A battery cover plate feature measurement method and device based on line laser scanning and a storage medium
Patent Information
- Application Number
- CN202410185467.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-02-19
- Publication Date
- 2026-10-30
- Estimated Expiration
- 2044-02-19
AI Technical Summary
[0017]本发明利用线激光扫描点云的特性，更为快速高效地提取出电池盖板工件点云的特征点云，解决了复杂工件参数测量难的问题；
Smart Images
Figure CN118196458B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of 3D point cloud measurement technology, and in particular to a method, apparatus and storage medium for measuring the features of a battery cover plate based on line laser scanning. Background Technology
[0002] The battery cover plays a crucial role in protecting the stable operation of the battery module and effectively reducing overcurrent. The battery cover is mainly made of steel or aluminum alloy and serves as a protective shell for the battery pack. Its main structure includes an upper aluminum plate, which is equipped with sealing rings, positive and negative terminals, explosion-proof valves, and other components. During the processing of the battery cover, the surface quality directly affects the overall quality and reliability of the product.
[0003] For example, if parameters such as the flatness of the positive and negative terminals and the aluminum base plate, the size of the terminal welding pad, and the height of the battery terminals in the battery cover do not meet the standards, it may lead to the battery cover failing to meet design requirements or even producing defects during welding and other processes. If a defective battery cover is used to protect the battery pack, it will lead to a series of problems such as increased scrap rate. In modern industry, measurement technology is mainly divided into two-dimensional measurement and three-dimensional measurement. Two-dimensional measurement cannot solve the problem of measuring the three-dimensional dimensions of the cover. Traditional workpiece inspection uses a coordinate measuring machine, but this contact inspection method is costly, inefficient, and prone to damaging the workpiece being measured. Therefore, there is an urgent need for a battery cover feature measurement method that can accurately inspect the battery cover, enabling manufacturers to measure the parameters of the battery cover more accurately, ensuring that the parameters of each part of the battery cover meet the standards, and effectively reducing the risk of battery pack damage in industrial production. Summary of the Invention
[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method, device and storage medium for measuring the features of battery cover based on line laser scanning, so that manufacturers can measure the parameters of battery cover more accurately, and make the parameters of each part of the battery cover meet the standards, effectively reducing the risk of battery pack damage in industrial production.
[0005] To achieve the above objectives, the present invention is implemented using the following technical solution:
[0006] In a first aspect, the present invention provides a method for measuring the features of a battery cover plate based on line laser scanning, comprising: The battery cover workpiece is scanned by a line laser 3D scanner to extract the first target point cloud data of the battery cover workpiece. The first target point cloud data is filtered and preprocessed to obtain the second point cloud data; The point cloud features of the second point cloud data are segmented using an improved Euclidean clustering algorithm, including: Construct a kd-tree topology based on the second point cloud data, calculate feature parameters for each point in the second point cloud data, set an empty clustering list D, and a point cloud queue Q that needs to be checked; Take any point from the second point cloud data Place it into the point cloud queue Q that needs to be inspected, so that... Centered on, with Find the radius neighborhood element set ,examine Characteristic parameters of each element Whether it is less than the threshold, Elements whose feature parameters are less than a threshold are added to the point cloud queue Q, while elements whose feature parameters are greater than a threshold are not added to the point cloud queue Q. For each newly added element in the point cloud queue Q, take a neighborhood with the same radius and check whether the feature parameters of all elements in the neighborhood are less than the threshold and whether they have been processed. Add the elements that are less than the threshold and have not been processed back to the point cloud queue Q. When all points in the point cloud queue Q have been processed and no more elements are added, all points are placed into the cluster set D and then the point cloud queue Q is cleared. When all points in the second point cloud data have been processed and are located in the various clusters formed by the cluster set D, the segmentation ends. Any point in the second cloud data mentioned above Feature parameter definition: ; Where M is a spherical neighborhood of radius R, and N is the total number of data points in the M neighborhood. Within the neighborhood of M, and Different points, Let Z be the mean of all points Z in the M neighborhood. Let M be the standard deviation of all points in the neighborhood of M in the Z direction. It is a scaling factor; Obtain the positive electrode point cloud dataset, the negative electrode point cloud dataset, and the aluminum base plate point cloud dataset. Extract the edge contour points of the electrode point cloud based on the threshold of the angle between the normal vectors of the positive electrode point cloud dataset and the negative electrode point cloud dataset. Based on the edge contour points of the pole point cloud, the pad circle is fitted with a three-dimensional circle using an improved RANSAC algorithm to obtain the pad circle size information. Then, the pole plane is extracted according to the pad circle size information to obtain the pole plane point cloud. The pole flatness is measured by the pole plane point cloud. The pole height is obtained by calculating the difference in the Z direction between the positive pole point cloud dataset, the negative pole point cloud dataset and the aluminum base plate point cloud dataset.
[0007] Furthermore, the second point cloud data is obtained by filtering and preprocessing the first target point cloud data, including: After rotating and translating the first target point cloud data so that the scanning platform coincides with the XOY plane, a pass-through filter is used to filter out noise and the point cloud of the scanning platform. Outliers in the first target point cloud data are filtered out by statistical filtering. Then, voxel filtering is used to downsample the data to obtain the preprocessed second point cloud data.
[0008] Further, extracting the edge contour points of the pole point cloud based on the threshold of the angle between the normal vectors of the positive pole point cloud dataset and the negative pole point cloud dataset includes: Normal vector estimation is performed on the positive pole point cloud dataset and the negative pole point cloud dataset to obtain the angle between the normals of each point cloud data and its neighboring points, where the normals are all parallel in the plane. Points with normal angles greater than a set threshold are extracted as edge contour points of the pole point cloud.
[0009] Furthermore, based on the edge contour points of the pole point cloud, an improved RANSAC algorithm is used to perform spatial three-dimensional circle fitting on the pad circle to obtain the pad circle size information, including: Input the dataset of polar edge contour points, the error threshold, the threshold for the minimum number of points required by the model, and the number of iterations; By substituting randomly selected initial points into the standard equation of a circle, a unique parametric equation for the circle is obtained, and the error equation of the circle is defined. ; The plane equation and plane normal vector are obtained based on randomly selected initial points; The dataset of polar contour points is traversed, and each point is substituted into the error equation of the circle. Points whose obtained values are less than the error threshold and are located within the plane equation are considered to conform to the current model. They are added to the preset current model, and the number of elements contained in the current model is detected until the number of elements in the current model is greater than the preset optimal model. The iteration terminates, and the current model is considered to be the optimal model. The radius and center of the pad circle are obtained by fitting the data in the optimal model using the least squares method.
[0010] Further, the process of extracting the pole plane and obtaining the pole plane point cloud based on the pad circle size information includes: Select any point from the positive pole point cloud dataset and the negative pole point cloud dataset. ,calculate Distance to the center O Where R is the radius of the solder pad circle, and all points in the positive and negative electrode point cloud datasets are traversed to delete them. The point cloud of the solder pads is used to obtain the point cloud of the polar plane.
[0011] Furthermore, the measurement of pole flatness using the pole plane point cloud includes: Fitting the polar plane using the least squares method Using this plane as a reference plane, for any point within the point cloud of the polar plane... Difference from the reference plane: ; in for The Z value, z is At the projection point on the reference plane, As a normalization factor, Iterate through and calculate each point in the polar plane point cloud. pole flatness .
[0012] Furthermore, the measurement parameters of the battery cover plate include the radius of the terminal pad, the center, the flatness of the terminal, and the height.
[0013] Secondly, the present invention provides a battery cover feature measurement device based on line laser scanning, comprising:
[0014] Acquisition module: used to scan the battery cover workpiece with a line laser 3D scanner and extract the first target point cloud data of the battery cover workpiece; Preprocessing module: used to perform filtering preprocessing on the first target point cloud data to obtain the second point cloud data; Feature extraction module: used to segment the point cloud features of the second point cloud data using an improved Euclidean clustering algorithm, including: Construct a kd-tree topology based on the second point cloud data, calculate feature parameters for each point in the second point cloud data, set an empty clustering list D, and a point cloud queue Q that needs to be checked; Take any point from the second point cloud data Place it into the point cloud queue Q that needs to be inspected, so that... Centered on, with Find the radius neighborhood element set ,examine Characteristic parameters of each element Whether it is less than the threshold, Elements whose feature parameters are less than a threshold are added to the point cloud queue Q, while elements whose feature parameters are greater than a threshold are not added to the point cloud queue Q. For each newly added element in the point cloud queue Q, take a neighborhood with the same radius and check whether the feature parameters of all elements in the neighborhood are less than the threshold and whether they have been processed. Add the elements that are less than the threshold and have not been processed back to the point cloud queue Q. When all points in the point cloud queue Q have been processed and no more elements are added, all points are placed into the cluster set D and then the point cloud queue Q is cleared. When all points in the second point cloud data have been processed and are located in the various clusters formed by the cluster set D, the segmentation ends. Any point in the second cloud data mentioned above Feature parameter definition: ; Where M is a spherical neighborhood of radius R, and N is the total number of data points in the M neighborhood. Within the neighborhood of M, and Different points, Let Z be the mean of all points Z in the M neighborhood. Let M be the standard deviation of all points in the neighborhood of M in the Z direction. It is a scaling factor; Obtain the positive electrode point cloud dataset, the negative electrode point cloud dataset, and the aluminum base plate point cloud dataset. Extract the edge contour points of the electrode point cloud based on the threshold of the angle between the normal vectors of the positive electrode point cloud dataset and the negative electrode point cloud dataset. Measurement module: Based on the edge contour points of the pole point cloud, the pad circle is fitted with a three-dimensional circle using an improved RANSAC algorithm to obtain the pad circle size information. Then, the pole plane is extracted based on the pad circle size information to obtain the pole plane point cloud. The pole flatness is measured by the pole plane point cloud. The pole height is calculated by the difference in the Z direction between the positive pole point cloud dataset, the negative pole point cloud dataset and the aluminum base plate point cloud dataset.
[0015] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the aforementioned method.
[0016] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:
[0017] This invention utilizes the characteristics of line laser scanning point clouds to extract the feature point cloud of the battery cover workpiece more quickly and efficiently, solving the problem of difficult parameter measurement for complex workpieces.
[0018] Traditional Euclidean clustering has good real-time segmentation performance, but it is prone to undersegmentation and oversegmentation when dealing with point cloud scenarios where there are stacked and adhered points (where the point cloud of the pole and the aluminum base plate connects). To address this problem, this invention uses feature parameters as conditions to restrict the segmentation process, preventing erroneous clustering while preserving real-time performance, thus solving the problem of difficult segmentation of the features of the pole and aluminum base plate in the point cloud of the battery cover workpiece.
[0019] The classic RANSAC algorithm can fit a model well in a dataset containing noisy points, but when measuring circular features in space, it can only fit a sphere and cannot obtain accurate circular measurement parameters. When using the RANSAC algorithm to fit circular features in space, coplanarity judgment needs to be added to achieve the function of fitting circular features in space and obtaining the corresponding parameters. Through the improved RANSAC algorithm, the problem of fitting spatial feature circles in pole point clouds is solved. This invention does not require manual selection of point clouds, making the measurement of cover plate parameters more intelligent and suitable for the scenario of measuring the three-dimensional dimensions of battery cover plates. Attached Figure Description
[0020] Figure 1 A flowchart of a battery cover feature measurement method based on line laser scanning provided in an embodiment of the present invention.
[0021] Figure 2 This is a scanned point cloud image of a battery cover workpiece provided in an embodiment of the present invention.
[0022] Figure 3 This is an improved Euclidean clustering flowchart provided in an embodiment of the present invention.
[0023] Figure 4 This is a schematic diagram of a point cloud edge determination method provided by an embodiment of the present invention when p is a boundary point.
[0024] Figure 5 This is a schematic diagram of a point cloud edge determination method provided by an embodiment of the present invention when p is an internal point.
[0025] Figure 6 This is a flowchart of the improved RANSAC algorithm provided in an embodiment of the present invention.
[0026] Figure 7 This is the pole plane segmentation provided in the embodiment of the present invention.
[0027] Figure 8 This is a comparison between the measurement results and actual values provided in the embodiments of this invention. Detailed Implementation
[0028] For ease of understanding, additional explanations are provided for the technical terms used in this embodiment:
[0029] RANSAC is an abbreviation for RandomSampleConsensus. It is an algorithm that estimates the parameters of a mathematical model from a set of sample datasets containing outliers through an iterative process, calculates the mathematical model parameters of the data, and obtains the valid sample data.
[0030] Point cloud: In reverse engineering, the collection of point data on the surface of a product obtained through measuring instruments is also called a point cloud.
[0031] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments and specific features in the embodiments are detailed descriptions of the technical solution of the present application, rather than limitations thereof. In the absence of conflict, the embodiments and technical features in the embodiments can be combined with each other.
[0032] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.
[0033] Example 1:
[0034] Figure 1 This is a flowchart of the battery cover feature measurement method based on line laser scanning according to Embodiment 1 of the present invention. This flowchart only illustrates the logical sequence of the method described in this embodiment. Provided there are no conflicts, different methods may be used in other possible embodiments of the present invention. Figure 1 Complete the steps shown or described in the order indicated.
[0035] The battery cover feature measurement method based on line laser scanning provided in this embodiment can be applied to a terminal and can be executed by a battery cover feature measurement device based on line laser scanning. This device can be implemented by software and / or hardware and can be integrated into the terminal, such as any smartphone, tablet computer or computer device with communication function.
[0036] This embodiment mainly uses C++ programming to call the PCL library to perform basic operations on point clouds.
[0037] See Figure 1 The method in this embodiment specifically includes the following steps:
[0038] Step 1: Scan the battery cover workpiece using a line laser 3D scanner, set the camera ROI, and use a high-precision fixture to extract the first target point cloud data of the battery cover workpiece. The first target point cloud data is shown below. Figure 2As shown, in deep learning, ROI usually refers to the Region of Interest. ROI technology is widely used in the fields of object detection and object recognition, especially when processing image and video data. Step 2: Perform filtering preprocessing on the first target point cloud data to obtain the second point cloud data: After rotating and translating the first target point cloud data so that the scanning platform coincides with the XOY plane, a pass-through filter is used to filter out noise and the point cloud of the scanning platform. Outliers in the first target point cloud data are filtered out by statistical filtering. Then, voxelization filtering is used to downsample the data to obtain the preprocessed second point cloud data. Step 3: Segment the point cloud features of the second point cloud data using the improved Euclidean clustering algorithm, such as... Figure 3 As shown: Construct a kd-tree topology based on the second point cloud data, calculate feature parameters for each point in the second point cloud data, set an empty clustering list D, and a point cloud queue Q that needs to be checked; Take any point from the second point cloud data Place it into the point cloud queue Q that needs to be inspected, so that... Centered on, with Find the radius neighborhood element set ,examine Characteristic parameters of each element Whether it is less than the threshold, Elements whose feature parameters are less than a threshold are added to the point cloud queue Q, while elements whose feature parameters are greater than a threshold are not added to the point cloud queue Q. For each newly added element in the point cloud queue Q, take a neighborhood with the same radius and check whether the feature parameters of all elements in the neighborhood are less than the threshold and whether they have been processed. If the feature parameters are less than the threshold and have not been processed, add them back to the point cloud queue Q. Repeat the above steps until the point cloud queue Q no longer adds elements. When all points in the point cloud queue Q have been processed and no more elements are added, all points are placed into the cluster set D and then the point cloud queue Q is cleared. Then, any point is selected and the above steps are repeated. When all points in the second point cloud data have been processed and are located in the various clusters formed by the cluster set D, the segmentation ends and the positive pole point cloud dataset, negative pole point cloud dataset, and aluminum base plate point cloud dataset are obtained. Any point in the second cloud data mentioned above Feature parameter definition: ; Where M is a spherical neighborhood of radius R, and N is the total number of data points in the M neighborhood. Within the neighborhood of M, and Different points, Let Z be the mean of all points Z in the M neighborhood. Let M be the standard deviation of all points in the neighborhood of M in the Z direction. It is a scaling factor; Since the patent targets battery cover workpieces, which are multi-planar point clouds, their point cloud characteristics are: the point cloud is basically uniformly distributed in the X and Y directions, and can be distinguished in the Z direction. In the above formula, by introducing the method of single-direction standard deviation in three-dimensional space, the correlation of data in the Z direction is increased in the clustering judgment process, achieving a more accurate clustering effect. At the same time, 1 / R focuses on the influence of the neighborhood size to adjust the spatial scale, and 1 / N focuses on the number of points in the neighborhood, providing a standardization for density changes.
[0039] like Figure 4 and Figure 5 As shown, the extraction of polar point cloud edge contour points based on the threshold of the angle between the normal vectors of the positive and negative polar point cloud datasets includes: Normal vector estimation is performed on the positive pole point cloud dataset and the negative pole point cloud dataset to obtain the angle between the normals of each point cloud data and its neighboring points, where the normals are all parallel in the plane. Points with normal angles greater than a set threshold are extracted as edge contour points of the pole point cloud. Step Four: As Figure 6 As shown, based on the edge contour points of the pole point cloud, the pad circle is fitted with a spatial three-dimensional circle using an improved RANSAC algorithm to obtain the pad circle size information: Specifically, the input is the dataset cloud of pole edge contour points, the error threshold and the threshold of the minimum number of points required by the model, as well as the number of iterations. In this embodiment, the initial number of selected points is 3. The randomly selected coordinates are as follows Substituting the three initial points into the standard equation of a circle yields a unique parametric equation for the circle: ; Define the error equation of a circle The plane equation and plane normal vector are obtained based on the randomly selected initial points. The vector can be obtained from the three initial points: , , Therefore, normal vector ; Substituting any point into the equation yields the plane equation. ; The dataset of polar contour points is traversed, and each point is substituted into the error equation of the circle. Points whose values are less than the error threshold and lie within the plane equation are considered to fit the current model. These points are added to `cur_model`, and the number of elements in `cur_model` is checked. If this number exceeds the minimum number of points required by the model, `cur_model` is set to `best_model`, and its elements are deleted. A second iteration is performed, selecting points again to fit the second circular model. The error threshold check is repeated, and points that fit the current model are added to `cur_model`. If the number of elements in `cur_model` is greater than the number of elements in `best_model`, `cur_model` is set to `best_model` again, `cur_model` is deleted, and the next iteration begins. If the number of elements in `cur_model` is less than the number of elements in `best_model`, the next iteration begins directly.
[0040] Repeat the above random point selection fitting circle model iteratively, find the interior points and put them into cur_model, until the number of elements in cur_model is no longer greater than the number of elements in best_model, or best_model reaches the number of iterations, stop the iteration, and output best_model; The radius and center of the pad circle are obtained by fitting the data in the best_model using the least squares method; like Figure 7 As shown, the pole plane point cloud is then obtained by extracting the pole plane based on the pad circle size information: Select any point from the positive pole point cloud dataset and the negative pole point cloud dataset. ,calculate Distance to the center O Where R is the radius of the solder pad circle, and all points in the positive and negative electrode point cloud datasets are traversed to delete them. The point cloud of the pads is used to obtain the point cloud of the pole plane; The flatness of the pole is measured using the point cloud of the pole plane. Fitting the polar plane using the least squares method Using this plane as a reference plane, for any point within the point cloud of the polar plane... Difference from the reference plane: ; in for The Z value, z is At the projection point on the reference plane, As a normalization factor, when the coordinate values in the x and y directions are large, The above calculation can be simplified to: ; traverse and calculate for each point in the pole plane point cloud , the pole flatness , where is the maximum difference, is the minimum difference; the height is obtained by calculating the difference in the Z direction between the positive pole point cloud dataset, the negative pole point cloud dataset and the aluminum base plate point cloud dataset.
[0041] as Figure 8 shown, the measurement parameters of the battery cover plate include the pole pad radius, the center of circle, the pole flatness and the height, where the center of circle is a process variable during the execution of the method, and the specific usage methods and steps after obtaining the center of circle and the pad radius are all mentioned above, so the center of circle Figure 8 will not be represented again in the following part.
[0042] Embodiment 2:
[0043] The second embodiment of the present invention further provides a battery cover plate feature measurement device based on line laser scanning, comprising:
[0044] Acquisition module: configured to scan the battery cover plate workpiece through a line laser 3D scanner, and extract first target point cloud data of the battery cover plate workpiece; Preprocessing module: configured to perform filtering preprocessing on the first target point cloud data to obtain second point cloud data; Feature extraction module: configured to segment point cloud features of the second point cloud data by an improved Euclidean clustering algorithm, including: constructing a kd-tree topological structure according to the second point cloud data, calculating a feature parameter for each point in the second point cloud data, setting an empty clustering list D, and a point cloud queue Q to be checked; taking any point in the second point cloud data and putting it into the point cloud queue Q to be checked, taking as the center, and taking as the radius to find the neighborhood element set , checking whether the feature parameter of each element in is less than a threshold, adding the elements with the feature parameter less than the threshold in to the point cloud queue Q, and not adding the elements with the feature parameter greater than the threshold to the point cloud queue Q; then taking the neighborhood of the same radius for each newly added element in the point cloud queue Q, checking whether the feature parameter of all elements in the neighborhood is less than the threshold and whether they have been processed, and adding the elements that are less than the threshold and have not been processed to the point cloud queue Q again; When all points in the point cloud queue Q have been processed and no more elements are added, all points are placed into the cluster set D and then the point cloud queue Q is cleared. When all points in the second point cloud data have been processed and are located in the various clusters formed by the cluster set D, the segmentation ends. Any point in the second cloud data mentioned above Feature parameter definition: ; Where M is a spherical neighborhood of radius R, and N is the total number of data points in the M neighborhood. Within the neighborhood of M, and Different points, Let Z be the mean of all points Z in the M neighborhood. Let M be the standard deviation of all points in the neighborhood of M in the Z direction. It is a scaling factor; Obtain the positive electrode point cloud dataset, the negative electrode point cloud dataset, and the aluminum base plate point cloud dataset. Extract the edge contour points of the electrode point cloud based on the threshold of the angle between the normal vectors of the positive electrode point cloud dataset and the negative electrode point cloud dataset. Measurement module: Based on the edge contour points of the pole point cloud, the pad circle is fitted with a three-dimensional circle using an improved RANSAC algorithm to obtain the pad circle size information. Then, the pole plane is extracted based on the pad circle size information to obtain the pole plane point cloud. The pole flatness is measured by the pole plane point cloud. The pole height is calculated by the difference in the Z direction between the positive pole point cloud dataset, the negative pole point cloud dataset and the aluminum base plate point cloud dataset.
[0045] The battery cover feature measurement device based on line laser scanning provided in this embodiment of the invention can execute the battery cover feature measurement method based on line laser scanning provided in this embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.
[0046] Example 3:
[0047] This invention also provides a computer-readable storage medium storing a computer program thereon. When executed by a processor, the program implements the steps of the method described in Embodiment 1, and has the corresponding functional modules and beneficial effects of the method.
[0048] Those skilled in the art will understand that embodiments of this application can be provided as methods, apparatus, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0049] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (devices), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0050] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0051] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0052] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.
Claims
1. A method for measuring the features of a battery cover plate based on line laser scanning, characterized in that, include: The battery cover workpiece is scanned by a line laser 3D scanner to extract the first target point cloud data of the battery cover workpiece. The first target point cloud data is filtered and preprocessed to obtain the second point cloud data; The point cloud features of the second point cloud data are segmented using an improved Euclidean clustering algorithm, including: Construct a kd-tree topology based on the second point cloud data, calculate feature parameters for each point in the second point cloud data, set an empty clustering list D, and a point cloud queue Q that needs to be checked; Take any point from the second point cloud data Place it into the point cloud queue Q that needs to be inspected, so that... Centered on, with Find the radius neighborhood element set ,examine Characteristic parameters of each element Whether it is less than the threshold, Elements whose feature parameters are less than a threshold are added to the point cloud queue Q, while elements whose feature parameters are greater than a threshold are not added to the point cloud queue Q. For each newly added element in the point cloud queue Q, take a neighborhood with the same radius and check whether the feature parameters of all elements in the neighborhood are less than the threshold and whether they have been processed. Add the elements that are less than the threshold and have not been processed back to the point cloud queue Q. When all points in the point cloud queue Q have been processed and no more elements are added, all points are placed into the cluster set D and then the point cloud queue Q is cleared. When all points in the second point cloud data have been processed and are located in the various clusters formed by the cluster set D, the segmentation ends. Any point in the second cloud data mentioned above Feature parameter definition: ； Where M is a spherical neighborhood of radius R, and N is the total number of data points in the M neighborhood. Within the neighborhood of M, and Different points, Let Z be the mean of all points Z in the M neighborhood. Let M be the standard deviation of all points in the neighborhood of M in the Z direction. It is a scaling factor; Obtain the positive electrode point cloud dataset, the negative electrode point cloud dataset, and the aluminum base plate point cloud dataset. Extract the edge contour points of the electrode point cloud based on the threshold of the angle between the normal vectors of the positive electrode point cloud dataset and the negative electrode point cloud dataset. Based on the edge contour points of the pole point cloud, the pad circle is fitted with a three-dimensional circle using an improved RANSAC algorithm to obtain the pad circle size information. Then, the pole plane is extracted according to the pad circle size information to obtain the pole plane point cloud. The pole flatness is measured by the pole plane point cloud. The pole height is obtained by calculating the difference in the Z direction between the positive pole point cloud dataset, the negative pole point cloud dataset and the aluminum base plate point cloud dataset.
2. The battery cover feature measurement method based on line laser scanning according to claim 1, wherein filtering and preprocessing the first target point cloud data to obtain the second point cloud data includes: After rotating and translating the first target point cloud data so that the scanning platform coincides with the XOY plane, a pass-through filter is used to filter out noise and the point cloud of the scanning platform. Outliers in the first target point cloud data are filtered out by statistical filtering. Then, voxel filtering is used to downsample the data to obtain the preprocessed second point cloud data.
3. The battery cover feature measurement method based on line laser scanning according to claim 1, characterized in that, Extracting the edge contour points of the polar point cloud based on the threshold of the angle between the normal vectors of the positive and negative polar point cloud datasets includes: Normal vector estimation is performed on the positive pole point cloud dataset and the negative pole point cloud dataset to obtain the angle between the normals of each point cloud data and its neighboring points, where the normals are all parallel in the plane. Points with normal angles greater than a set threshold are extracted as edge contour points of the pole point cloud.
4. The battery cover feature measurement method based on line laser scanning according to claim 1, characterized in that, Based on the edge contour points of the pole point cloud, an improved RANSAC algorithm is used to perform spatial three-dimensional circle fitting on the pad circle to obtain the pad circle size information, including: Input the dataset of polar edge contour points, the error threshold, the threshold for the minimum number of points required by the model, and the number of iterations; By substituting randomly selected initial points into the standard equation of a circle, a unique parametric equation for the circle is obtained, and the error equation of the circle is defined. ; The plane equation and plane normal vector are obtained based on randomly selected initial points; The dataset of polar contour points is traversed, and each point is substituted into the error equation of the circle. Points whose obtained values are less than the error threshold and are located within the plane equation are considered to conform to the current model. They are added to the preset current model, and the number of elements contained in the current model is detected until the number of elements in the current model is greater than the preset optimal model. The iteration terminates, and the current model is considered to be the optimal model. The radius and center of the pad circle are obtained by fitting the data in the optimal model using the least squares method.
5. The battery cover feature measurement method based on line laser scanning according to claim 1, characterized in that, The process of extracting the polarity plane point cloud based on the pad circle size information includes: Select any point from the positive pole point cloud dataset and the negative pole point cloud dataset. ,calculate Distance to the center O Where R is the radius of the solder pad circle, and all points in the positive and negative electrode point cloud datasets are traversed to delete them. The point cloud of the solder pads is used to obtain the point cloud of the pole plane.
6. The battery cover feature measurement method based on line laser scanning according to claim 1, characterized in that, The polarity flatness measurement is completed using the polarity plane point cloud, including: Fitting the polar plane using the least squares method Using this plane as a reference plane, for any point within the point cloud of the polar plane... Difference from the reference plane: ； in for The Z value, z is At the projection point on the reference plane, As a normalization factor, Iterate through and calculate each point in the polar plane point cloud. pole flatness .
7. The battery cover feature measurement method based on line laser scanning according to claim 1, characterized in that, The measurement parameters of the battery cover include the radius of the terminal pad, the center, the flatness of the terminal, and the height.
8. A battery cover feature measurement device based on line laser scanning, characterized in that, include: Acquisition module: used to scan the battery cover workpiece with a line laser 3D scanner and extract the first target point cloud data of the battery cover workpiece; Preprocessing module: used to perform filtering preprocessing on the first target point cloud data to obtain the second point cloud data; Feature extraction module: used to segment the point cloud features of the second point cloud data using an improved Euclidean clustering algorithm, including: Construct a kd-tree topology based on the second point cloud data, calculate feature parameters for each point in the second point cloud data, set an empty clustering list D, and a point cloud queue Q that needs to be checked; Take any point from the second point cloud data Place it into the point cloud queue Q that needs to be inspected, so that... Centered on, with Find the radius neighborhood element set ,examine Characteristic parameters of each element Whether it is less than the threshold, Elements whose feature parameters are less than a threshold are added to the point cloud queue Q, while elements whose feature parameters are greater than a threshold are not added to the point cloud queue Q. For each newly added element in the point cloud queue Q, take a neighborhood with the same radius and check whether the feature parameters of all elements in the neighborhood are less than the threshold and whether they have been processed. Add the elements that are less than the threshold and have not been processed back to the point cloud queue Q. When all points in the point cloud queue Q have been processed and no more elements are added, all points are placed into the cluster set D and then the point cloud queue Q is cleared. When all points in the second point cloud data have been processed and are located in the various clusters formed by the cluster set D, the segmentation ends. Any point in the second cloud data mentioned above Feature parameter definition: ； Where M is a spherical neighborhood of radius R, and N is the total number of data points in the M neighborhood. Within the neighborhood of M, and Different points, Let Z be the mean of all points Z in the M neighborhood. Let M be the standard deviation of all points in the neighborhood of M in the Z direction. It is a scaling factor; Obtain the positive electrode point cloud dataset, the negative electrode point cloud dataset, and the aluminum base plate point cloud dataset. Extract the edge contour points of the electrode point cloud based on the threshold of the angle between the normal vectors of the positive electrode point cloud dataset and the negative electrode point cloud dataset. Measurement module: Based on the edge contour points of the pole point cloud, the pad circle is fitted with a three-dimensional circle using an improved RANSAC algorithm to obtain the pad circle size information. Then, the pole plane is extracted based on the pad circle size information to obtain the pole plane point cloud. The pole flatness is measured by the pole plane point cloud. The pole height is calculated by the difference in the Z direction between the positive pole point cloud dataset, the negative pole point cloud dataset and the aluminum base plate point cloud dataset.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the method according to any one of claims 1 to 8.