A wheel hub polishing robot trajectory planning method based on 3D vision curve technology

Abstract

The present application relates to the technical field of hub polishing robots, and particularly relates to a hub polishing robot trajectory planning method based on 3D vision curve technology; by comparing an actual 3D curved surface graph with a preset 3D curved surface graph, a to-be-polished region model graph of each to-be-polished region in a target hub is generated; according to the to-be-polished region model graph of each to-be-polished region, optimal polishing parameters for polishing processing of each to-be-polished region in the target hub are obtained by pairing in a database; according to the optimal polishing parameters for polishing processing of each to-be-polished region in the target hub, optimal polishing trajectories of a polishing robot are planned; and the optimal polishing trajectories of the polishing robot and the optimal polishing parameters for polishing processing of each to-be-polished region in the target hub are sent to a control terminal. More accurate, self-adaptive, efficient and consistent polishing effects can be achieved, and production efficiency and product quality are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of hub polishing robots, and in particular to a hub polishing robot trajectory planning method based on 3D vision curve technology. BACKGROUND

[0002] In recent years, with the increasing labor costs in China year by year, automation and less people have become the common requirement of the hub manufacturing industry. And under the pressure of safety, environmental protection and many other aspects, the factory must improve and upgrade as soon as possible, especially some posts with poor environment and easy to cause occupational health hazards, so it is urgent to replace workers with robots. The die casting blank of the hub will produce many irregular burrs and flash after turning on the machine tool, which are distributed in the spoke, flange weight reduction recess, cap slot, center hole, PCD hole and valve hole, etc. The shape is multi-arc surface, and the machining range is narrow. It is difficult to rely on manual polishing, and it needs to change tools many times, which is labor-intensive, low in efficiency, large in dust, and uneven in polishing, and poor in consistency. The on-site labor intensity is large, the environmental noise is large, dust is generated, and the health of the operators is directly affected. If the robot automatically plans the polishing burr technology, it can reduce the labor intensity, achieve the purpose of reducing the number of employees, and effectively guarantee the product quality and consistency. However, the existing polishing robot planning method often has low precision in actual application, which cannot meet the demand of high-precision polishing; and the existing planning method has poor adaptability, cannot effectively adapt to the polishing demand of hubs of various shapes, and cannot plan an efficient polishing processing scheme, resulting in low hub production efficiency. SUMMARY

[0003] The present application overcomes the shortcomings of the prior art and provides a hub polishing robot trajectory planning method based on 3D vision curve technology.

[0004] To achieve the above-mentioned purpose, the technical scheme adopted by the present application is as follows:

[0005] The present application discloses a hub polishing robot trajectory planning method based on 3D vision curve technology, comprising the following steps:

[0006] Obtain the hub image of the target hub at a plurality of preset position nodes, and obtain the actual 3D curved surface image of the target hub according to the hub image; obtain the engineering drawing information of the target hub after polishing, and obtain the preset 3D curved surface image of the target hub according to the engineering drawing information;

[0007] Compare the actual 3D curved surface image with the preset 3D curved surface image to obtain a plurality of areas to be polished, and generate a model image of each area to be polished in the target hub;

[0008] According to the to-be-polished region model diagram of each to-be-polished region, the optimal polishing parameters for polishing each to-be-polished region in the target wheel hub are obtained by matching in the database; and the optimal polishing trajectory of the polishing robot is planned according to the optimal polishing parameters for polishing each to-be-polished region in the target wheel hub.

[0009] The optimal polishing trajectory of the polishing robot and the optimal polishing parameters for polishing each to-be-polished region in the target wheel hub are sent to a control terminal to control the polishing robot to polish the target wheel hub.

[0010] Further, in a preferred embodiment of the present application, wheel hub images of the target wheel hub are obtained at a plurality of preset position nodes, and an actual 3D curved surface diagram of the target wheel hub is obtained according to the wheel hub images, specifically as follows.

[0011] The wheel hub images of the target wheel hub are obtained at a plurality of preset position nodes, and the edge points of the wheel hub in each wheel hub image are obtained by performing feature extraction processing on each wheel hub image based on a Canny edge detection algorithm.

[0012] The descriptors of the edge points of the wheel hub in each wheel hub image are obtained by using a SIFT algorithm, and the Euclidean distances between the descriptors of the edge points of the wheel hub in different wheel hub images are calculated; the edge points between the wheel hubs in different wheel hub images with a Euclidean distance less than a preset Euclidean distance are matched to obtain a plurality of pairs of matching points.

[0013] A preset pair of matching points is randomly selected from the plurality of pairs of matching points as a sample pair, and a geometric model is established using these randomly selected sample pairs; the distance between the actual position and the predicted position of each matching point pair in the geometric model is calculated.

[0014] If the distance between the actual position and the predicted position of a matching point pair is greater than a preset threshold, the matching point pair is marked as an outlier; if the distance between the actual position and the predicted position of a matching point pair is not greater than the preset threshold, the matching point pair is marked as an inlier.

[0015] The matching point pairs marked as outliers are removed to remove incorrect matching point pairs and retain correct matching point pairs, to obtain screened matching point pairs, and a matching result is generated according to the screened matching point pairs.

[0016] The edge points of the wheel hub in each wheel hub image are matched and fused according to the matching result to obtain an edge point model, and an actual 3D curved surface diagram of the target wheel hub is generated according to the edge point model.

[0017] Further, in a preferred embodiment of the present application, the actual 3D curved surface diagram is compared with a preset 3D curved surface diagram to obtain a plurality of to-be-polished regions, and a to-be-polished region model diagram of each to-be-polished region in the target wheel hub is generated, specifically as follows.

[0018] constructing a three-dimensional space coordinate system, importing the actual 3D curved surface graph and the preset 3D curved surface graph into the three-dimensional space coordinate system, and aligning the actual 3D curved surface graph and the preset 3D curved surface graph in the three-dimensional space coordinate system based on an ICP algorithm;

[0019] after alignment, discretely processing the actual 3D curved surface graph based on a discretization method to obtain a plurality of first discrete points and obtain three-dimensional coordinate values of each first discrete point, and discretely processing the preset 3D curved surface graph to obtain a plurality of second discrete points and obtain three-dimensional coordinate values of each second discrete point;

[0020] calculating the Mahalanobis distance between the first discrete point and the second discrete point at each corresponding position in the three-dimensional space coordinate system according to the corresponding three-dimensional coordinate values;

[0021] if the Mahalanobis distance between the first discrete point and the second discrete point at a certain corresponding position is greater than a preset Mahalanobis distance, marking the first discrete point and the second discrete point at the corresponding position as non-coincidence points; if the Mahalanobis distance between the first discrete point and the second discrete point at a certain corresponding position is not greater than the preset Mahalanobis distance, marking the first discrete point and the second discrete point at the corresponding position as coincidence points;

[0022] obtaining a plurality of independent non-coincidence regions according to the non-coincidence regions between the actual 3D curved surface graph and the preset 3D curved surface graph formed by the non-coincidence points, marking each independent non-coincidence region as a to-be-polished region, and obtaining a to-be-polished region model graph of the to-be-polished region.

[0023] Further, in a preferred embodiment of the present application, the optimal polishing parameters for polishing each to-be-polished region in the target hub are obtained by matching each to-be-polished region model graph in the database, specifically:

[0024] obtaining the running log of all polishing robots in the production workshop, extracting the historical polishing parameters corresponding to the historical polishing region model graphs polished by the polishing robots in the running log, and obtaining the average polishing accuracy of the corresponding historical polishing regions polished by each historical polishing parameter;

[0025] sorting the average polishing accuracy of the corresponding historical polishing regions polished by each historical polishing parameter to obtain the maximum average polishing accuracy, and marking the historical polishing parameter corresponding to the maximum average polishing accuracy as the optimal polishing parameter of the corresponding historical polishing region model graph;

[0026] constructing a database, importing each historical polishing region model graph and its corresponding optimal polishing parameter into the database;

[0027] The target hub is obtained, and each to-be-polished region model graph of each to-be-polished region in the target hub is obtained. The similarity between each to-be-polished region model graph and each historical polishing region model graph in the database is calculated one by one through a Euclidean distance algorithm, and the historical polishing region model graph with the greatest similarity to each to-be-polished region model graph is screened out.

[0028] According to the historical polishing region model graph with the greatest similarity to each to-be-polished region model graph, the corresponding optimal polishing parameter in the database is matched, and the optimal polishing parameter for polishing processing of each to-be-polished region in the target hub is obtained.

[0029] Further, in a preferred embodiment of the present application, the optimal polishing trajectory of the polishing robot is planned according to the optimal polishing parameter for polishing processing of each to-be-polished region in the target hub, specifically:

[0030] The origin position information of the polishing robot is obtained, and the position information of each to-be-polished region is obtained. According to the origin position information of the polishing robot and the position information of each to-be-polished region, the to-be-polished region closest to the origin of the polishing robot is screened out, and the to-be-polished region closest to the origin of the polishing robot is marked as the first to-be-polished region.

[0031] The optimal polishing parameter of the first to-be-polished region and the optimal polishing parameter of the remaining to-be-polished region are obtained. The correlation coefficient between the optimal polishing parameter of the first to-be-polished region and the optimal polishing parameter of the remaining to-be-polished region is calculated through a Pearson correlation coefficient algorithm, and a plurality of correlation coefficients are obtained.

[0032] A size sorting table is constructed, and the plurality of correlation coefficients are imported into the size sorting table for descending order sorting processing, and a sorting result is obtained.

[0033] The maximum correlation coefficient is extracted from the sorting result, and the remaining to-be-polished region corresponding to the maximum correlation coefficient is obtained, and the remaining to-be-polished region corresponding to the maximum correlation coefficient is marked as the second to-be-polished region. The remaining to-be-polished region corresponding to the second largest correlation coefficient is obtained, and the remaining to-be-polished region corresponding to the second largest correlation coefficient is marked as the third to-be-polished region. In succession, until all the remaining to-be-polished regions are planned, the Nth to-be-polished region is obtained.

[0034] The position information of the first to-be-polished region to the Nth to-be-polished region is obtained in sequence, and the optimal polishing trajectory of the polishing robot is generated according to the position information of the first to-be-polished region to the Nth to-be-polished region.

[0035] Further, in a preferred embodiment of the present application, the following steps are further included:

[0036] If two or more of the plurality of correlation coefficients are equal, obtain the position distances between each of the remaining polishing areas with equal correlation coefficients and the last planned polishing area, to obtain a plurality of position distances;

[0037] Sort the position distances from small to large, mark the remaining polishing area corresponding to the smallest position distance as the Nth polishing area, mark the remaining polishing area corresponding to the second smallest position distance as the N+1th polishing area, and so on, until the N+Nth polishing area is obtained, until all the remaining polishing areas with equal correlation coefficients are planned;

[0038] Obtain the position information of the first polishing area to the N+Nth polishing area in turn, and generate the optimal polishing trajectory of the polishing robot according to the position information of the first polishing area to the N+Nth polishing area.

[0039] The second aspect of the present application discloses a hub polishing robot trajectory planning system based on 3D vision curve technology, which comprises a memory and a processor, and the memory stores a hub polishing robot trajectory planning method program.

[0040] Obtain the hub image of the target hub at a plurality of preset position nodes, obtain the actual 3D surface graph of the target hub according to the hub image, obtain the engineering drawing information of the target hub after polishing, and obtain the preset 3D surface graph of the target hub according to the engineering drawing information;

[0041] Compare the actual 3D surface graph with the preset 3D surface graph to obtain a plurality of polishing areas, and generate a polishing area model graph of each polishing area in the target hub;

[0042] According to the polishing area model graph of each polishing area, the optimal polishing parameters for polishing each polishing area in the target hub are obtained by matching in the database, and the optimal polishing trajectory of the polishing robot is planned according to the optimal polishing parameters for polishing each polishing area in the target hub;

[0043] The optimal polishing trajectory of the polishing robot and the optimal polishing parameters for polishing each polishing area in the target hub are sent to a control terminal to control the polishing robot to polish the target hub.

[0044] The present application solves the technical defects in the background art, and has the following beneficial effects: obtaining hub images of a target hub at multiple preset position nodes, obtaining an actual 3D curved surface map of the target hub according to the hub images; obtaining engineering drawing information of the target hub after polishing, obtaining a preset 3D curved surface map of the target hub according to the engineering drawing information; comparing the actual 3D curved surface map with the preset 3D curved surface map to obtain a plurality of polishing areas, and generating a polishing area model map of each polishing area in the target hub; pairing the polishing area model map of each polishing area in the database to obtain optimal polishing parameters for polishing each polishing area in the target hub; planning an optimal polishing trajectory of a polishing robot according to the optimal polishing parameters for polishing each polishing area in the target hub; and sending the optimal polishing trajectory of the polishing robot and the optimal polishing parameters for polishing each polishing area in the target hub to a control terminal to control the polishing robot to polish the target hub. The hub polishing robot trajectory planning method based on 3D visual curve technology can achieve more accurate, adaptive, efficient and consistent polishing effect, and improves production efficiency and product quality. BRIEF DESCRIPTION OF DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.

[0046] Figure 1 The overall method flowchart of the present hub polishing robot trajectory planning method;

[0047] Figure 2 The partial method flowchart of the present hub polishing robot trajectory planning method;

[0048] Figure 3 The system block diagram of the present hub polishing robot trajectory planning system. DETAILED DESCRIPTION

[0049] In order to more clearly illustrate the above-mentioned purposes, features and advantages of the present application, the following will further describe the present application in combination with the drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0050] In the following description, a lot of specific details are set forth in order to provide a thorough understanding of the present application, however, the present application can be practiced in other manners different from those described herein, therefore, the scope of protection of the present application is not limited by the specific embodiments disclosed below.

[0051] The first aspect of the present application discloses a hub polishing robot trajectory planning method based on 3D vision curve technology, comprising the following steps:

[0052] S102: Obtain hub images of the target hub at a plurality of preset position nodes, and obtain an actual 3D curved surface image of the target hub according to the hub images; obtain engineering drawing information of the target hub after polishing, and obtain a preset 3D curved surface image of the target hub according to the engineering drawing information;

[0053] S104: Compare the actual 3D curved surface image with the preset 3D curved surface image to obtain a plurality of polishing areas, and generate a polishing area model image of each polishing area in the target hub;

[0054] S106: Obtain optimal polishing parameters for polishing each polishing area in the target hub according to the polishing area model image of each polishing area in the database; and plan an optimal polishing trajectory of the polishing robot according to the optimal polishing parameters for polishing each polishing area in the target hub;

[0055] S108: Send the optimal polishing trajectory of the polishing robot and the optimal polishing parameters for polishing each polishing area in the target hub to a control terminal to control the polishing robot to polish the target hub.

[0056] Further, in a preferred embodiment of the present application, the hub images of the target hub are obtained at a plurality of preset position nodes, and the actual 3D curved surface image of the target hub is obtained according to the hub images, specifically:

[0057] The hub images of the target hub are obtained at a plurality of preset position nodes, and feature extraction processing is performed on each hub image based on a Canny edge detection algorithm to obtain edge points of the hub in each hub image;

[0058] The descriptors of the edge points of the hub in each hub image are obtained by using a SIFT algorithm, and the Euclidean distances between the descriptors of the edge points of the hub in different hub images are calculated; the edge points between the hubs in different hub images with a Euclidean distance less than a preset Euclidean distance are matched to obtain a plurality of pairs of matching points;

[0059] randomly select a preset number of matching point pairs as sample pairs from the matching point pairs, and use the sample pairs to establish a geometric model, and calculate the distance between the actual position and the predicted position of each matching point pair in the geometric model;

[0060] If the distance between the actual position and the predicted position of a matching point pair is greater than a preset threshold, the matching point pair is marked as an outlier; if the distance between the actual position and the predicted position of a matching point pair is not greater than a preset threshold, the matching point pair is marked as an inlier;

[0061] The matching point pairs marked as outliers are removed to remove incorrect matching point pairs and retain correct matching point pairs, and a filtered matching point pair is obtained, and a matching result is generated according to the filtered matching point pair;

[0062] According to the matching result, the edge points of the hub in each hub image are matched and fused to obtain an edge point model, and an actual 3D curved surface map of the target hub is generated according to the edge point model.

[0063] It should be noted that when the target hub is clamped and positioned, the industrial camera mounted on the robot is used to obtain hub images of the target hub at a plurality of preset position nodes, and the edge points of the hub in each image are extracted. After obtaining the edge points, a feature point extraction algorithm (such as SIFT, SURF or ORB) can be used to extract feature points with unique characteristics from the edge points. These feature points have rotation and scale invariance and can be identified and matched in images with different perspectives. For each feature point, a descriptor, i.e., a numerical vector, is generated to describe the image features around the point. The descriptor can effectively match the same feature points in different images. Feature matching algorithms (such as brute force matching and FLANN fast nearest neighbor search) are used to compare the descriptors of the feature points in different images to find matching pairs. This step is achieved by calculating and comparing the distances (such as Euclidean distance) between the descriptors. The smaller the distance, the more similar the two feature points. During the matching process, there may be false matches. RANSAC (Random Sample Consensus) algorithm can be used to remove false matches and retain correct matching point pairs, thereby improving the accuracy and reliability of the matching. Then, according to the matching result, the edge points of the hub in each hub image are matched and fused to obtain an edge point model, and an actual 3D curved surface map of the target hub is generated according to the edge point model. Through this method, the corresponding hub 3D curved surface map can be quickly generated according to the hub image.

[0064] As Figure 2As shown, further, in a preferred embodiment of the present application, the actual 3D curved surface graph is compared with the preset 3D curved surface graph to obtain a plurality of polishing areas, and a polishing area model graph of each polishing area in the target hub is generated, specifically:

[0065] S202: A three-dimensional space coordinate system is constructed, the actual 3D curved surface graph and the preset 3D curved surface graph are imported into the three-dimensional space coordinate system, and the actual 3D curved surface graph and the preset 3D curved surface graph are aligned in the three-dimensional space coordinate system based on the ICP algorithm;

[0066] S204: After alignment, the actual 3D curved surface graph is discretely processed based on the discretization method to obtain a plurality of first discrete points, and the three-dimensional coordinate values of each first discrete point are obtained; and the preset 3D curved surface graph is discretely processed to obtain a plurality of second discrete points, and the three-dimensional coordinate values of each second discrete point are obtained;

[0067] S206: The Mahalanobis distance between the first discrete point and the second discrete point at each corresponding position is calculated according to the corresponding three-dimensional coordinate values in the three-dimensional space coordinate system;

[0068] S208: If the Mahalanobis distance between the first discrete point and the second discrete point at a certain corresponding position is greater than a preset Mahalanobis distance, the first discrete point and the second discrete point at the corresponding position are marked as non-coincidence points; if the Mahalanobis distance between the first discrete point and the second discrete point at a certain corresponding position is not greater than the preset Mahalanobis distance, the first discrete point and the second discrete point at the corresponding position are marked as coincidence points;

[0069] S210: According to the non-coincidence area between the actual 3D curved surface graph and the preset 3D curved surface graph composed of each non-coincidence point, a plurality of independent non-coincidence areas are obtained, each independent non-coincidence area is marked as a polishing area, and a polishing area model graph of the polishing area is obtained.

[0070] It should be noted that the ICP algorithm is an algorithm for point cloud registration, which optimizes the rigid body transformation between two point clouds iteratively to make them as aligned as possible, thereby realizing the registration of point clouds. The basic idea of the ICP algorithm is to find the best rigid body transformation by minimizing the distance between two point clouds. In each iteration, the algorithm matches each point of one point cloud with the nearest point in the other point cloud, and then updates the rigid body transformation matrix according to the relationship of the matched points, and this process is iterated until the optimal rigid body transformation is converged.

[0071] If the Mahalanobis distance between the first discrete point and the second discrete point at a corresponding position is greater than the preset Mahalanobis distance, the first discrete point and the second discrete point at the corresponding position are marked as non-coincidence points; if the Mahalanobis distance between the first discrete point and the second discrete point at a corresponding position is not greater than the preset Mahalanobis distance, the first discrete point and the second discrete point at the corresponding position are marked as coincidence points. Then, the non-coincidence points are combined to obtain a plurality of non-coincidence regions, and each non-coincidence region is a burr region that needs to be polished. Then, the model graph of the non-coincidence region is separated, and the burr region model graph of each burr region to be polished is obtained. The method compares the actual 3D curved surface graph with the preset 3D curved surface graph, thereby quickly positioning the burr position region of the target wheel hub, and the burr region model graph can be separated. The function of quickly positioning the burr position region and obtaining the burr region model graph is realized through simple algorithm steps, which can improve the corresponding speed of the system and improve the processing efficiency.

[0072] Further, in a preferred embodiment of the present application, the optimal polishing parameters for polishing each burr region to be polished in the target wheel hub are obtained by matching each burr region model graph of the burr region to be polished in the database, specifically:

[0073] Obtain the running log of all polishing robots in the production workshop, and extract the historical polishing parameters corresponding to the historical polishing region model graphs polished by the polishing robots in the running log, and obtain the average polishing accuracy of the corresponding historical polishing regions polished by each historical polishing parameter;

[0074] Sort the average polishing accuracy of the corresponding historical polishing regions polished by each historical polishing parameter to obtain the maximum average polishing accuracy, and mark the historical polishing parameter corresponding to the maximum average polishing accuracy as the optimal polishing parameter of the corresponding historical polishing region model graph;

[0075] Construct a database, and import each historical polishing region model graph and its corresponding optimal polishing parameter into the database;

[0076] Obtain the burr region model graph of each burr region to be polished in the target wheel hub, calculate the similarity between each burr region model graph and each historical polishing region model graph in the database by using the Euclidean distance algorithm, and select the historical polishing region model graph with the greatest similarity to each burr region model graph;

[0077] According to the optimal polishing parameters corresponding to the historical polishing region model graphs with the greatest similarity to each burr region model graph in the database, the optimal polishing parameters for polishing each burr region to be polished in the target wheel hub are obtained.

[0078] The polishing parameters include a polishing speed, a polishing pressure, a polishing temperature, and a polishing time. The operation log of the polishing robot refers to a record file recording data and information generated by the robot in the polishing process, including an operation time, polishing parameters, an operation state, abnormal information, a polishing result, and the like.

[0079] It should be noted that the historical polishing area model graph and the corresponding optimal polishing parameters are extracted from the operation log of the polishing robot, and the database is constructed in advance, the historical polishing area model graph and the corresponding optimal polishing parameters are imported into the database, then the polishing area model graph of each polishing area to be polished in the target hub is obtained, and the optimal polishing parameters for polishing each polishing area to be polished are quickly matched in the database according to each polishing area model graph. In this way, the polishing parameters of each polishing area to be polished are automatically matched, without manual planning of polishing parameters, intelligent processing and production are realized, and the polishing efficiency and accuracy are effectively improved.

[0080] Further, in a preferred embodiment of the present application, the optimal polishing trajectory of the polishing robot is planned according to the optimal polishing parameters for polishing each polishing area to be polished in the target hub, specifically:

[0081] The origin position information of the polishing robot is obtained, and the position information of each polishing area to be polished is obtained, the polishing area to be polished closest to the origin of the polishing robot is selected according to the origin position information of the polishing robot and the position information of each polishing area to be polished, and the polishing area to be polished closest to the origin of the polishing robot is marked as the first polishing area to be polished;

[0082] The optimal polishing parameters of the first polishing area to be polished are obtained, and the optimal polishing parameters of the remaining polishing areas to be polished are obtained; the correlation coefficients between the optimal polishing parameters of the first polishing area to be polished and the optimal polishing parameters of the remaining polishing areas to be polished are calculated by a Pearson correlation coefficient algorithm, to obtain a plurality of correlation coefficients;

[0083] A size sorting table is constructed, the plurality of correlation coefficients are imported into the size sorting table for descending order sorting processing, to obtain a sorting result;

[0084] The maximum correlation coefficient is extracted from the sorting result, and the remaining polishing area corresponding to the maximum correlation coefficient is obtained, and the remaining polishing area corresponding to the maximum correlation coefficient is marked as the second polishing area to be polished; the remaining polishing area corresponding to the second largest correlation coefficient is obtained, and the remaining polishing area corresponding to the second largest correlation coefficient is marked as the third polishing area to be polished; and the process is repeated until all the remaining polishing areas to be polished are planned, to obtain the Nth polishing area to be polished;

[0085] The position information of the first to Nth polishing areas is acquired in sequence, and the optimal polishing track of the polishing robot is generated according to the position information of the first to Nth polishing areas.

[0086] It should be noted that, after the optimal polishing parameters for polishing each polishing area in the target hub are matched, the polishing area closest to the origin of the polishing robot is selected according to the position information of the origin of the polishing robot and the position information of each polishing area, and the polishing area closest to the origin of the polishing robot is marked as the first polishing area, then the optimal polishing parameters of the first polishing area are obtained, and the optimal polishing parameters of the remaining polishing areas are obtained; the correlation coefficient between the optimal polishing parameters of the first polishing area and the optimal polishing parameters of the remaining polishing areas is calculated by the Pearson correlation coefficient algorithm, and a plurality of correlation coefficients are obtained, wherein the larger the correlation coefficient, the higher the linear correlation between the two polishing parameters, that is, their trends are more consistent, so it can be considered that the higher the similarity between the two polishing parameters, when the Pearson correlation coefficient is 1, it indicates a complete positive correlation; when it is -1, it indicates a complete negative correlation, and when it is 0, it indicates no linear correlation. Then, the remaining polishing area corresponding to the maximum correlation coefficient is obtained, and the remaining polishing area corresponding to the maximum correlation coefficient is marked as the second polishing area; the remaining polishing area corresponding to the second largest correlation coefficient is obtained, and the remaining polishing area corresponding to the second largest correlation coefficient is marked as the third polishing area; in turn, until all the remaining polishing areas are planned, the Nth polishing area is obtained, and the position information of the first polishing area to the Nth polishing area is obtained in turn, and the optimal polishing trajectory of the polishing robot is generated according to the position information of the first polishing area to the Nth polishing area. For example, if there are three burr areas in the target hub, that is, there are three polishing areas, the polishing area closest to the polishing head is obtained first and marked as the first polishing area, then the correlation coefficient between the optimal polishing parameters of the remaining two polishing areas and the optimal polishing parameters of the first polishing area is calculated, the larger the correlation coefficient, the higher the similarity of the optimal polishing parameters of the first polishing area, at this time the polishing area with the largest correlation coefficient is marked as the second polishing area, then the last remaining polishing area is marked as the third polishing area, and the optimal polishing trajectory of the polishing head is planned according to the polishing order of the three polishing areas, that is, the first polishing area is polished, then the second polishing area is polished, and finally the third polishing area is polished.The purpose of planning the polishing path is to reduce the polishing parameter state switching range of the polishing head during polishing of each to-be-polished area by the polishing head, so that the polishing head can smoothly transition between different areas, avoiding uneven surface and traces caused by sudden parameter changes; reducing the switching range can reduce the frequency of parameter adjustment, thereby reducing energy consumption; reducing the pause and adjustment time of the polishing head between different areas can improve polishing efficiency; in general, the method can make the polishing process smoother and more efficient, while reducing energy consumption and protecting the workpiece surface.

[0087] Further, in a preferred embodiment of the present application, the following steps are further included:

[0088] If the calculated plurality of correlation coefficients exist two or more correlation coefficients equal to each other, the position distance between each correlation coefficient equal to the remaining to-be-polished area and the last planned to-be-polished area is obtained, and a plurality of position distances are obtained.

[0089] The position distances are sorted from small to large, the remaining to-be-polished area corresponding to the smallest position distance is marked as the Nth to-be-polished area, the remaining to-be-polished area corresponding to the second smallest position distance is marked as the N+1th to-be-polished area, and so on, until the N+Nth to-be-polished area is obtained, until all correlation coefficient equal to the remaining to-be-polished area is planned.

[0090] The position information of the first to-be-polished area to the N+Nth to-be-polished area is obtained in sequence, and the optimal polishing trajectory of the polishing robot is generated according to the position information of the first to-be-polished area to the N+Nth to-be-polished area.

[0091] It should be noted that if the calculated plurality of correlation coefficients exist two or more correlation coefficients equal to each other, such as two correlation coefficients equal to 1, it means that the polishing parameters of two to-be-polished areas are the same as the polishing parameters of the first to-be-polished area, at this time, by analyzing the positions of the two to-be-polished areas with correlation coefficient equal to 1, the to-be-polished area closest to the first to-be-polished area is selected first and marked as the second to-be-polished area, then the remaining to-be-polished area with correlation coefficient equal to 1 is marked as the third to-be-polished area, and then the to-be-polished areas with correlation coefficient not equal to 1 are planned according to the size order of the correlation coefficient.

[0092] In summary, the 3D visual curve technology can provide more accurate hub surface three-dimensional shape information, so as to more accurately plan the polishing scheme to adapt to the hub surface of different shapes, thereby realizing higher quality polishing effect; realizing the adaptability of robot trajectory planning, which can dynamically adjust the polishing trajectory according to the geometric characteristics of the hub surface, thereby polishing the target hub with high precision; can intelligently generate the corresponding polishing scheme according to the burr condition of the target hub, can effectively adapt to the polishing needs of hubs of various shapes, and can effectively improve the production efficiency.

[0093] In addition, the method further comprises the following steps:

[0094] During the polishing of each to-be-polished area, real-time polishing parameters are obtained at a preset machining time node, and the real-time polishing parameters are compared with the corresponding optimal polishing parameters to obtain a polishing parameter deviation value;

[0095] If the polishing parameter deviation value is greater than a preset deviation value, an ultrasonic detector is used to scan the polishing area being polished by the polishing robot, and ultrasonic information fed back by the polishing area is obtained, and a three-dimensional model diagram of the polishing area is constructed according to the ultrasonic information fed back by the polishing area;

[0096] The three-dimensional model diagram of the polishing area is analyzed to determine whether the three-dimensional model diagram of the polishing area has a polishing crack;

[0097] If there is a polishing crack, it is further analyzed whether the polishing crack extends to a non-polishing area; if the polishing crack extends to the non-polishing area, polishing machining of the current target hub is stopped, and the current target hub is scrapped;

[0098] If there is no polishing crack, a control instruction is generated according to the polishing parameter deviation value, and the real-time polishing parameters of the polishing robot are controlled according to the control instruction.

[0099] Further comprising the following steps:

[0100] If the polishing crack does not extend to the non-polishing area, crack characteristic data of each crack in the three-dimensional model diagram of the polishing area is calculated, the volume value of each crack is calculated according to the crack characteristic data of each crack, and the total volume value of the cracks in the three-dimensional model diagram of the polishing area is obtained according to the volume value of each crack;

[0101] The crack proportion is calculated according to the total volume value of the cracks and the total volume of the three-dimensional model diagram of the polishing area; if the crack proportion is greater than a preset proportion, polishing machining of the current target hub is stopped, and the current target hub is scrapped;

[0102] If the crack proportion is not greater than the preset proportion, finite element analysis is performed on each crack in the polishing area three-dimensional model diagram to obtain a critical crack propagation force of each crack; the critical crack propagation forces of the cracks are compared to obtain a minimum critical crack propagation force;

[0103] The maximum cutting force of the polishing head is calculated according to the minimum critical crack propagation force, the maximum polishing parameter of the polishing head is determined according to the maximum cutting force, the maximum polishing parameter is compared with the real-time polishing parameter to obtain a second polishing parameter deviation value;

[0104] A second control instruction is generated according to the second polishing parameter deviation value, and the real-time polishing parameter of the polishing robot is controlled according to the second control instruction.

[0105] It should be noted that during the polishing process of each polishing area, if the polishing parameter deviation value is greater than the preset deviation value, it indicates that the parameter of the polishing head has abnormally changed at this time. At this time, the polishing area three-dimensional model diagram is obtained, and then it is judged whether there is a polishing crack in the polishing area three-dimensional model diagram. If there is no polishing crack, it indicates that the abnormal polishing parameter does not cause a polishing crack. At this time, a control instruction is generated according to the polishing parameter deviation value, and the real-time polishing parameter of the polishing robot is controlled according to the control instruction, so that the polishing parameter returns to normal, thereby avoiding causing a polishing crack. If there is a polishing crack, it is further analyzed whether the polishing crack extends to the non-polishing area; if the polishing crack extends to the non-polishing area, it indicates that the hub being polished is a waste product at this time, and the polishing process of the current target hub is stopped, and the current target hub is scrapped. In this way, the semi-finished product is immediately scrapped, which can reduce the waste of processing resources and save costs.

[0106] If the polishing crack does not extend to the non-polishing area, the crack ratio is calculated, and if the crack ratio is greater than the preset ratio, it means that a large number of cracks have been generated in the polishing area, and the probability of crack instability extension is extremely high, and the probability of product processing being a waste product is extremely high, at this time, the polishing processing of the current target hub is stopped, and the current target hub is scrapped. If the crack ratio is not greater than the preset ratio, finite element analysis is performed on each crack in the polishing area three-dimensional model diagram to obtain the critical crack propagation force of each crack; compare the critical crack propagation forces of each crack to obtain the smallest critical crack propagation force, and then determine the maximum polishing parameters such as the maximum polishing speed, the maximum polishing temperature, etc. according to the smallest critical crack propagation force. Then compare the maximum polishing parameters with the real-time polishing parameters to obtain a second polishing parameter deviation value, generate a second control instruction according to the second polishing parameter deviation value, and control the real-time polishing parameters of the polishing robot according to the second control instruction. In this way, the polishing parameters of the polishing robot are adjusted, which can effectively avoid the further extension of the polishing crack and reduce the product scrap rate.

[0107] In addition, the present method further comprises the following steps:

[0108] Obtain the limit cutting force that the target hub can withstand under various preset temperature field conditions through a big data network;

[0109] Construct a knowledge graph, and import the limit cutting force that the target hub can withstand under various preset temperature field conditions into the knowledge graph;

[0110] When polishing each polishing area, obtain the real-time temperature field of the polishing area in real time, import the real-time temperature field into the knowledge graph for pairing, and obtain the limit cutting force that the target hub can withstand under the real-time temperature field condition;

[0111] Obtain the real-time cutting force of the polishing head, and compare the real-time cutting force with the limit cutting force that the target hub can withstand;

[0112] If the real-time cutting force is greater than the limit cutting force that the target hub can withstand, calculate the difference between the real-time cutting force and the limit cutting force that the target hub can withstand to obtain a cutting force difference;

[0113] Control the real-time cutting force of the polishing head according to the cutting force difference.

[0114] It should be noted that, in the grinding process, when the cutting force is too large, it will generate excessive stress on the hub material, and if the stress exceeds the yield limit or fatigue limit of the material, it may cause micro-cracks inside or on the surface of the material, and with the continuation of the grinding process, the micro-cracks may expand, and eventually form significant cracks. In order to avoid the initiation of grinding cracks as much as possible, the cutting force needs to be precisely controlled when grinding the hub to ensure that it can effectively remove the material without damaging the hub.

[0115] As shown in Figure 3 The second aspect of the present application discloses a hub grinding robot trajectory planning system based on 3D vision curve technology, which comprises a memory 20 and a processor 80, wherein the memory 20 stores a hub grinding robot trajectory planning method program, and when the hub grinding robot trajectory planning method program is executed by the processor 80, the following steps are implemented:

[0116] Obtaining hub images of the target hub at a plurality of preset position nodes, obtaining an actual 3D curved surface diagram of the target hub according to the hub images, obtaining engineering drawing information of the target hub after grinding, and obtaining a preset 3D curved surface diagram of the target hub according to the engineering drawing information;

[0117] Comparing the actual 3D curved surface diagram with the preset 3D curved surface diagram to obtain a plurality of to-be-grounded areas, and generating a to-be-grounded area model diagram of each to-be-grounded area in the target hub;

[0118] According to the to-be-grounded area model diagram of each to-be-grounded area, the optimal grinding parameters for grinding each to-be-grounded area in the target hub are matched in the database; and according to the optimal grinding parameters for grinding each to-be-grounded area in the target hub, the optimal grinding trajectory of the grinding robot is planned;

[0119] The optimal grinding trajectory of the grinding robot and the optimal grinding parameters for grinding each to-be-grounded area in the target hub are sent to a control terminal to control the grinding robot to grind the target hub.

[0120] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only illustrative, for example, the division of the units is only a logical function division, and in actual implementation, another division mode can be used, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the various components shown or discussed can be through some interface, indirect coupling or communication connection between devices or units, which can be electrical, mechanical or other forms.

[0121] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0122] In addition, in the various embodiments of the present invention, each functional unit can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0123] 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.

[0124] Alternatively, if the integrated units of this invention are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product 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 of the various embodiments of this 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.

[0125] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A trajectory planning method for a wheel hub grinding robot based on 3D vision curve technology, characterized in that, Includes the following steps: The target wheel hub image is acquired at multiple preset position nodes, and the actual 3D surface map of the target wheel hub is obtained based on the wheel hub image; Obtain the engineering drawing information of the target wheel hub after polishing, and obtain the preset 3D surface drawing of the target wheel hub based on the engineering drawing information; The actual 3D surface map is compared with the preset 3D surface map to obtain several areas to be polished, and a model map of the areas to be polished in each area of ​​the target wheel hub is generated. Based on the model diagrams of each area to be polished, the optimal polishing parameters for polishing each area in the target wheel hub are obtained by matching them in the database; based on the optimal polishing parameters for polishing each area in the target wheel hub, the optimal polishing trajectory of the polishing robot is planned. The optimal grinding trajectory of the grinding robot and the optimal grinding parameters for grinding each area to be ground in the target wheel hub are sent to the control terminal to control the grinding robot to grind the target wheel hub. Specifically, the optimal grinding parameters for grinding each area in the target wheel hub are obtained by matching the model diagrams of each area to be ground in the database. Obtain the operation logs of all grinding robots in the production workshop, extract the historical grinding parameters corresponding to the historical grinding area model diagrams of various historical grinding areas by the grinding robots from the operation logs, and obtain the average grinding accuracy of the corresponding historical grinding area after grinding with each historical grinding parameter. The average grinding accuracy of the corresponding historical grinding area after grinding with each historical grinding parameter is sorted by grinding accuracy to obtain the maximum average grinding accuracy. The historical grinding parameters corresponding to the maximum average grinding accuracy are marked as the optimal grinding parameters of the corresponding historical grinding area model. Build a database and import the model images of each historical polishing area and their corresponding optimal polishing parameters into the database; Obtain the model image of each area to be polished in the target wheel hub. Calculate the similarity between each model image of the area to be polished and each historical model image of the polished area in the database using the Euclidean distance algorithm. Then, select the historical model image of the polished area that has the highest similarity to each model image of the area to be polished. Based on the historical grinding area model map that has the highest similarity to each grinding area model map, the corresponding optimal grinding parameters are matched in the database to obtain the optimal grinding parameters for grinding each grinding area in the target wheel hub. Specifically, the optimal grinding trajectory of the grinding robot is planned based on the optimal grinding parameters for grinding each area to be ground in the target wheel hub, as follows: Obtain the origin position information of the polishing robot and the position information of each area to be polished. Based on the origin position information of the polishing robot and the position information of each area to be polished, filter out the area to be polished that is closest to the origin of the polishing robot and mark the area to be polished that is closest to the origin of the polishing robot as the first area to be polished. Obtain the optimal polishing parameters for the first area to be polished, and obtain the optimal polishing parameters for the remaining areas to be polished; calculate the correlation coefficient between the optimal polishing parameters of the first area to be polished and the optimal polishing parameters of the remaining areas to be polished using the Pearson correlation coefficient algorithm, and obtain multiple correlation coefficients; Construct a size sorting table, import multiple correlation coefficients into the size sorting table and sort them in descending order to obtain the sorting result; Extract the maximum correlation coefficient from the sorting results, and obtain the remaining area to be polished corresponding to the maximum correlation coefficient. Mark the remaining area to be polished corresponding to the maximum correlation coefficient as the second area to be polished. Obtain the remaining area to be polished corresponding to the second maximum correlation coefficient, and mark the remaining area to be polished corresponding to the second maximum correlation coefficient as the third area to be polished. Continue in this manner until all remaining areas to be polished are planned, and obtain the Nth area to be polished. The position information of the first area to be polished to the Nth area to be polished is obtained sequentially, and the optimal polishing trajectory of the polishing robot is generated based on the position information of the first area to be polished to the Nth area to be polished.

2. The trajectory planning method for a wheel hub grinding robot based on 3D vision curve technology according to claim 1, characterized in that, The target wheel hub image is acquired at multiple preset location nodes, and the actual 3D surface map of the target wheel hub is obtained based on the wheel hub image. Specifically: The wheel hub images of the target wheel hub are acquired at multiple preset location nodes. Based on the Canny edge detection algorithm, feature extraction processing is performed on each wheel hub image to obtain the edge points of the wheel hub in each wheel hub image. The SIFT algorithm is used to obtain the descriptors of the edge points of the wheel hub in each wheel hub image, and the Euclidean distance between the descriptors of the edge points of the wheel hub in different wheel hub images is calculated; the edge points between the wheel hubs in different wheel hub images with Euclidean distances less than the preset Euclidean distance are matched to obtain several pairs of matching points; Randomly select a preset pair of matching point pairs from several pairs of matching point pairs as sample pairs, use these randomly selected sample pairs to build a geometric model, and calculate the distance between the actual position and the predicted position of each pair of matching point pairs in the geometric model. If the distance between the actual position and the predicted position of a matching point pair is greater than a preset threshold, the matching point pair is marked as an outside point; if the distance between the actual position and the predicted position of a matching point pair is not greater than the preset threshold, the matching point pair is marked as an inside point. Remove the matching point pairs marked as outside points to eliminate incorrect matching point pairs and retain the correct matching point pairs to obtain the filtered matching point pairs, and generate the matching results based on the filtered matching point pairs; Based on the matching results, the edge points of the wheel hubs in each wheel hub image are matched and fused to obtain an edge point model. Based on the edge point model, the actual 3D surface map of the target wheel hub is generated.

3. The trajectory planning method for a wheel hub grinding robot based on 3D vision curve technology according to claim 1, characterized in that, The actual 3D surface image is compared with the preset 3D surface image to obtain several areas to be polished, and a model image of each area to be polished in the target wheel hub is generated, specifically as follows: A three-dimensional spatial coordinate system is constructed, and the actual 3D surface map and the preset 3D surface map are imported into the three-dimensional spatial coordinate system. Based on the ICP algorithm, the actual 3D surface map and the preset 3D surface map are aligned in the three-dimensional spatial coordinate system. After alignment, the actual 3D surface map is discretized using a discretization method to obtain several first discrete points, and the three-dimensional coordinate values ​​of each first discrete point are obtained; and the preset 3D surface map is discretized to obtain several second discrete points, and the three-dimensional coordinate values ​​of each second discrete point are obtained. In the three-dimensional spatial coordinate system, the Mahalanobis distance between the first discrete point and the second discrete point at each corresponding position is calculated based on the corresponding three-dimensional coordinate values. If the Mahalanobis distance between a first discrete point and a second discrete point at a certain corresponding position is greater than a preset Mahalanobis distance, then the first discrete point and the second discrete point at that corresponding position are marked as non-coincident points; if the Mahalanobis distance between a first discrete point and a second discrete point at a certain corresponding position is not greater than a preset Mahalanobis distance, then the first discrete point and the second discrete point at that corresponding position are marked as coincident points. Based on the non-overlapping areas formed by the non-overlapping points between the actual 3D surface map and the preset 3D surface map, several independent non-overlapping areas are obtained, and each independent non-overlapping area is marked as a polishing area, and a polishing area model diagram of the polishing area is obtained.

4. The trajectory planning method for a wheel hub grinding robot based on 3D vision curve technology according to claim 1, characterized in that, It also includes the following steps: If two or more of the calculated correlation coefficients are equal, then the positional distance between the remaining areas to be polished with equal correlation coefficients and the previously planned areas to be polished is obtained, resulting in several positional distances. Sort the distances of each location from smallest to largest. Mark the remaining area to be polished corresponding to the smallest location distance as the Nth area to be polished, the remaining area to be polished corresponding to the second smallest location distance as the N+1th area to be polished, and so on, until the N+Nth area to be polished is obtained, until all remaining areas to be polished with equal correlation coefficients are planned. The position information of the first area to be polished to the N+Nth area to be polished is obtained sequentially, and the optimal polishing trajectory of the polishing robot is generated based on the position information of the first area to be polished to the N+Nth area to be polished.

5. A trajectory planning system for a wheel hub grinding robot based on 3D vision curve technology, characterized in that, The wheel hub grinding robot trajectory planning system includes a memory and a processor. The memory stores a wheel hub grinding robot trajectory planning method program. When the wheel hub grinding robot trajectory planning method program is executed by the processor, the steps of the wheel hub grinding robot trajectory planning method as described in any one of claims 1 to 4 are implemented.

Images