wear condition of the shield tool so that they can take corresponding measures to determine whether the shield tool needs to be replaced; for example, if the shield tool is judged to be lightly worn, then the warning information is a yellow warning.

[0148] In order to allow operators to fully understand the condition of shield cutters, it is necessary to integrate and display the cutter type, wear degree, wear type and warning information of the shield cutters, so that operators can obtain key information at a glance and make reasonable decisions; for example, "Cutter type: hob, wear degree: light wear, wear type: uniform wear, warning information: yellow warning" is displayed on the display screen, and this information is output in the form of data reports for easy recording and archiving.

[0149] In order to ensure that relevant operators can know the wear of shield cutters in a timely manner, it is necessary to send early warning information through various means, which helps to take maintenance measures in time to avoid the impact of excessive wear of shield cutters on construction progress and safety; such as through the built-in notification function of the system, a push notification is sent to the mobile phone APP of the relevant operator, with the content of "The roller cutter has slight wear and the wear type is uniform wear. Please pay attention to its usage". At the same time, a sound prompt is issued in the monitoring center at the construction site to remind the on-duty personnel to pay attention.

[0150] By integrating and displaying the tool type, wear degree, wear type and warning information of the shield cutter and providing multi-channel notifications, relevant operators can obtain the status information of the shield cutter in a timely and accurate manner. Whether at the construction site or through remote monitoring, they can easily understand the condition of the shield cutter, thereby improving the efficiency and accuracy of information communication. Clear warning information and detailed displays of wear degree and wear type provide operators with a clear basis for decision-making. In the stage of mild wear, operators can arrange maintenance plans in advance and prepare for tool replacement to avoid increased wear of the shield cutter. This helps to ensure the safety and efficiency of shield construction and reduce construction delays and accident risks caused by tool problems.

Claims

1. The shield tool image recognition system based on FPGA morphological operator is characterized by: include: Image acquisition module, morphological processing module, image recognition module and result output module; The image acquisition module is used to acquire the shield tool image and divide the shield tool image into shield image blocks according to the computing resources of the FPGA to pre-process the shield tool image in parallel; The morphological processing module is used to adjust the structural elements according to the edge direction and texture characteristics of the shield image block, construct a morphological operation array, process the shield tool image through the structural elements, so as to extract the contour information and edge gradient information of the shield tool, determine whether the pixel point of the shield image block is an edge pixel based on the contour information and edge gradient information of the shield tool, obtain the unrefined edge pixels, detect the neighborhood connectivity of the edge pixels in the shield image block to refine and extract the topological structure information, and fuse the shield image block in combination with the topological structure information; The image recognition module is used to determine the tool type, wear degree and wear type of the shield tool based on the profile information, edge gradient information and topological structure information of the shield tool; The result output module is used to display the tool type, wear degree and wear type of the shield tool and generate early warning information.

2. The shield tool image recognition system based on FPGA morphological operator according to claim 1, characterized in that: The processing strategy of the shield tool image includes: Adaptively adjust the shape and size of the structural element according to the edge direction and texture characteristics of the shield image block; Constructing a morphological operation array to perform morphological operations through structural elements to obtain the contour information of the shield tool and extract the edge gradient information of the shield tool; Refine and extract the topological structure information of the shield tool based on its contour information and edge gradient information; The shield image blocks are fused according to the block index of the shield image blocks and combined with the topological structure information of the shield tool.

3. The shield tool image recognition system based on FPGA morphological operator according to claim 2, characterized in that: The logic for adaptively adjusting the shape and size of the structural element includes: Determine the texture feature of each shield image block based on the local contrast of the shield image block, and determine the edge direction of each shield image block based on the gradient direction clustering of the shield image block; Adjust the size of the structural element according to the texture characteristics of the shield image block; The shape of the structural element is adjusted according to the edge direction of the shield image block.

4. The shield tool image recognition system based on FPGA morphological operator according to claim 3 is characterized in that: The execution logic of the morphological operation includes: For each shield image block, according to the shape and size of the structural element, the upper left corner pixel of each shield image block is used as the sliding starting point, and the shield image blocks are traversed one by one; In the process of traversing the shield image block, the erosion operation array, the expansion operation array, the opening operation array and the closing operation array are sequentially executed to obtain the contour information of the shield tool; The edge gradient information of the shield tool is extracted according to the contour information of the shield tool.

5. The shield tool image recognition system based on FPGA morphological operator according to claim 4 is characterized in that: The refinement and extraction method of the topology structure information includes: Based on the contour information and edge gradient information of the shield tool, it is determined whether the pixel point of each shield image block is an edge pixel; Taking the edge pixel as the center, count the number of edge pixels in the neighborhood of the edge pixel to determine whether the edge pixel is thinned, and continue iterating until the thinning termination condition is met; Obtain unrefined edge pixels and detect the neighborhood connectivity of edge pixels in each shield image block to determine the topological structure information of the shield tool; Organize and mark the topological structure information of shield tools.

6. The shield tool image recognition system based on FPGA morphological operator according to claim 5, characterized in that: The detection logic of the neighborhood connection situation includes: Traverse and count the unrefined edge pixels, and determine the information of bifurcation points and intersection points through threshold judgment; The search is for edge pixels with a neighborhood connectivity of 2 and a stable gradient direction as the starting point of the straight line segment. The gradient direction is used as the initial direction of the straight line segment. The unrefined edge pixels in the next neighborhood along the gradient direction are sequentially checked, as well as the deviation angle between the edge pixel and the initial direction of the straight line segment, to determine the end point of the straight line segment and obtain the straight line segment information. The edge pixel with a search neighborhood connection number of 2 and an unstable gradient direction is used as the starting point of the curve segment. The gradient direction is used as the initial direction of the curve segment. The unrefined edge pixels in the next neighborhood and the curvature between adjacent unrefined edge pixels are checked along the gradient direction to determine the end point of the curve segment and obtain the curve segment information.

7. The shield tool image recognition system based on FPGA morphological operators according to claim 6, characterized in that: The judgment logic of the tool type, wear degree and wear type of the shield tool includes: Build a standard shield tool model library; Perform similarity calculation and comparison on the profile information of the shield tool and the profile information in the standard shield tool model library to obtain a candidate list of shield tools; Compare the edge gradient information of the shield tool with the edge gradient information in the standard shield tool model library to narrow down the candidate list of shield tools; Deeply match the topological structure information of the shield tool with the topological structure information in the standard shield tool model library to determine the tool type of the shield tool; The wear degree and wear type of shield tool are comprehensively judged based on the contour information and edge gradient information.

8. The shield tool image recognition system based on FPGA morphological operators according to claim 7, characterized in that: The shield tool image is acquired through a camera, and the preprocessing logic of the shield tool image includes: Segment the shield tool image based on the FPGA's computing resources to obtain shield image blocks; Establish a block index for each shield image block; The denoising and enhancement operations of each shield image block are performed in parallel in the FPGA.

9. The shield tool image recognition system based on FPGA morphological operators according to claim 8, characterized in that: The judgment sub-logic of the wear degree of the shield tool includes: Extract the contour area from the contour information, and calculate the contour deformation variable of the shield tool's contour area and the contour area in the standard shield tool model library through Euclidean distance; Extract the gradient intensity from the edge gradient information and calculate the gradient attenuation between the gradient intensity of the shield tool and the gradient intensity in the standard shield tool model library; Configure the contour deformation threshold and the gradient attenuation threshold, compare the contour deformation amount with the contour deformation threshold to obtain the contour deformation degree, and compare the gradient attenuation amount with the gradient attenuation threshold to obtain the gradient attenuation degree; The degree of shield tool wear is comprehensively judged based on the degree of profile deformation and gradient attenuation.

10. The shield tool image recognition system based on FPGA morphological operators according to claim 9, characterized in that: The judgment sub-logic of the wear type of the shield tool includes: Mirror the shield cutter along its central axis and calculate the symmetry of the profile before and after flipping. Extract the gradient direction from the edge gradient information, count the distribution of the gradient direction on the shield tool contour, and calculate the mean value of the gradient direction of different shield image blocks to determine the gradient direction change; The wear type of shield tool is comprehensively judged based on the profile symmetry and gradient direction change. The wear types include eccentric wear, chipping and uniform wear.

Images