Linear flexible body composite detection method, system, equipment, medium and product
Through the detection method of fusing two-dimensional and three-dimensional feature information, the problem of low detection accuracy and efficiency of linear flexible bodies in complex environments is solved, and a more efficient and accurate detection effect is achieved.
Patent Information
- Application Number
- CN202510341631.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-21
- Publication Date
- 2025-07-11
AI Technical Summary
Traditional visual detection systems are difficult to effectively deal with the difficulty in accurately capturing and identifying the appearance and position of linear flexible bodies due to deformation such as bending and distortion during movement or positioning, which affects detection accuracy and efficiency.
The fusion method of two-dimensional feature information and three-dimensional feature information is adopted to obtain RGB images through CCD cameras and point cloud data is obtained through 3D cameras, and linear flexible bodies are identified in combination with model-level fusion technology, including improved Otsu’s method, bilateral filtering, Canny edge detection, region growth algorithm, statistical outlier removal, convolution operation and maximum pooling method and other technical means.
It improves the detection accuracy and efficiency of linear flexible bodies in complex environments, reduces the detection cost and burden, and provides a more accurate and reliable detection solution for industrial automation and intelligence.
Smart Images
Figure CN120298324A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of object detection, and in particular, to a composite detection method, system, device, medium, and product for linear flexible bodies. Background Art
[0002] A linear flexible body refers to an object with a linear or one-dimensional structure, having high flexibility and low stiffness, and can undergo large-scale bending, stretching, or twisting deformations under the action of external forces. Such objects widely exist in nature and engineering technologies, such as cables, ropes, fiber materials, etc. Due to their flexible characteristics, linear flexible bodies can adapt to environmental changes during application and deform freely under certain conditions, having important application values and research significances.
[0003] With the continuous development of automation and intelligent manufacturing, industrial robots are increasingly widely used in production lines, especially in complex assembly, processing, quality inspection and other fields. In these application scenarios, visual detection technology, as an important quality monitoring means, has become an indispensable part of industrial robot systems. However, linear flexible bodies are prone to bending, twisting and other deformations during movement or positioning, which makes it difficult to accurately capture and identify their appearance and position. Traditional visual detection systems often have difficulty effectively coping with these changes, resulting in a decrease in detection accuracy and affecting production quality. Summary of the Invention
[0004] The purpose of this application is to provide a composite detection method, system, device, medium, and product for linear flexible bodies, which can improve the detection efficiency and accuracy of linear flexible bodies in complex environments.
[0005] To achieve the above purpose, this application provides the following solutions:
[0006] In the first aspect, this application provides a composite detection method for linear flexible bodies, including:
[0007] Obtain the RGB image of the linear flexible body;
[0008] Detect the two-dimensional feature information of the linear flexible body based on the RGB image;
[0009] Obtain the three-dimensional point cloud data of the linear flexible body;
[0010] Detect the three-dimensional feature information of the linear flexible body based on the three-dimensional point cloud data;
[0011] Adopt a model-level fusion method to fuse the two-dimensional feature information and the three-dimensional feature information to identify the linear flexible body.
[0012] In the second aspect, this application provides a composite detection system for linear flexible bodies, including:
[0013] A CCD camera for acquiring RGB images of a linear flexible body;
[0014] A two-dimensional feature information detection module for detecting two-dimensional feature information of a linear flexible body based on the RGB image;
[0015] A 3D camera for acquiring three-dimensional point cloud data of a linear flexible body;
[0016] A three-dimensional feature information detection module for detecting three-dimensional feature information of a linear flexible body based on the three-dimensional point cloud data;
[0017] A linear flexible body recognition module for fusing the two-dimensional feature information and the three-dimensional feature information by using a model-level fusion method to recognize the linear flexible body.
[0018] In a third aspect, the present application provides a computer device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the computer program to implement the above-mentioned linear flexible body composite detection method.
[0019] In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the above-mentioned linear flexible body composite detection method is implemented.
[0020] In a fifth aspect, the present application provides a computer program product, including a computer program, and when the computer program is executed by a processor, the above-mentioned linear flexible body composite detection method is implemented.
[0021] According to the specific embodiments provided by the present application, the present application has the following technical effects:
[0022] The present application provides a linear flexible body composite detection method, system, device, medium and product, which identifies the linear flexible body through the two-dimensional feature information and the three-dimensional feature information of the linear flexible body. The present application combines the two-dimensional detection technology and the three-dimensional detection technology, can more comprehensively perceive and analyze the shape and state of the linear flexible body, can detect the linear flexible body more accurately and efficiently in a complex environment, reduces the cost and burden of the industrial robot in the detection process, and provides a more accurate and reliable detection solution for industrial automation and intelligence. Description of the Drawings
[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.
[0024] Figure 1 Flow schematic diagram of the linear flexible body composite detection method provided in an embodiment of the present application;
[0025] Figure 2 Effect schematic diagram of fitting the linear flexible body;
[0026] Figure 3 Structure schematic diagram of a computer device provided in an embodiment of the present application. Specific embodiments
[0027] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.
[0028] In recent years, with the continuous development of machine vision technology, the composite detection technology that fuses two-dimensional feature information and three-dimensional feature information has received increasing attention in key applications such as industrial manufacturing and the field of robotics. By fusing two-dimensional and three-dimensional features, the deformation and motion trajectory of an object can be better captured. The core steps of this technology include feature extraction, feature description, and classification and recognition. However, so far, in the field of linear flexible body detection, there are few applications of the technology that fuses two-dimensional feature information and three-dimensional feature information. If this technology can be applied to the real-time detection of linear flexible bodies and break through the traditional linear flexible body detection methods, the detection efficiency and accuracy of linear flexible bodies in complex environments will be improved. Therefore, the present application provides a linear flexible body composite detection method, system, device, medium, and product that fuse two-dimensional detection technology and three-dimensional detection technology.
[0029] To make the above objects, features, and advantages of the present application more obvious and understandable, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.
[0030] In an exemplary embodiment, as Figure 1 shown, a linear flexible body composite detection method is provided. This method is executed by a computer device, which can specifically be executed alone by a computer device such as a terminal or a server, or jointly executed by a terminal and a server. In the embodiments of the present application, this method is described by taking it as an example applied to a server, and includes the following steps S1 to S5. Among them:
[0031] S1: Obtain the RGB image of the linear flexible body.
[0032] In a specific embodiment, a CCD camera is used to obtain the RGB image of the linear flexible body.
[0033] S2: Detect the two-dimensional feature information of the linear flexible body based on the RGB image.
[0034] In a specific embodiment, step S2 includes the following steps S21 to S28.
[0035] S21: Preprocess the RGB image to obtain an enhanced image.
[0036] (1) Use an improved Otsu's method to perform adaptive threshold segmentation on the RGB image to obtain a binary image.
[0037] In this application, a local area adaptive adjustment mechanism is introduced based on the Otsu's method, making the image segmentation effect more accurate in scenes with complex lighting conditions and avoiding the problem of regional failure in the case of uneven background in traditional methods. Specifically, it includes the following steps:
[0038] 1) Use the classic Otsu's method to calculate the global threshold. This step can effectively perform preliminary segmentation in the case of relatively uniform lighting.
[0039] 2) After calculating the global threshold, divide the image into multiple local areas and perform dynamic division according to the local brightness change of the image. Specifically, calculate the brightness standard deviation of each local area. A larger standard deviation of the area indicates a larger brightness change, which may have more texture information or local details. Based on these brightness features, the image can be divided into multiple local areas, making the brightness change relatively uniform within each local area and avoiding the situation of global threshold failure.
[0040] According to the local features such as the brightness change and texture features of the image, divide the image into multiple small areas. The division method is to judge the regional boundary through texture or brightness change, so as to divide the image. The image features (brightness, texture, etc.) of each area have relative uniformity and consistency. This means that the brightness information of each local area is relatively stable, which helps to avoid the failure or error in global threshold calculation.
[0041] 3) For each local area, calculate the histogram of the local image and apply the Otsu's method to obtain the local threshold of the area.
[0042] For each local area, calculate the gray value distribution of all pixels in the area. The histogram of the image is obtained by counting the frequency of each gray value (from 0 to 255) in the local area. Once the histogram of the local area is obtained, the Otsu's method can be used to calculate the local threshold of the area.
[0043] 4) In regions with large illumination variations, local thresholds and global thresholds may conflict. In such cases, a dynamic adjustment mechanism is introduced. The strategy of the dynamic adjustment mechanism is as follows: for regions with uniform local illumination, the global threshold is mainly used for image segmentation; for regions with uneven or large illumination variations, local thresholds are preferentially considered for image segmentation; if the difference between the local threshold and the global threshold is large, a weighted average method or an adaptive adjustment strategy is adopted. This improves the accuracy and robustness of image segmentation. The output of the dynamic adjustment mechanism is an adjusted threshold, and then image segmentation is performed.
[0044] (2) The binary image is smoothed using a bilateral filtering method to obtain a denoised image, which can remove detailed noise while retaining edge information.
[0045] (3) The denoised image is enhanced by combining the Laplacian operator and an adaptive sharpening method to obtain an enhanced image.
[0046] Specifically, the Laplacian operator is applied to the denoised image to extract high-frequency detail information in the image and generate a detail intensity map, which reflects the richness of details in each region of the image; based on the detail intensity map, a sharpening intensity is dynamically assigned to each pixel position through a non-linear mapping function. Higher sharpening intensities are assigned to regions with rich details to enhance the details, while lower sharpening intensities are assigned to smooth regions to avoid over-enhancing noise; according to the sharpening intensity of each pixel, the output of the Laplacian operator is combined with the original image for adaptive sharpening processing to obtain an enhanced image.
[0047] This application combines the Laplacian operator with an adaptive sharpening method and automatically selects an appropriate sharpening intensity according to the detail level of the denoised image, which can avoid the problem of over-enhancing noise during the sharpening process.
[0048] S22: The Canny edge detection method is used to perform edge detection on the enhanced image to obtain a Canny edge map. This can improve the robustness of the shape of linear flexible bodies in complex scenarios.
[0049] S23: Based on the Canny edge map, a region growing algorithm is used to extract connected regions to obtain a labeled image.
[0050] Specifically, based on the Canny edge detection results, the edge information in the image is extracted, and potential seed points are screened out by setting a threshold. Starting from these seed points, the region growing algorithm is used to expand the connected regions. By setting growth conditions such as pixel gray value similarity and spatial neighborhood relationship, the adjacent pixels that meet the conditions are gradually incorporated into the current connected region. During this process, combined with the constraints of the Canny edge map, it is ensured that the region growth does not cross the edge boundary, thus avoiding misidentification caused by noise interference. The expanded connected regions are marked to generate a marked image.
[0051] This application uses the region growing algorithm to extract connected regions in combination with seed points. By expanding the seed points in the Canny edge map region, the extraction of connected regions is made more accurate, avoiding misidentification caused by noise interference.
[0052] S24: Identify the area of the connected region based on the marked image.
[0053] Traverse the marked image, count the number of pixels in each connected region, and obtain the area of the connected region.
[0054] S25: According to the area of the connected region, use the clustering algorithm to divide the connected domains and determine the connected domain with the largest area.
[0055] S26: Draw an ellipse for the connected domain with the largest area on the Canny edge map.
[0056] On the Canny edge map, use the cv2.ellipse() function of OpenCV to draw an ellipse for the connected domain with the largest area.
[0057] S27: Fit the drawn ellipse and use the edge detection algorithm to obtain the contour of the linear flexible body.
[0058] Merge the drawn ellipses and fit a new fused figure. Use the edge detection algorithm to process this fused figure and extract the edge of the fused figure as the contour of the linear flexible body, as Figure 2 shown.
[0059] S28: Determine the two-dimensional characteristic parameters of the linear flexible body based on the contour; the two-dimensional characteristic parameters are two-dimensional characteristic information.
[0060] The contour shape of the linear flexible body characterizes the morphology of the linear flexible body. The two-dimensional characteristic parameters include information such as curvature, tangent direction, degree of bend, and centroid coordinates.
[0061] S3: Obtain the three-dimensional point cloud data of the linear flexible body.
[0062] In a specific embodiment, a linear flexible body is scanned by a 3D camera to generate three-dimensional point cloud data.
[0063] S4: Detect the three-dimensional feature information of the linear flexible body based on the three-dimensional point cloud data.
[0064] In a specific embodiment, step S4 includes the following steps S41 to S44.
[0065] S41: Preprocess the three-dimensional point cloud data.
[0066] Use the Statistical Outlier Removal (SOR) algorithm to calculate the standard deviation of the distances of the neighboring points of each point in the three-dimensional point cloud data, and remove the points with a standard deviation greater than the set threshold to obtain the denoised three-dimensional point cloud data, effectively removing noise.
[0067] Adopt the Farthest Point Sampling (FPS) method to uniformly sample the denoised three-dimensional point cloud data, and select points far from the existing sampled points from the three-dimensional point cloud data to uniformly cover the distribution of the three-dimensional point cloud data.
[0068] Normalize the three-dimensional point cloud data after uniform distribution sampling. Specifically, map the coordinates of the three-dimensional point cloud data to the interval [0, 1], calculate the minimum and maximum values of each coordinate axis (x-axis, y-axis, z-axis), and use the normalization formula to convert each coordinate value to perform the normalization operation.
[0069] S42: Extract features from the preprocessed three-dimensional point cloud data to obtain the initial features of the linear flexible body; the initial features include geometric features, morphological features, and local directional features.
[0070] (1) The geometric features mainly include the degree of bending, the deformed shape, the bending moment distribution, and the deflection. Obtain the normal vectors of the neighboring points and analyze their rate of change. The greater the change in the normal vector, the greater the local curvature. By fitting the local surface or curve of the neighboring points, the curvature value at that point is estimated. The curvature reflects the degree of bending of the linear flexible body at a certain point. Combining the relationship between the curvature and the bending moment, the bending moment distribution and the stress distribution can be deduced. The change in curvature helps to calculate the local deflection and deformation mode of the linear flexible body. By integrating the curvature, the deformation path and geometric shape of the linear flexible body can be reconstructed.
[0071] (2) Obtain the discrete point cloud data of the surface or curve of the linear flexible body through methods such as sampling and grid division. Analyze the morphological characteristics of the local area by calculating the distribution density of the neighborhood points of each point, that is, the number of points within a certain neighborhood range. Areas with higher density usually correspond to key parts with curvature changes or morphological changes. By analyzing the topological relationship between neighborhood points, construct a local connection graph to reveal the relative position changes and connection methods between points, and then obtain the morphological characteristics of the linear flexible body. The morphological characteristics refer to the bending, twisting, contraction, etc. characteristics of the linear flexible body.
[0072] (3) Extract the local directional characteristics of the linear flexible body by calculating the direction vector of each point relative to the neighborhood points and combining the normal vector and tangent vector information.
[0073] The local directional characteristics of the linear flexible body are essentially to reveal information such as the geometric morphology, deformation trend, curvature change, and mechanical response of the point and its neighborhood by analyzing the geometric relationships between points (such as the changes in the direction vector, tangent vector, and normal vector). The direction vector refers to the vector from the current point to its neighborhood point. The normal vector refers to the vector perpendicular to the surface or curve of the linear flexible body, perpendicular to the surface. The tangent vector refers to the vector tangent to the local tangent plane of the curve or surface, representing the local orientation of the surface or curve.
[0074] S43: Aggregate the initial features using convolution operations and the max-pooling method to obtain the aggregated features.
[0075] Perform convolution operations using a method based on neighborhood sampling, and use the max-pooling method to perform feature aggregation on the geometric features, morphological features, and local directional features obtained in step S42 to extract useful features.
[0076] S44: Perform fusion and dimensionality reduction processing on the aggregated features through a fully connected layer to obtain the three-dimensional feature parameters of the linear flexible body; the three-dimensional feature parameters are three-dimensional feature information.
[0077] Feature fusion is achieved through a fully connected layer. The fully connected layer will fuse the features extracted by the convolutional layer into a more compact representation, retaining higher-level semantic information. Dimensionality reduction is usually achieved through the weight matrix and bias term in the fully connected layer. In the fully connected layer, the dimension of the features will be reduced, thereby performing dimensionality reduction.
[0078] S5: Use a model-level fusion method to fuse the two-dimensional feature information and the three-dimensional feature information to identify the linear flexible body.
[0079] In a specific embodiment, step S5 specifically includes the following steps S51 to S52.
[0080] S51: Concatenate and perform weighted operations on the two-dimensional feature information and the three-dimensional feature information to obtain a fused feature vector.
[0081] Dynamically adjust the fusion weight of the two-dimensional feature information and the three-dimensional feature information according to the needs of the task. When the shape information of the linear flexible body is important, increase the weight of the three-dimensional feature information to strengthen the role of the three-dimensional data; when the appearance, color, or texture feature information of the linear flexible body is important, increase the weight of the two-dimensional feature information to strengthen the role of the two-dimensional data. Ensure that various features can be reasonably integrated according to their importance during fusion.
[0082] S52: Input the fused feature vector into a fully connected layer and a softmax activation function in sequence to identify the linear flexible body. Display the detected linear flexible body on the host computer interface so that the operator can clearly see it.
[0083] The linear flexible body composite detection method provided by this application has high detection accuracy for linear flexible bodies, excellent efficiency, strong portability and stability, overcomes the disadvantages of complex traditional linear flexible body detection tasks, high cost, and low efficiency, and improves the accuracy and efficiency of linear flexible body detection.
[0084] Based on the same inventive concept, the embodiments of this application also provide a system for implementing the above-mentioned linear flexible body composite detection method. The implementation solutions provided by this system to solve problems are similar to the implementation solutions described in the above method. Therefore, the specific limitations in one or more embodiments of the following linear flexible body composite detection system can refer to the limitations on the linear flexible body composite detection method in the above text and will not be repeated here.
[0085] In an exemplary embodiment, a linear flexible body composite detection system is provided, including:
[0086] A CCD camera for acquiring the RGB image of the linear flexible body.
[0087] A two-dimensional feature information detection module for detecting the two-dimensional feature information of the linear flexible body based on the RGB image.
[0088] A 3D camera for acquiring the three-dimensional point cloud data of the linear flexible body.
[0089] A three-dimensional feature information detection module for detecting the three-dimensional feature information of the linear flexible body based on the three-dimensional point cloud data.
[0090] A linear flexible body recognition module for fusing the two-dimensional feature information and the three-dimensional feature information using a model-level fusion method to identify the linear flexible body.
[0091] In an exemplary embodiment, a computer device is provided, which includes a memory and a processor. A computer program is stored in the memory. When the processor executes the computer program, the steps in the above method embodiments are implemented. The computer device may be a server or a terminal, and its internal structure diagram may be as shown in Figure 3 shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data to be processed. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals through a network connection. When the computer program is executed by the processor, a linear flexible body composite detection method is implemented.
[0092] Those skilled in the art can understand that Figure 3 the structure shown in is only a block diagram of a part of the structure related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements. In an exemplary embodiment, a computer device is provided, which includes a memory and a processor. A computer program is stored in the memory. When the processor executes the computer program, the steps in the above method embodiments are implemented.
[0093] In an exemplary embodiment, a computer-readable storage medium is provided, which stores a computer program. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.
[0094] In an exemplary embodiment, a computer program product is provided, which includes a computer program. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.
[0095] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.
[0096] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memories can include read-only memory (ROM), magnetic tapes, floppy disks, flash memories, optical memories, high-density embedded non-volatile memories, resistive random access memories (ReRAMs), magnetoresistive random access memories (MRAMs), ferroelectric random access memories (FRAMs), phase change memories (PCMs), graphene memories, etc. Volatile memories can include random access memories (RAMs) or external caches, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.
[0097] The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logics, data processing logics based on quantum computing, etc., without limitation.
[0098] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.
[0099] Specific examples are used in this article to elaborate on the principles and implementation manners of the present application. The descriptions of the above embodiments are only used to help understand the methods and core ideas of the present application; at the same time, for those of ordinary skill in the art, according to the ideas of the present application, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present application.
Claims
1. A linear flexible body composite detection method, characterized in that, Comprising: Obtain the RGB image of the linear flexible body; Detect the two-dimensional feature information of the linear flexible body based on the RGB image; Obtain the three-dimensional point cloud data of the linear flexible body; Detect the three-dimensional feature information of the linear flexible body based on the three-dimensional point cloud data; Adopt a model-level fusion method to fuse the two-dimensional feature information and the three-dimensional feature information to identify the linear flexible body.
2. The linear flexible body composite detection method according to claim 1, wherein Detect the two-dimensional feature information of the linear flexible body based on the RGB image, specifically including: Preprocess the RGB image to obtain an enhanced image; Use the Canny edge detection method to perform edge detection on the enhanced image to obtain a Canny edge map; Based on the Canny edge map, use the region growing algorithm to extract connected regions to obtain a labeled image; Identify the area of the connected region based on the labeled image; According to the area of the connected region, use a clustering algorithm to divide the connected domain and determine the connected domain with the largest area; Draw an ellipse on the Canny edge map for the connected domain with the largest area; Fit the drawn ellipse and use an edge detection algorithm to obtain the contour of the linear flexible body; Determine the two-dimensional feature parameters of the linear flexible body based on the contour; the two-dimensional feature parameters are two-dimensional feature information.
3. The linear flexible body composite detection method according to claim 2, wherein, Preprocess the RGB image, specifically including: Use an improved Otsu's method to perform adaptive threshold segmentation on the RGB image to obtain a binary image; Use a bilateral filtering method to perform smoothing processing on the binary image to obtain a denoised image; Combine the Laplace operator and an adaptive sharpening method to perform enhancement processing on the denoised image to obtain an enhanced image.
4. The linear flexible body composite detection method according to claim 1, characterized in that, Detect the three-dimensional feature information of the linear flexible body based on the three-dimensional point cloud data, specifically including: Preprocess the three-dimensional point cloud data; Extract features from the preprocessed three-dimensional point cloud data to obtain the initial features of the linear flexible body; the initial features include geometric features, morphological features, and local directional features; Use convolution operations and max pooling to aggregate the initial features to obtain aggregated features; Perform fusion and dimensionality reduction processing on the aggregated features through a fully connected layer to obtain the three-dimensional feature parameters of the linear flexible body; the three-dimensional feature parameters are three-dimensional feature information.
5. The linear flexible body composite detection method according to claim 4, characterized in that Preprocess the three-dimensional point cloud data, specifically including: Use a statistical outlier removal algorithm to calculate the standard deviation of the distances of the neighborhood points of each point in the three-dimensional point cloud data, and remove the points with a standard deviation greater than a set threshold to obtain denoised three-dimensional point cloud data; Use the farthest point sampling method to perform uniform distribution sampling on the denoised three-dimensional point cloud data; Perform normalization processing on the uniformly distributed sampled three-dimensional point cloud data.
6. The linear flexible body composite detection method according to claim 1, wherein Adopt a model-level fusion method to fuse the two-dimensional feature information and the three-dimensional feature information to identify the linear flexible body, specifically including: Perform splicing and weighting operations on the two-dimensional feature information and the three-dimensional feature information to obtain a fused feature vector; Input the fused feature vector into a fully connected layer and a softmax activation function in sequence to identify the linear flexible body.
7. A linear flexible body composite detection system, characterized in that, Comprising: A CCD camera for obtaining the RGB image of the linear flexible body; A two-dimensional feature information detection module for detecting two-dimensional feature information of a linear flexible body based on the RGB image; A 3D camera for acquiring three-dimensional point cloud data of a linear flexible body; A three-dimensional feature information detection module for detecting three-dimensional feature information of a linear flexible body based on the three-dimensional point cloud data; A linear flexible body recognition module for fusing the two-dimensional feature information and the three-dimensional feature information by using a model-level fusion method to recognize the linear flexible body.
8. A computer device, comprising: A memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor executes the computer program to implement the linear flexible body composite detection method according to any one of claims 1-6.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the linear flexible body composite detection method according to any one of claims 1-6.
10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the linear flexible body composite detection method according to any one of claims 1-6.