A wiring harness bend radius measurement method, device, equipment and medium

Abstract

This application discloses a method, apparatus, device, and medium for measuring the bending radius of a wire harness, relating to the field of aircraft design technology, and is used to solve the problems of low efficiency and poor accuracy in the existing technology for measuring the bending radius of wire harnesses. The method includes: acquiring images of a target wire harness using a binocular structured light camera to obtain a two-dimensional image and corresponding three-dimensional point cloud data of the target; inputting the two-dimensional image and corresponding three-dimensional point cloud data of the target into a trained image segmentation model to obtain multiple contour points of the target wire harness; extracting feature points from the multiple contour points to obtain two feature points corresponding to each of the two wire harness contours on both sides of the target wire harness; calculating the bending radius corresponding to each of the two wire harness contours on both sides based on the two feature points corresponding to each of the two wire harness contours on both sides; and calculating the bending radius of the target wire harness based on the bending radius corresponding to each of the two wire harness contours on both sides, so as to measure the bending radius of the wire harness more accurately and efficiently.

Description

Technical Field

[0001] This application relates to the field of aircraft manufacturing technology, and provides a method, apparatus, equipment and medium for measuring the bending radius of a wire harness. Background Technology

[0002] As is well known, in the quality control of aircraft wiring harnesses, if the bending radius of the wiring harness is too small, it will not only fail to meet design requirements, but may also lead to wire harness breakage, thereby causing aircraft safety risks. Therefore, accurately measuring the bending radius of the wiring harness is a crucial step in ensuring overall product quality. In traditional methods, the bending radius of the wiring harness is often measured by "gauge comparison." However, this measurement method has problems such as long operation time, low efficiency, low accuracy, and susceptibility to errors.

[0003] Therefore, how to measure the bending radius of wire harnesses more accurately and efficiently has become an urgent problem to be solved. Summary of the Invention

[0004] This application provides a method, apparatus, device, and medium for measuring the bending radius of a wire harness, which addresses the problems of low efficiency and poor accuracy in measuring the bending radius of wire harnesses in existing technologies.

[0005] On the one hand, a method for measuring the bending radius of a wire harness is provided, the method comprising:

[0006] A binocular structured light camera is used to acquire images of the target wire bundle, obtaining two-dimensional images and corresponding three-dimensional point cloud data of the target;

[0007] The two-dimensional image and the corresponding target three-dimensional point cloud data are input into a trained image segmentation model to obtain multiple contour points of the target line bundle; wherein, the trained image segmentation model is generated based on an instance segmentation model of a convolutional neural network;

[0008] Feature point extraction is performed on the multiple contour points to obtain two feature points corresponding to the contours of the two sides of the target line bundle.

[0009] Calculate the bending radius of each of the two feature points corresponding to the outlines of the two wire harnesses on each side;

[0010] The bending radius of the target wire harness is calculated based on the bending radius corresponding to the respective contours of the wire harnesses on both sides.

[0011] Optionally, before acquiring images of the target line bundle using a binocular structured light camera to obtain two-dimensional images and corresponding target three-dimensional point cloud data, the method further includes:

[0012] The binocular structured light camera is used to acquire images of the wire bundle to be detected from different angles and distances, obtaining multiple acquired images;

[0013] The multiple acquired images are imported into the segmentation and annotation software in batches, and the line bundle contours in the multiple acquired images are annotated to obtain multiple annotated acquired images.

[0014] The initial image segmentation model is trained using the multiple labeled acquired images to obtain a trained image segmentation model.

[0015] Optionally, the step of acquiring images of the target line bundle using a binocular structured light camera to obtain a two-dimensional image and corresponding target three-dimensional point cloud data includes:

[0016] A binocular structured light camera is used to acquire images of the target wire bundle, thereby obtaining the two-dimensional image and the corresponding initial three-dimensional point cloud data.

[0017] The initial 3D point cloud data is filtered to obtain the target 3D point cloud data;

[0018] The two-dimensional image is matched and aligned with the target three-dimensional point cloud data to obtain the matched two-dimensional image;

[0019] The matched 2D image and the target 3D point cloud data are stored.

[0020] Optionally, the step of storing the matched two-dimensional image and the target three-dimensional point cloud data includes:

[0021] The matched 2D image is stored in its original size and PNG format;

[0022] The target 3D point cloud data is stored in a preset data structure in xyz format.

[0023] Optionally, the step of extracting feature points from the plurality of contour points to obtain two feature points corresponding to the contours of each of the two sides of the target line bundle includes:

[0024] For any two adjacent contour points, determine the direction of the contour point vector corresponding to the two adjacent contour points;

[0025] For any two adjacent contour point vectors, determine the angular change of the two adjacent contour point vectors.

[0026] The K-means clustering method is used to perform cluster analysis on the multiple contour points to determine multiple first contour point groups; wherein, the first contour point group includes the contour points of the two side contours of the target wire bundle and other points.

[0027] Anomaly removal is performed on the plurality of first contour point groups to obtain a plurality of second contour point groups;

[0028] Based on the angular change of any two adjacent contour point vectors, feature points are extracted from the multiple second contour points in groups to obtain two feature points corresponding to the contours of the two sides of the target line bundle; wherein, the feature points are endpoints or inflection points.

[0029] Optionally, the step of calculating the bending radius corresponding to each of the two feature points corresponding to the contours of the two side wire harnesses includes:

[0030] Based on the two feature points corresponding to the outlines of the two wire harnesses on each side, the key points of the outlines of the two wire harnesses on both sides are extracted in sequence to determine the first wire harness outline and the second wire harness outline corresponding to the target wire harness; wherein, the first wire harness outline and the second wire harness outline are located on both sides of the target wire harness respectively.

[0031] The bending radius of the first segment of the wire harness profile is calculated using curve fitting or curvature calculation methods.

[0032] The bending radius of the second segment of the wire harness profile is calculated using curve fitting or curvature calculation methods.

[0033] Optionally, the step of calculating the bending radius of the target wire harness based on the bending radii corresponding to the respective contours of the two wire harnesses includes:

[0034] The bending radius of the target wire harness is calculated by averaging the bending radii of the first wire harness profile and the second wire harness profile.

[0035] On the one hand, a wire harness bending radius measuring device is provided, the device comprising:

[0036] The point cloud data acquisition unit is used to acquire images of the target line bundle using a binocular structured light camera, and obtain two-dimensional images and corresponding target three-dimensional point cloud data.

[0037] The contour point acquisition unit is used to input the two-dimensional image and the corresponding target three-dimensional point cloud data into a trained image segmentation model to obtain multiple contour points of the target line bundle; wherein, the trained image segmentation model is generated based on an instance segmentation model of a convolutional neural network;

[0038] The feature point acquisition unit is used to extract feature points from the plurality of contour points to obtain two feature points corresponding to the contours of the two sides of the target line bundle.

[0039] The first bending radius calculation unit is used to calculate the bending radius of each of the two side wire harness contours based on two feature points corresponding to each of the two side wire harness contours.

[0040] The second bending radius calculation unit is used to calculate the bending radius of the target wire harness based on the bending radii corresponding to the respective contours of the wire harnesses on both sides.

[0041] On one hand, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the methods described above.

[0042] On the one hand, a computer storage medium is provided that stores computer program instructions thereon, which, when executed by a processor, implement any of the methods described above.

[0043] Compared with the prior art, the beneficial effects of this application are as follows:

[0044] In this embodiment, when measuring the bending radius of the wire harness, firstly, a binocular structured light camera can be used to acquire images of the target wire harness to obtain a two-dimensional image and corresponding target three-dimensional point cloud data; then, the two-dimensional image and corresponding target three-dimensional point cloud data can be input into a trained image segmentation model to obtain multiple contour points of the target wire harness; wherein, the trained image segmentation model is generated based on an instance segmentation model of a convolutional neural network; next, feature points can be extracted from the multiple contour points to obtain two feature points corresponding to the contours of both sides of the target wire harness; then, the bending radius corresponding to each of the two contours of both sides of the wire harness is calculated based on the two feature points corresponding to each of the two contours of both sides of the wire harness; finally, the bending radius of the target wire harness can be calculated based on the bending radius corresponding to each of the two contours of both sides of the wire harness. Therefore, in this embodiment, by using a binocular structured light camera to acquire images of the target wire bundle to obtain two-dimensional images and corresponding three-dimensional point cloud data of the target, and by using a trained image segmentation model to identify and extract the wire bundle contour points to calculate the bending radius of the target wire bundle, this application overcomes the limitations of the traditional "gauge comparison" method by combining binocular vision and structured light technology, and achieves a more accurate and efficient measurement of the bending radius of the wire bundle. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0046] Figure 1 This is a schematic diagram of an application scenario provided by an embodiment of this application;

[0047] Figure 2A schematic flowchart of a wire harness bending radius measurement method provided in an embodiment of this application;

[0048] Figure 3 A schematic diagram illustrating a connection between a cabin installation wiring harness and a binocular structured light camera provided in an embodiment of this application;

[0049] Figure 4 This is a detection schematic diagram of an image segmentation model provided in an embodiment of this application;

[0050] Figure 5 A schematic diagram illustrating the direction angle value of the contour point vector provided in an embodiment of this application;

[0051] Figure 6 A schematic diagram illustrating the change in the direction angle of the contour point vector provided in an embodiment of this application;

[0052] Figure 7 A schematic diagram illustrating the extraction of feature points from the installation harness provided in an embodiment of this application;

[0053] Figure 8 This is a schematic diagram of a wire harness bending radius measuring device provided in an embodiment of this application.

[0054] The diagram is labeled as follows: 10-Wire harness bending radius measuring device, 101-Processor, 102-Memory, 103-I / O interface, 104-Database, 80-Wire harness bending radius measuring device, 801-Point cloud data acquisition unit, 802-Contour point acquisition unit, 803-Feature point acquisition unit, 804-First bending radius calculation unit, 805-Second bending radius calculation unit, 806-Model training unit. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. Unless otherwise specified, the embodiments and features in the embodiments of this application can be arbitrarily combined with each other. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than that shown here.

[0056] As is well known, in the quality control of aircraft wiring harnesses, if the bending radius of the wiring harness is too small, it will not only fail to meet design requirements, but may also lead to wire harness breakage, thereby causing aircraft safety risks. Therefore, accurately measuring the bending radius of the wiring harness is a crucial step in ensuring overall product quality. In traditional methods, the bending radius of the wiring harness is often measured by "gauge comparison." However, this measurement method has problems such as long operation time, low efficiency, low accuracy, and susceptibility to errors.

[0057] Based on this, this application provides a method for measuring the bending radius of a wire harness. In this method, firstly, a binocular structured light camera is used to acquire images of the target wire harness to obtain a two-dimensional image and corresponding three-dimensional point cloud data. Then, the two-dimensional image and corresponding three-dimensional point cloud data are input into a trained image segmentation model to obtain multiple contour points of the target wire harness. The trained image segmentation model is generated based on an instance segmentation model using a convolutional neural network. Next, feature points are extracted from the multiple contour points to obtain two feature points corresponding to each of the two wire harness contours on both sides of the target wire harness. Then, the bending radius corresponding to each of the two wire harness contours on both sides is calculated based on these two feature points. Finally, the bending radius of the target wire harness is calculated based on the bending radius corresponding to each of the two wire harness contours on both sides. Therefore, in this embodiment, by using a binocular structured light camera to acquire images of the target wire bundle to obtain two-dimensional images and corresponding three-dimensional point cloud data of the target, and by using a trained image segmentation model to identify and extract the wire bundle contour points to calculate the bending radius of the target wire bundle, this application overcomes the limitations of the traditional "gauge comparison" method by combining binocular vision and structured light technology, and achieves a more accurate and efficient measurement of the bending radius of the wire bundle.

[0058] After introducing the design concept of the embodiments of this application, the following is a brief introduction to the application scenarios to which the technical solutions of the embodiments of this application can be applied. It should be noted that the application scenarios described below are only for illustrating the embodiments of this application and are not intended to limit the scope. In specific implementation, the technical solutions provided by the embodiments of this application can be flexibly applied according to actual needs.

[0059] like Figure 1 The diagram shown illustrates an application scenario provided by an embodiment of this application. This application scenario may include a wire harness bending radius measuring device 10.

[0060] The wire harness bending radius measuring device 10 can be used to measure the bending radius of a wire harness, and can be, for example, a personal computer (PC), server, or laptop. The wire harness bending radius measuring device 10 may include one or more processors 101, memory 102, I / O interfaces 103, and databases 104. Specifically, the processor 101 may be a central processing unit (CPU) or a digital processing unit, etc. The memory 102 may be volatile memory, such as random-access memory (RAM); the memory 102 may also be non-volatile memory, such as read-only memory, flash memory, hard disk drive (HDD), or solid-state drive (SSD); or the memory 102 may be any other medium capable of carrying or storing desired program code in the form of instructions or data structures that can be accessed by a computer, but is not limited thereto. The memory 102 may be a combination of the above-mentioned memories. The memory 102 can store some program instructions for the wire harness bending radius measurement method provided in this application embodiment. When these program instructions are executed by the processor 101, they can be used to implement the steps of the wire harness bending radius measurement method provided in this application embodiment, so as to solve the problems of low efficiency and poor accuracy in measuring the bending radius of wire harnesses in the prior art. The database 104 can be used to store data such as two-dimensional images, target three-dimensional point cloud data, trained image segmentation models, multiple contour points of the target wire harness, bending radii corresponding to the contours of the wire harnesses on both sides, bending radius of the target wire harness, and initial three-dimensional point cloud data involved in the solution provided in this application embodiment.

[0061] In this embodiment, the wire harness bending radius measuring device 10 can acquire a two-dimensional image and the corresponding target three-dimensional point cloud data through the I / O interface 103. Then, the processor 101 of the wire harness bending radius measuring device 10 will measure the wire harness bending radius more accurately and efficiently according to the program instructions of the wire harness bending radius measuring method provided in this embodiment of the application stored in the memory 102. In addition, the two-dimensional image, target three-dimensional point cloud data, trained image segmentation model, multiple contour points of the target wire harness, bending radii corresponding to the contours of the two wire harnesses on both sides, bending radius of the target wire harness, and initial three-dimensional point cloud data can be stored in the database 104.

[0062] Of course, the methods provided in the embodiments of this application are not limited to... Figure 1The application scenarios shown can also be used in other possible scenarios, and this application embodiment does not impose any limitations. Figure 1 The functions that the various devices in the application scenarios shown can achieve will be described in subsequent method embodiments, and will not be elaborated on here. Below, the methods of the embodiments of this application will be described in conjunction with the accompanying drawings.

[0063] like Figure 2 The diagram shown is a flowchart illustrating a wire harness bending radius measurement method provided in an embodiment of this application. This method can... Figure 1 The wire harness bending radius measuring device 10 is used to perform the measurement. Specifically, the process of this method is described as follows.

[0064] Step 201: Use a binocular structured light camera to acquire images of the target wire bundle, obtaining two-dimensional images and corresponding three-dimensional point cloud data of the target.

[0065] In the embodiments of this application, such as Figure 3 The diagram shown illustrates a connection between the cabin installation wiring harness and the binocular structured light camera provided in this embodiment. The binocular structured light camera is mounted on a bracket. In practical applications, the binocular structured light camera can be configured to connect to a control processing unit (e.g., a personal computer, server, etc.) to capture images of the cabin installation wiring harness. Specifically, the control program of the control processing unit can control the image capture, parameters, and storage results of the binocular structured light camera.

[0066] Therefore, when measuring the bending radius of the wiring harness, firstly, a binocular structured light camera can be used to scan the wiring harness area within the compartment to acquire images of the target wiring harness, obtaining a two-dimensional image and corresponding initial three-dimensional point cloud data. During image acquisition, it is necessary to ensure that the angle and position of the binocular structured light camera can cover the entire wiring harness area. Then, the initial three-dimensional point cloud data can be processed and filtered to obtain the target three-dimensional point cloud data, removing any noise or outliers to ensure the accuracy of subsequent data analysis. Next, the two-dimensional image and the target three-dimensional point cloud data can be matched and aligned to ensure a one-to-one correspondence, thus obtaining a matched two-dimensional image. Finally, a corresponding data structure can be created to ensure that the matched two-dimensional image and the target three-dimensional point cloud data can be correctly called and processed in subsequent steps.

[0067] Finally, the matched 2D image and the target 3D point cloud data can be stored. Specifically, the matched 2D image can be stored in its original size and PNG format; the target 3D point cloud data can be stored in a preset data structure in XYZ format.

[0068] Step 202: Input the two-dimensional image and the corresponding target three-dimensional point cloud data into the trained image segmentation model to obtain multiple contour points of the target line bundle.

[0069] In the embodiments of this application, the trained image segmentation model is generated based on an instance segmentation model of a convolutional neural network (e.g., a Yolov8 segmentation model).

[0070] Specifically, first, the matched 2D image and corresponding target 3D point cloud data can be read from the memory to prepare for subsequent feature extraction. Then, the previously trained image segmentation model can be read from the memory; this trained image segmentation model is stored as a best.pt file. Next, the 2D image and corresponding target 3D point cloud data can be input into the trained image segmentation model to identify the target line bundle and its contour, thereby obtaining multiple contour points of the target line bundle, such as... Figure 4 The diagram shown is a detection schematic of an image segmentation model provided in an embodiment of this application. After inputting a two-dimensional image and the corresponding target three-dimensional point cloud data into the trained image segmentation model, the trained image segmentation model can identify the target line bundle and its contour in the two-dimensional image through the recognition box.

[0071] Step 203: Extract feature points from multiple contour points to obtain two feature points corresponding to the contours of the two sides of the target wire bundle.

[0072] Specifically, firstly, for any two adjacent contour points, the direction of the contour point vectors corresponding to the two adjacent contour points can be determined (expressed as angle values), thus determining the direction and orientation of the line bundle contour, such as... Figure 5 The diagram shown is a schematic representation of the direction angle value of the contour point vector provided in an embodiment of this application.

[0073] Then, for any two adjacent contour point vectors, the angular change between them is determined. This provides data for subsequent analysis to filter feature values ​​based on angular changes. Figure 6 The diagram shown is a schematic representation of the change in the direction angle of the contour point vector provided in an embodiment of this application.

[0074] Next, the K-means clustering method can be used to perform cluster analysis on multiple contour points to divide these multiple contour points into groups of contour points of the two side wire bundle contours and other points. That is, multiple first contour point groups can be determined. For example, they can be divided into 3 first contour point groups. Among them, the first contour point group includes the contour points of the two side wire bundle contours of the target wire bundle and other points.

[0075] Then, multiple first contour points can be grouped together to remove outliers, thereby obtaining multiple second contour point groups, thus ensuring that the obtained contour points can accurately represent the contour features on both sides of the target wire bundle.

[0076] Finally, based on cluster analysis, multiple second contour points can be grouped and feature points extracted according to the angular change of any two adjacent contour point vectors. This further extracts feature points of the contour points on both sides of the target bundle, thereby obtaining two feature points corresponding to each side of the bundle contour, i.e., the inflection points of the two side contours of the bundle to the diameter edge, resulting in a total of four feature points extracted. Figure 7 The diagram shown is a schematic representation of extracting feature points of the wiring harness according to an embodiment of this application. The feature points are endpoints or inflection points. Therefore, by determining the accuracy of the feature points, an important data foundation is provided for the subsequent calculation of the bending radius of the target wiring harness, enabling accurate calculation of the bending radius of the target wiring harness.

[0077] Step 204: Calculate the bending radius of each side of the wire harness profile based on the two feature points corresponding to each side of the wire harness profile.

[0078] Specifically, firstly, based on the two feature points corresponding to the contours of the two wire harnesses on each side, the key points of the contours of the two wire harnesses on both sides can be extracted in sequence to determine the first and second wire harness contours corresponding to the target wire harness, so as to calculate the bending radius of the target wire harness in the future; wherein, the first and second wire harness contours are located on both sides of the target wire harness.

[0079] Then, curve fitting or curvature calculation methods can be used to calculate the bending radius of the first wire harness profile. For example, the bending radius of the first wire harness profile can be calculated by "fitting a plane circle". The specific calculation process is as follows:

[0080] Assume the set of contour points of the first line harness is {pt1, pt2, ..., pt...} n}, the i-th contour point pt i It can be represented as (x i ,y i ,z i Based on this, by minimizing the objective function f as shown in formula (1), the best fitting plane parameters (a,b,c,d) of the wire harness contour points can be obtained, and the normal vector of the plane is n=(a,b,c).

[0081]

[0082] Then, based on the best-fit plane of the wire harness contour points, the set of projection points {p1, p2, ..., p} on that plane can be calculated. n}. Where the i-th projection point p i It can be obtained through the following formula (2).

[0083]

[0084] Next, the parameters of the plane circle can be calculated based on the projection points on the plane. Specifically, the parameters of the plane circle can be calculated by minimizing the objective function f as shown in the following formula (3):

[0085]

[0086] Here, params are the optimization parameters, which include the center coordinates c and the radius r of the circle, n·p. i Point p i The projection onto the normal vector n, where d represents the offset of the plane, i.e., d = -n·p i .

[0087] Furthermore, in this embodiment of the application, the radius and center position of the circle can be fitted by the objective function in the above formula (3). Then, the objective function is minimized by adjusting the center coordinates and radius, thereby obtaining the best-fit circle radius (i.e., the bending radius of the first segment of the wire harness profile).

[0088] Similarly, curve fitting or curvature calculation methods can be used to calculate the bending radius of the second wire harness profile. For example, the bending radius of the second wire harness profile can also be calculated by "fitting a plane circle".

[0089] Step 205: Calculate the bending radius of the target wire harness based on the bending radii corresponding to the contours of the wire harnesses on both sides.

[0090] Specifically, after calculating the bending radius of each of the two wire harness contours, that is, after calculating the bending radius of the first wire harness contour and the second wire harness contour, the bending radius of the target wire harness can be calculated by averaging the bending radius of the first wire harness contour and the second wire harness contour.

[0091] Finally, the calculated bending radius of the target wire harness can be output as the result, which may include the recorded bending radius value and related information.

[0092] Furthermore, since this application overcomes the limitations of the traditional "gauge comparison" method by combining binocular vision and structured light technology, it can achieve more accurate and efficient measurement of the bending radius of the wire harness.

[0093] In one possible implementation, in order to improve the segmentation accuracy of the image segmentation model, in this embodiment of the application, before using a binocular structured light camera to acquire images of the target line bundle and obtain two-dimensional images and corresponding target three-dimensional point cloud data, the image segmentation model can also be trained.

[0094] Specifically, firstly, in order to ensure the normal operation of devices such as binocular structured light cameras (or industrial cameras), parameters of the binocular structured light camera can be set. For example, parameters such as exposure time, resolution, and frame rate can be set to ensure accurate imaging of the wire harness area.

[0095] Then, to improve the segmentation accuracy of the image segmentation model, a binocular structured light camera can be used to acquire images of the wire bundle to be detected from different angles and distances, obtaining multiple images. For example, a total of 104 images can be acquired, which can include various parts of the wire bundle. During the acquisition process, it is necessary to ensure that the wire bundle within the detection area is placed in a clearly visible position, without any objects obstructing or hindering the view. Furthermore, image quality can be checked to eliminate problems such as blurriness, overexposure, or underexposure, and any abnormal images can be recorded for subsequent processing.

[0096] Next, multiple acquired images can be imported in batches into segmentation and annotation software (e.g., the segmentation and annotation tool Labelme) to annotate the line bundle contours in the multiple acquired images, and ensure that the boundaries and details of each line bundle are accurately marked, thereby obtaining multiple annotated acquired images.

[0097] Then, multiple labeled collected images can be saved, that is, the labeling results can be saved as a dataset in the form of a file for subsequent analysis and model training. The labeled collected images are saved in JSON file format.

[0098] Next, we can prepare the labeled image dataset for training, that is, prepare these labeled acquired images and perform necessary data cleaning and preprocessing on these labeled acquired images to improve the model training effect. Then, we can determine the specific algorithm, network structure and training parameters of the training model, making full preparations for training the image segmentation model.

[0099] Finally, the initial image segmentation model was trained using multiple labeled acquired images after data cleaning and preprocessing to obtain a trained image segmentation model. This trained image segmentation model was stored in the format of a best.pt file.

[0100] In summary, by using a binocular structured light camera to acquire images of the target wire bundle to obtain two-dimensional images and corresponding three-dimensional point cloud data of the target, and by using a trained image segmentation model to identify and extract the wire bundle contour points to calculate the bending radius of the target wire bundle, this application overcomes the limitations of the traditional "gauge comparison" method by combining binocular vision and structured light technology, and achieves a more accurate and efficient measurement of the bending radius of the wire bundle.

[0101] Furthermore, addressing the problems of time-consuming operation, low efficiency, low accuracy, and error susceptibility inherent in traditional "gauge comparison" methods for aircraft, this application not only enables automated and precise measurement of wiring harnesses but also allows for real-time monitoring of the wiring harness manufacturing process. This facilitates the identification of production problems and timely adjustments, improving production efficiency and wiring harness assembly quality. Compared to traditional "gauge comparison" methods, the accuracy reaches 0.5 mm from the centimeter level, and the detection time is reduced to within 10 seconds (verification results), demonstrating higher accuracy and efficiency.

[0102] Furthermore, this application proposes using a model to identify wire harness contour points. The model is trained using deep learning and other technologies, enabling it to intelligently identify and extract wire harness contours. By combining two-dimensional images and three-dimensional point cloud data, the contour features of the wire harness are comprehensively extracted and analyzed. This multi-dimensional data fusion method can more accurately capture the shape and detail information of the wire harness. The technical solution of this application offers rapid detection with high accuracy and reliability. It eliminates the need for actual contact with the wire harness, avoiding potential damage or deformation. It is not only suitable for measuring the bending radius of wire harnesses but can also be applied to other fields requiring shape measurement and feature extraction, such as parts processing, assembly, and quality control, providing new possibilities and prospects for the development of related fields.

[0103] Based on the same inventive concept, embodiments of this application provide a wire harness bending radius measuring device 80, such as... Figure 8 As shown, the wire harness bending radius measuring device 80 includes:

[0104] The point cloud data acquisition unit 801 is used to acquire images of the target line bundle using a binocular structured light camera, and obtain two-dimensional images and corresponding target three-dimensional point cloud data.

[0105] The contour point acquisition unit 802 is used to input the two-dimensional image and the corresponding target three-dimensional point cloud data into the trained image segmentation model to obtain multiple contour points of the target line bundle; wherein, the trained image segmentation model is generated based on the instance segmentation model of the convolutional neural network;

[0106] The feature point acquisition unit 803 is used to extract feature points from multiple contour points and obtain two feature points corresponding to the contours of the two sides of the target line bundle.

[0107] The first bending radius calculation unit 804 is used to calculate the bending radius of each of the two wire harness contours based on the two feature points corresponding to each of the two wire harness contours.

[0108] The second bending radius calculation unit 805 is used to calculate the bending radius of the target wire harness based on the bending radii corresponding to the respective wire harness contours on both sides.

[0109] Optionally, the wire harness bending radius measuring device 80 also includes a model training unit 806, used for:

[0110] A binocular structured light camera was used to acquire images of the wire bundle to be inspected from different angles and distances, resulting in multiple acquired images;

[0111] Multiple acquired images are imported into the segmentation and annotation software in batches, and the line bundle contours in the multiple acquired images are annotated to obtain multiple annotated acquired images;

[0112] The initial image segmentation model is trained using multiple labeled acquired images to obtain a trained image segmentation model.

[0113] Optionally, the point cloud data acquisition unit 801 is also used for:

[0114] A binocular structured light camera is used to acquire images of the target wire bundle, obtaining two-dimensional images and corresponding initial three-dimensional point cloud data;

[0115] The initial 3D point cloud data is filtered to obtain the target 3D point cloud data;

[0116] The two-dimensional image is matched and aligned with the target three-dimensional point cloud data to obtain the matched two-dimensional image;

[0117] The matched 2D image and the target 3D point cloud data are stored.

[0118] Optionally, the point cloud data acquisition unit 801 is also used for:

[0119] The matched 2D image is stored in its original size and PNG format;

[0120] The target 3D point cloud data is stored in a preset data structure in xyz format.

[0121] Optionally, the feature point acquisition unit 803 is also used for:

[0122] For any two adjacent contour points, determine the direction of the contour point vector corresponding to the two adjacent contour points;

[0123] For any two adjacent contour point vectors, determine the angular change between the two adjacent contour point vectors.

[0124] K-means clustering was used to perform cluster analysis on multiple contour points to determine multiple first contour point groups; wherein, the first contour point group includes the contour points of the two sides of the target wire bundle and other points.

[0125] Multiple first contour points are grouped together to remove outliers, and multiple second contour point groups are obtained.

[0126] Based on the angular change of any two adjacent contour point vectors, multiple second contour points are grouped and feature points are extracted to obtain two feature points corresponding to the contours of the two sides of the target line bundle; where the feature points are endpoints or inflection points.

[0127] Optionally, the first bending radius calculation unit 804 is also used for:

[0128] Based on the two feature points corresponding to the contours of the two wire harnesses on each side, the key points of the contours of the two wire harnesses on both sides are extracted in sequence to determine the first and second wire harness contours corresponding to the target wire harness; wherein, the first and second wire harness contours are located on both sides of the target wire harness.

[0129] The bending radius of the first segment of the wire harness profile is calculated using curve fitting or curvature calculation methods.

[0130] The bending radius of the second segment of the wire harness profile is calculated using curve fitting or curvature calculation methods.

[0131] Optionally, the second bending radius calculation unit 805 is also used for:

[0132] The bending radius of the target wire harness is calculated by averaging the bending radii of the first and second wire harness profiles.

[0133] The wire harness bending radius measuring device 80 can be used to perform... Figures 2-7 The method performed in the illustrated embodiment is described above. Therefore, the functions that each functional module of the wire harness bending radius measuring device 80 can achieve can be referred to. Figures 2-7 The embodiments shown are described in detail below.

[0134] In some possible implementations, various aspects of the methods provided in this application can also be implemented as a program component comprising program code that, when run on a computer device, causes the computer device to perform the steps of the methods according to the various exemplary embodiments of this application described above. For example, the computer device may perform actions such as... Figures 2-7 The method performed in the illustrated embodiment.

[0135] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks. Alternatively, if the integrated unit of the present invention is implemented as a software functional module and sold or used as an independent part, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software part. This computer software part is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

[0136] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0137] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for measuring the bending radius of a wire harness, characterized in that, The method includes: A binocular structured light camera is used to acquire images of the target wire bundle, obtaining two-dimensional images and corresponding three-dimensional point cloud data of the target; The two-dimensional image and the corresponding target three-dimensional point cloud data are input into a trained image segmentation model to obtain multiple contour points of the target line bundle; wherein, the trained image segmentation model is generated based on an instance segmentation model of a convolutional neural network; For any two adjacent contour points, determine the direction of the contour point vectors corresponding to the two adjacent contour points; for any two adjacent contour point vectors, determine the angular change of the two adjacent contour point vectors; use K-means clustering to perform cluster analysis on the multiple contour points to determine multiple first contour point groups; remove outliers from the multiple first contour point groups to obtain multiple second contour point groups; extract feature points from the multiple second contour point groups based on the angular change of all two adjacent contour point vectors to obtain two feature points corresponding to the contours of the two sides of the target line bundle; wherein, the first contour point group includes the contour points of the contours of the two sides of the target line bundle and other points; the feature points are endpoints or inflection points. Calculate the bending radius of each of the two feature points corresponding to the outlines of the two wire harnesses on each side; The bending radius of the target wire harness is calculated based on the bending radius corresponding to the respective contours of the wire harnesses on both sides.

2. The method as described in claim 1, characterized in that, Before acquiring images of the target line bundle using a binocular structured light camera to obtain two-dimensional images and corresponding target three-dimensional point cloud data, the method further includes: The binocular structured light camera is used to acquire images of the wire bundle to be detected from different angles and distances, obtaining multiple acquired images; The multiple acquired images are imported into the segmentation and annotation software in batches, and the line bundle contours in the multiple acquired images are annotated to obtain multiple annotated acquired images. The initial image segmentation model is trained using the multiple labeled acquired images to obtain a trained image segmentation model.

3. The method as described in claim 1, characterized in that, The step of acquiring images of the target line bundle using a binocular structured light camera to obtain a two-dimensional image and the corresponding three-dimensional point cloud data of the target includes: A binocular structured light camera is used to acquire images of the target wire bundle, thereby obtaining the two-dimensional image and the corresponding initial three-dimensional point cloud data. The initial 3D point cloud data is filtered to obtain the target 3D point cloud data; The two-dimensional image is matched and aligned with the target three-dimensional point cloud data to obtain the matched two-dimensional image; The matched 2D image and the target 3D point cloud data are stored.

4. The method as described in claim 3, characterized in that, The step of storing the matched two-dimensional image and the target three-dimensional point cloud data includes: The matched 2D image is stored in its original size and PNG format; The target 3D point cloud data is stored in a preset data structure in xyz format.

5. The method as described in claim 1, characterized in that, The step of calculating the bending radius corresponding to each of the two feature points corresponding to the contours of the two side wire harnesses includes: Based on the two feature points corresponding to the outlines of the two wire harnesses on each side, the key points of the outlines of the two wire harnesses on both sides are extracted in sequence to determine the first wire harness outline and the second wire harness outline corresponding to the target wire harness; wherein, the first wire harness outline and the second wire harness outline are located on both sides of the target wire harness respectively. The bending radius of the first segment of the wire harness profile is calculated using curve fitting or curvature calculation methods. The bending radius of the second segment of the wire harness profile is calculated using curve fitting or curvature calculation methods.

6. The method as described in claim 5, characterized in that, The step of calculating the bending radius of the target wire harness based on the bending radii corresponding to the respective contours of the two wire harnesses includes: The bending radius of the target wire harness is calculated by averaging the bending radii of the first wire harness profile and the second wire harness profile.

7. A device for measuring the bending radius of a wire harness, characterized in that, The device includes: The point cloud data acquisition unit is used to acquire images of the target line bundle using a binocular structured light camera, and obtain two-dimensional images and corresponding target three-dimensional point cloud data. The contour point acquisition unit is used to input the two-dimensional image and the corresponding target three-dimensional point cloud data into a trained image segmentation model to obtain multiple contour points of the target line bundle; wherein, the trained image segmentation model is generated based on an instance segmentation model of a convolutional neural network; A feature point acquisition unit is used to: determine the direction of the contour point vector corresponding to any two adjacent contour points; determine the angular change of any two adjacent contour point vectors; perform cluster analysis on the multiple contour points using the K-means clustering method to determine multiple first contour point groups; remove outliers from the multiple first contour point groups to obtain multiple second contour point groups; extract feature points from the multiple second contour point groups based on the angular change of all any two adjacent contour point vectors to obtain two feature points corresponding to the contours of each side of the target line bundle; wherein, the first contour point group includes the contour points of the contours of the two side line bundles of the target line bundle and other points; the feature points are endpoints or inflection points. The first bending radius calculation unit is used to calculate the bending radius of each of the two side wire harness contours based on two feature points corresponding to each of the two side wire harness contours. The second bending radius calculation unit is used to calculate the bending radius of the target wire harness based on the bending radii corresponding to the respective contours of the wire harnesses on both sides.

8. An electronic device, characterized in that, The device includes: Memory, used to store program instructions; A processor is configured to invoke program instructions stored in the memory and execute the method described in any one of claims 1-6 according to the obtained program instructions.

9. A storage medium, characterized in that, The storage medium stores computer-executable instructions for causing a computer to perform the method described in any one of claims 1-6.

Images