Injection mold wear detection system for automobile part processing

Abstract

The invention belongs to the technical field of image processing, and particularly relates to an injection mold wear detection system for automobile part processing, and the system comprises an injection mold data acquisition module which is used for obtaining point cloud data of an injection mold at a plurality of sampling moments; the injection mold surface model building module is used for building an initial surface model of the injection mold according to the point cloud data of the injection mold and obtaining a surface model of the injection mold at each sampling moment; the injection mold easy-to-wear area acquisition module is used for acquiring an easy-to-wear index of each data point and an easy-to-wear area of the injection mold; the wear area and corresponding wear degree acquisition module is used for acquiring the wear degree and wear direction of each data point at each sampling moment and acquiring the wear area of the injection mold; and obtaining the wear degree of each current wear area of the injection mold. According to the invention, the accuracy and efficiency of wear detection of the injection mold for automobile part processing are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing. More specifically, the present invention relates to an injection mold wear detection system for automobile part processing. Background Art

[0002] Plastics are used in the production of automobile parts due to their lightweight, durability, and low cost. For example, skeleton parts made of glass fiber-reinforced polypropylene, door panels made of polypropylene, and headlight housings made of high-temperature polycarbonate. These automobile parts are mass-produced through processes such as injecting injection molding raw materials into an injection mold, molding, cooling, and demolding. Due to reasons such as an increase in the number of stamping times, different stresses on different positions of the mold, and thermal fatigue, the injection mold will experience wear during the production process. The worn positions will affect the quality of subsequent injection-molded parts and cause economic losses. Therefore, it is necessary to monitor the precision of the injection mold in real time so that the staff can repair it in a timely and accurate manner, avoiding the failure and scrapping of the injection mold and the continuous loss of injection-molded parts.

[0003] In order to obtain the real-time wear condition of an injection mold, machine vision technology is currently commonly used to monitor the mold surface. In related technologies, for example, a Chinese patent document with the authorization announcement number CN104626492B discloses an injection molding processing monitoring and detection system based on machine vision, which discloses obtaining the cavity image of an injection mold in real time through an optical device and comparing it with a marked cavity image for the detection of the injection mold. However, the distance between the moving mold and the fixed mold of the injection mold is small, resulting in insufficient presentation of the details of the injection mold in the cavity image, so the detection effect is not good. A Chinese patent document with the authorization announcement number CN105414247B discloses a mold wear detection device for a numerically controlled turret punch press, which discloses detecting the mold wear condition by installing a turntable and setting reserved holes, ensuring the presentation of cavity details.

[0004] However, collecting images from multiple angles will result in repeated detection of the same position of the injection mold, causing waste of resources and unable to form an overall wear model of the injection mold. And if the overall image of the injection mold is obtained by real-time stitching of images and then compared with the marked cavity image, it will affect the real-time performance of the injection mold wear detection and reduce the detection efficiency of the injection mold. Summary of the Invention

[0005] To solve the above technical problems of insufficient accuracy and real-time performance in the wear detection of injection molds, the present invention provides an injection mold wear detection system for automobile part processing, and the system includes the following modules: An injection mold data acquisition module is used to obtain the point cloud data of the injection mold at several sampling moments. The point cloud data includes several data points and their corresponding three-dimensional coordinates in the mold coordinate system. An injection mold surface model establishment module is used to construct the initial surface model of the injection mold according to the point cloud data of the injection mold, and to obtain the surface models of the injection mold at each sampling moment. An injection mold easily worn area acquisition module is used to obtain the easily worn index of each data point according to the coordinates and included angles of each data point in the initial surface model of the injection mold. The easily worn area of the injection mold is obtained according to the easily worn index and gray value of each data point. A worn area and corresponding wear degree acquisition module is used to obtain the wear degree of each data point at each sampling moment and the wear direction of each data point at each sampling moment according to the differences between the surface models of the injection mold at each sampling moment. The worn area of the injection mold is obtained according to the wear degree and wear direction of the neighboring data points of the data point. The wear degree of each current worn area of the injection mold is obtained according to the coordinates of the data points in the worn area of the injection mold at each sampling moment and the distance between the worn area and the easily worn area of the injection mold.

[0006] By screening the pixel point data obtained by the depth camera, the present invention constitutes the point cloud data of the injection mold, reduces the calculation amount of the surface analysis of the injection mold, and improves the efficiency of the wear detection of the injection mold. The present invention obtains the data of the continuous displacement of the surface data points of the injection mold during the operation of the injection mold, can obtain the wear changes at each position on the surface of the injection mold, can repair the injection mold in time, reduce economic losses, and ensure the real-time performance of the wear detection of the injection mold.

[0007] Preferably, the constructing the initial surface model of the injection mold according to the point cloud data of the injection mold and obtaining the surface models of the injection mold at each sampling moment includes: denoting any data point at the first sampling moment as the target data point, drawing a ray from the origin of the mold coordinate system in the direction of the target data point, and denoting the data point closest to the origin on this ray as the type-I data point. Using the Delaunay triangulation method to generate a triangular mesh for all type-I data points, which is denoted as the initial surface model of the injection mold. For the point cloud at the i-th sampling moment, screening out the data points with three-dimensional coordinates different from those of all the point clouds at the first sampling moment, which are denoted as the type-II data points at the i-th sampling moment. The initial surface model of the injection mold and all the type-II data points at the i-th sampling moment constitute the surface model of the injection mold at the i-th sampling moment.

[0008] The present invention reconstructs the surface of the injection mold by using the Delaunay triangulation method for the point cloud data, and obtains the initial surface model of the injection mold, avoiding the inability to perform local and overall analysis of the surface of the injection mold for discrete point cloud data.

[0009] Preferably, the easily worn index of each data point satisfies the expression: ; In the formula, represents the wear index of the c-th data point; represents the number of included angles of all adjacent triangles in the triangle formed by the c-th data point and all adjacent data points; , represent the a-th and b-th included angles of the adjacent triangles formed by the c-th data point and all adjacent data points; represents the number of adjacent data points of the c-th data point; represents the distance between the h-th adjacent data point of the c-th data point and the adjacent plane of the c-th data point; represents the normalization function.

[0010] Preferably, the obtaining of the easily worn area of the injection mold includes: taking any triangle as the target triangle, calculating the possibility that the target triangle belongs to the easily worn area based on the wear index and gray value of the data points included in the target triangle; if the possibility that the target triangle belongs to the easily worn area is greater than the first threshold, setting the target triangle as an easily worn area; if the possibility that the adjacent triangle of the easily worn area belongs to the easily worn area is greater than the first threshold, incorporating the adjacent triangle of the easily worn area into the easily worn area until the possibility that all adjacent triangles of the easily worn area belong to the easily worn area is not greater than the first threshold.

[0011] The present invention obtains more easily worn data points through the position and gray level of the data points, enabling the staff to make key references and improving the efficiency of wear detection of the injection mold.

[0012] Preferably, the possibility that the target triangle belongs to the easily worn area satisfies the expression: ; In the formula, represents the possibility that the target triangle belongs to the easily worn area; represents the average value of the wear indexes of the data points included in the target triangle; represents the range size of the gray values of the data points included in the target triangle; represents the number of data points included in the target triangle; represents the normalization function.

[0013] Preferably, the obtaining of the wear degree of each data point at each sampling moment includes: denoting the k-th data point of the surface model at the i-th sampling moment as , and denoting the straight line connected to the origin of the mold coordinate system as ; obtaining Among the data points from the first sampling moment to the (i - 1)-th sampling moment, the data point with the closest time distance to is denoted as the pre-wear data point of the k-th data point of the surface model at the i-th sampling moment, denoted as ; Obtain the sampling moment; Take the Euclidean distance between and , and compare it with the time distance between the sampling moment of and the i-th sampling moment, denoted as the wear degree at the i-th sampling moment.

[0014] Preferably, the obtaining of the wear direction of each data point at each sampling moment includes: If the distance of from the origin is greater than the distance of from the origin, then the wear direction of at the i-th sampling moment is denoted as the positive wear direction; if the distance of from the origin is less than the distance of from the origin, then the wear direction of at the i-th sampling moment is denoted as the negative wear direction.

[0015] The present invention confirms the wear direction of the data points, avoids the deviation in the wear detection of the injection mold caused by plastic residues and the residues of mold wear, and improves the accuracy of the wear detection of the injection mold.

[0016] Preferably, the obtaining of the wear area of the injection mold includes: Obtain the data points with a wear degree greater than 0 among the data points from the first sampling moment to the (i - 1)-th sampling moment on , denoted as the relevant data points of ; Obtain the intersection points of and the initial surface model of the injection mold, denoted as the initial coordinates of ; Calculate the wear accumulation amount of each data point according to the coordinates of the data point and the relevant data points; Obtain the wear accumulation amount of ; Take any data point with a wear accumulation amount greater than 0 as a wear area. If there are data points with a wear accumulation amount greater than 0 in the neighborhood of the wear area, then incorporate them into the wear area to obtain several wear areas at the i-th sampling moment; The neighborhood of the wear area is a spherical space within a preset radius range of any data point in the wear area.

[0017] Preferably, the wear accumulation amount of the data point satisfies the expression: ; In the formula, represents the wear accumulation amount of ; represents The number of relevant data points; Represents a direction function. If The s-th relevant data point of is in the positive wear direction, then the value of is 1. If the s-th relevant data point of is in the negative wear direction, then

[0018] The present invention calculates the wear accumulation amount of the data point positions, analyzes from the overall change of the data points over time, avoids the problem of reduced accuracy of the wear detection of the injection mold caused by the deviation of the data at a single moment, and improves the robustness of the wear detection of the injection mold.

[0019] Preferably, the wear degrees of the current wear regions of the injection mold satisfy the expression: ; In the formula, represents the wear degree of the current v-th wear region; represents the ordinal number of the last sampling moment; represents the average wear accumulation amount of all data points in the v-th wear region at the I-th moment; represents the number of data points belonging to the easily worn region in the v-th wear region at the I-th moment; represents the number of data points in the v-th wear region at the I-th moment; represents the number of relevant data points of the u-th data point in the v-th wear region at the I-th moment; 、 represent the distances between the (w + 1)-th and w-th relevant data points, and between the w-th and (w - 1)-th relevant data points of the u-th data point in the v-th wear region at the I-th moment.

[0020] The beneficial effects of the present invention are as follows: (1) The present invention analyzes the worn regions of the injection mold by using a small number of data points with position changes in units of data points, improving the speed and efficiency of the wear detection of the injection mold; (2) In the present invention, the worn regions of the injection mold are obtained, and all the worn regions of the injection mold are detected, avoiding missed detection and misdetection, and improving the accuracy of the wear detection of the injection mold; (3) The present invention analyzes different structures and different wear speeds of the injection mold, calculates the wear degree of the wear region according to the position change speed of the data points, the distance from the easily worn region, and the wear accumulation amount, avoiding singularity and improving the applicable range of the wear detection of the injection mold. Description of the Drawings

[0021] Figure 1 It schematically shows the system block diagram of an injection mold wear detection system for automobile part processing in the present invention; Figure 2 It schematically shows the schematic diagram of the sampling device; Figure 3 It schematically shows the structural schematic diagram of the c-th data point and adjacent data points. Detailed implementation manners

[0022] The present invention provides an injection mold wear detection system for automobile part processing. As Figure 1 shown, an injection mold wear detection system for automobile part processing includes an injection mold data acquisition module 100, an injection mold surface model establishment module 200, an injection mold easily worn area acquisition module 300, and a worn area and corresponding wear degree acquisition module 400, which are specifically described below.

[0023] The injection mold data acquisition module 100 is used to acquire the point cloud data of the injection mold at several sampling moments.

[0024] It should be noted that in the process of injection molding of automobile parts, the moving mold and the stationary mold are closed, then the plastic enters the cavity, cools and solidifies to obtain the injection molded part, and then the ejection device ejects the injection molded part, and the picking device takes out the injection molded part. In order to improve production efficiency and reduce the movement range of the moving mold, the distance between the moving mold and the stationary mold is often relatively narrow, resulting in insufficient light and a small visible range. Therefore, the quality of the surface image of the injection mold collected by the RGB camera is not good. Considering that automobile parts often have complex structures, there are concave and convex changes on the surface of the injection mold, and the distances between different positions on the surface and the camera lens are different. Therefore, the present invention uses a depth camera to obtain the surface image of the injection mold. In addition, since the surface image of the injection mold obtained by the depth camera contains a large number of redundant pixel points, by screening some pixel points as the point cloud, it can represent the surface of the injection mold and reduce the calculation amount.

[0025] Specifically, a sampling device is installed between the moving mold and the stationary mold. The schematic diagram of the sampling device is as Figure 2 , and the sampling device includes a plurality of industrial RGB-D cameras. The industrial RGB-D cameras are equidistantly installed on the bracket, and the bracket is fixed on the periphery of the injection mold and is parallel to both the moving mold and the stationary mold.

[0026] Record the time to produce an injection-molded part as a cycle. Denote the position where the moving mold reaches the farthest distance from the stationary mold as the starting point of the moving mold, and the position where the moving mold merges with the stationary mold as the ending point of the moving mold. A cycle includes the waiting time T1 of the moving mold at the starting point of the moving mold, the moving time T2 of the moving mold from the starting point to the ending point, the waiting time T3 of the moving mold at the ending point for mold closing with the stationary mold, and the moving time T4 of the moving mold from the ending point to the starting point. Denote the start moment of the waiting time T1 of each cycle as the sampling moment. Sample the injection mold through a sampling device once every other cycle to obtain a number of pixel points at a sampling moment. Each pixel point includes the R value, G value, B value of the pixel point in the RGB color space, the two-dimensional coordinates of the pixel point in the image plane, and the depth relative to the industrial RGB-D camera.

[0027] Combine the two-dimensional coordinates of the pixel points with the depth through the camera internal parameters to convert them into spatial coordinates in the three-dimensional space, forming a point cloud.

[0028] Due to the different positions of the industrial RGB-D cameras, the spatial coordinates of the point clouds are not unified. Therefore, set up a mold coordinate system. Set the centroid of the bracket as the origin of the mold coordinate system, set the direction parallel to the bracket and vertically upward as the positive direction of the Z axis, set the direction perpendicular to the stationary mold and facing the stationary mold as the positive direction of the Y axis, and set the X axis perpendicular to the YOZ plane. A vector is formed from any industrial RGB-D camera to the origin of the mold coordinate system, which is the deviation vector of the industrial RGB-D camera. Move all the pixel points obtained by the industrial RGB-D camera by the deviation vector of the industrial RGB-D camera to obtain the point cloud in the mold coordinate system.

[0029] During the real-time injection production process, obtain the point cloud data at a number of sampling moments. The point cloud data includes a number of data points, three-dimensional coordinates in the mold coordinate system, and gray values.

[0030] So far, the point cloud data of the injection mold at a number of sampling moments has been obtained.

[0031] The injection mold surface model building module 200 constructs the initial surface model of the injection mold according to the point cloud data of the injection mold, and obtains the surface models of the injection mold at each sampling moment.

[0032] It should be noted that discrete point clouds cannot directly represent continuous surfaces, while triangulation forms a triangular mesh by connecting adjacent points and establishes a topological relationship, which can realize the continuous modeling of the injection mold surface and retain the features of the injection mold. Delaunay triangulation can significantly improve the mesh quality of the point cloud triangular mesh due to its characteristics of maximizing the minimum angle and result uniqueness. Therefore, the present invention uses Delaunay triangulation to establish the triangular mesh of the injection mold surface. At the same time, in order to avoid point cloud redundancy, it is necessary to delete the point cloud.

[0033] It should be further noted that with the injection molding production, the wear change on the surface of the injection mold is a relatively long process. The average service life of automotive injection molds is 120,000 - 180,000 molding cycles. If a triangular mesh is established for each part produced, it will be a huge computational redundancy. Considering that the wear and defects on the surface of the injection mold change gradually and the change amount is small, the present invention considers screening the updated point cloud and only retaining the changed data points, which can retain the features for analyzing the wear of the injection mold to the greatest extent and improve the efficiency of analyzing the wear of the injection mold.

[0034] Specifically, for the point cloud at the first sampling moment, any data point is denoted as the target data point. A ray is drawn from the origin of the mold coordinate system in the direction of the target data point, and the data point on this ray that is closest to the origin is denoted as the type-I data point. The type-I data points in all data point directions are obtained; the Delaunay triangulation method is used for all type-I data points to generate a triangular mesh, which is denoted as the initial surface model of the injection mold.

[0035] For the point cloud at the i-th sampling moment, the data points with three-dimensional coordinates different from those of all the point clouds at the first sampling moment are screened out and denoted as the type-II data points at the i-th sampling moment. The initial surface model of the injection mold and all the type-II data points at the i-th sampling moment constitute the surface model of the injection mold at the i-th sampling moment.

[0036] Thus far, the initial surface model of the injection mold and the surface models at each sampling moment have been obtained.

[0037] The easily worn area acquisition module 300 of the injection mold is used to obtain the wear index of each data point according to the coordinates and the included angles formed by the data points in the initial surface model of the injection mold; and obtain the easily worn area of the injection mold according to the wear index and the gray value of each data point.

[0038] It should be noted that the structure of automotive parts is relatively complex, so the surface structure of the injection mold is also relatively complex. Some structures with acute angles and drastic concave-convex changes are also the structures that require high-precision accuracy for automotive parts. However, these structures are prone to stress concentration, resulting in material fatigue. These structures need to be monitored key points. Therefore, the present invention screens the easily worn areas.

[0039] It should be further noted that in the initial surface model of the injection mold, the structure of each triangle in the triangular mesh reflects the surface structure of the injection mold. For example, the smaller the included angle between two adjacent triangles, the greater the corresponding structural change, and the closer the included angle is to 180 degrees, the neater the corresponding structure. Therefore, the easily worn area can be determined according to the initial surface model of the injection mold.

[0040] Specifically, for the initial surface model of the injection mold, for the c-th data point, obtain the data points sharing an edge with the c-th data point, denoted as the adjacent data points of the c-th data point. The c-th data point and all its adjacent data points form a spatially irregular polygon.

[0041] Such as Figure 3 is a schematic structural diagram of the c-th data point and its adjacent data points, where is the c-th data point, 、 、 、 、 are all the adjacent data points of the c-th data point, where 、 and are coplanar, 、 have a certain distance from the plane There is a certain angle between any two adjacent triangles in Figure 3 . It should be noted that there are two angles for adjacent triangles. The angle with the opening towards the positive Z-axis direction is denoted as the angle of the adjacent triangle. The smaller the angles of these adjacent triangles and the worse the coplanarity of the data points, the higher the wear index of the c-th data point.

[0042] Preferably, according to the coordinates of each data point in the initial surface model of the injection mold, determine the wear index of each data point: Obtain the triangles formed by the c-th data point and all its adjacent data points, and obtain the angles between any two adjacent triangles; The three adjacent data points closest to the c-th data point form the adjacent plane of the c-th data point, and obtain the distances between all the adjacent data points of the c-th data point and the adjacent plane of the c-th data point; The wear index of the c-th data point satisfies the expression: ; In the formula, represents the wear index of the c-th data point; represents the number of angles of the adjacent triangles formed by the c-th data point and all its adjacent data points; 、 represent the a-th and b-th angles of the adjacent triangles formed by the c-th data point and all its adjacent data points; represents the number of adjacent data points of the c-th data point; represents the distance between the h-th adjacent data point of the c-th data point and the adjacent plane of the c-th data point; represents the normalization function.

[0043] In the formula, It represents the difference between the a-th and b-th included angles of the adjacent triangles formed by the c-th data point and all its adjacent data points. It represents the sum of the differences between any two included angles of the adjacent triangles formed by the c-th data point and all its adjacent data points, indicating the concavity and convexity of the c-th data point. The larger this value is, the greater the concavity and convexity of the c-th data point, the more prominent the structure, and thus the larger the wear susceptibility index. It represents the average distance from all adjacent data points of the c-th data point to the adjacent plane of the c-th data point. This value reflects the three-dimensionality of the spatial irregular polygon formed by the c-th data point and all its adjacent data points. The larger this value is, the more three-dimensional the position of the c-th data point, the more prominent the structure, and thus the larger the wear susceptibility index. 、 Normalize the two parts respectively, avoiding the situation that the wear susceptibility index of the c-th data point is only affected by one part of the data due to excessive numerical deviation between the two parts, which affects the accuracy of the conclusion.

[0044] Thus, the wear susceptibility index of each data point is obtained.

[0045] It should be noted that the data points of the easily worn structure do not appear alone. To obtain the complete easily worn area, clustering needs to be performed based on the wear susceptibility index of the data points. The overall wear susceptibility index of the easily worn area is relatively high and the gray-scale distribution is uneven. Therefore, the present invention obtains the easily worn area of the injection mold according to the wear susceptibility index and gray-scale value of each data point.

[0046] Preferably, obtaining the easily worn area of the injection mold includes: Taking any triangle as the target triangle, based on the wear susceptibility index and gray-scale value of the data points included in the target triangle, calculate the possibility that the target triangle belongs to the easily worn area: ; In the formula, represents the possibility that the target triangle belongs to the easily worn area; represents the average value of the wear susceptibility index of the data points included in the target triangle; represents the range size of the gray-scale value of the data points included in the target triangle; represents the number of data points included in the target triangle; represents the normalization function.

[0047] In the formula, represents the ratio of the range size of the gray-scale value of the data points included in the target triangle to the number of data points. This value reflects the gray-scale richness of the data points included in the target triangle. The larger this value is, the richer the gray-scale of the data points included in the target triangle, thus reflecting the uneven structure of the injection mold and the greater the possibility that the target triangle belongs to the easily worn area.

[0048] If the possibility that the target triangle belongs to the easily worn area is greater than the first threshold, the target triangle is set as an easily worn area; if the possibility that an adjacent triangle of the easily worn area belongs to the easily worn area is greater than the first threshold, the adjacent triangle of the easily worn area is incorporated into the easily worn area until the possibility that all adjacent triangles of the easily worn area belong to the easily worn area is not greater than the first threshold. It should be noted that the first threshold is set by the implementer according to the actual implementation situation. For example, the first threshold can be set to 0.5.

[0049] Thus, several easily worn areas of the injection mold are obtained.

[0050] The wear area and corresponding wear degree acquisition module 400 is used to obtain the wear degree and wear direction of each data point at each sampling moment according to the difference of the surface models of the injection mold at each sampling moment; obtain the wear area of the injection mold according to the wear degree and wear direction of the neighborhood data points of the data point; obtain the wear degree of each current wear area of the injection mold according to the coordinates of the data points in the wear area of the injection mold at each sampling moment and the distance between the wear area of the injection mold and the easily worn area.

[0051] It should be noted that wear is the small loss that appears on the surface structure of the injection mold. When the surface structure of the injection mold wears, the data points at the corresponding positions will move backward, which reflects the position after the surface of the injection mold wears. Therefore, the degree of backward movement of the data points can characterize the degree of surface wear of the injection mold. At the same time, the worn part will not disappear out of thin air. It may remain on the injection mold briefly and then fall off, or be fixed on the injection mold through high temperature, and due to the consistency of the injection mold structure, the backward movement of one side of a protruding structure may cause the forward movement of the other side. Although the other side is not worn, it will affect the manufacturing accuracy of the injection parts. In addition, there may be plastic residues on the injection mold for the injection parts, causing changes in the positions of the data points and interfering with the judgment of the wear of the injection mold. Therefore, the present invention analyzes the position changes of the data points. If the data points move forward and there are no data points moving backward in the neighborhood, it means that the data points are non-worn points. If the data points gradually move backward with the change of the sampling time, then the data points and the data points moving forward in the neighborhood together form a wear area that needs to be repaired by the staff.

[0052] It should be further noted that if the wear area appears in the easily worn area, the wear degree of the wear area will be greater, because the protruding structure in the easily worn area will have a greater impact on the injection parts due to minor wear, while the wear in the non-easily worn area is more integral. Although the wear area is larger, the wear degree is smaller. Therefore, the present invention adjusts the wear degree of the wear area in combination with the easily worn area of the injection mold.

[0053] Preferably, according to the differences in the surface models at each sampling moment of the injection mold, obtain the wear degree and wear direction of each data point at each sampling moment: Denote the k-th data point of the surface model at the i-th sampling moment as , and denote the straight line connecting it to the origin of the mold coordinate system as ; Obtain Among the data points from the 1st sampling moment to the (i - 1)-th sampling moment on , the data point with the closest time distance to is denoted as the data point before wear of the k-th data point of the surface model at the i-th sampling moment, denoted as ; Obtain the sampling moment of ; Divide the Euclidean distance between and by the time distance between the sampling moment of and the i-th sampling moment, and denote it as

[0054] If the distance from to the origin is greater than the distance from to the origin, then denote the wear direction of at the i-th sampling moment as the positive wear direction. If the distance from to the origin is less than the distance from to the origin, then denote the wear direction of at the i-th sampling moment as the negative wear direction.

[0055] Preferably, according to the wear degree and wear direction of the neighborhood data points of the data point, obtain the wear area of the injection mold: Obtain Among the data points from the 1st sampling moment to the (i - 1)-th sampling moment on , the data points with a wear degree greater than 0 are denoted as the relevant data points of; Obtain the intersection point of and the initial surface model of the injection mold, denoted as the initial coordinates of. It should be noted that has worn from the initial coordinates to the coordinates of

[0056] The wear accumulation of satisfies the expression: ; In the formula, represents the wear accumulation of; represents the number of relevant data points of; Represents a direction function. If the s-th relevant data point is in the positive wear direction, then the value is 1. If the s-th relevant data point is in the negative wear direction, then the value is -1; represents the wear degree of the s-th relevant data point of

[0057] It should be noted that, if the wear accumulation amount of is greater than 0, it means that there is wear on the surface of the injection mold, and it forms a wear area with the data points with wear in the neighborhood.

[0058] Cluster the data points with wear accumulation amount greater than 0 at the i-th sampling moment to obtain the wear area at the i-th sampling moment: Take any data point with wear accumulation amount greater than 0 as a wear area. If there are data points with wear accumulation amount greater than 0 in the neighborhood of the wear area, then incorporate them into the wear area to obtain several wear areas at the i-th sampling moment; the neighborhood of the wear area is the spherical space within the preset radius of any data point in the wear area. It should be noted that the preset radius is set by the implementer according to the actual implementation situation. For example, the preset radius can be set to 1 centimeter.

[0059] Preferably, according to the coordinates of the data points in the wear area of the injection mold at each sampling moment, and the distance between the wear area of the injection mold and the easily worn area, obtain the wear degree of each wear area of the current injection mold: Obtain all the data points and the corresponding all relevant data points of each wear area at the last sampling moment; For the u-th data point of the v-th wear area at the last sampling moment, connect it to the origin. If the connection intersects with the easily worn area of the injection mold, then mark this data point as a data point belonging to the easily worn area; obtain the data points belonging to the easily worn area of each wear area at the last sampling moment; The wear degree of any current wear area satisfies the expression: ; In the formula, represents the wear degree of the current v-th wear area; represents the ordinal number of the last sampling moment; represents the average wear accumulation amount of all the data points of the v-th wear area at the I-th moment; represents the number of data points belonging to the easily worn area in the v-th wear area at the I-th moment; represents the number of data points in the v-th wear area at the I-th moment; Denote the number of related data points of the \(u\)-th data point in the \(v\)-th wear area at the \(I\)-th moment; and Denote the distances between the \((w + 1)\)-th and \(w\)-th related data points, and between the \(w\)-th and \((w - 1)\)-th related data points of the \(u\)-th data point in the \(v\)-th wear area at the \(I\)-th moment.

[0060] In the formula, The larger it is, the greater the overall wear of the \(v\)-th wear area at the last sampling moment, so the greater the wear degree; Denote the proportion of data points belonging to the easily worn area in the \(v\)-th wear area at the last sampling moment. The larger this value is, the more likely the \(v\)-th wear area is to be worn, and the greater the impact of the wear, so the greater the wear degree; Denote the sum of the distance ratios of adjacent related data points of all data points in the \(v\)-th wear area at the last sampling moment. This value reflects the wear speed of the wear area. The larger this value is, the faster the current wear area wears, indicating that the wear degree of the \(v\)-th wear area is more serious.

[0061] Thus, the wear degree of each current wear area is obtained.

[0062] Set the second threshold and the third threshold. When the wear degree of the wear area is greater than the second threshold, injection molding needs to be stopped and the wear area of the injection mold needs to be repaired; when the wear degree of the wear area is greater than the third threshold, it means that the injection mold has reached its service life and needs to be replaced. It should be noted that the second threshold and the third threshold are set by the implementer according to the actual implementation situation. For example, the second threshold can be set to 0.5 and the third threshold can be set to 0.7.

[0063] Although this specification has shown and described multiple embodiments of the present invention, it is obvious to those skilled in the art that such embodiments are provided only by way of example. Those skilled in the art will think of many changes, alterations, and alternative ways without departing from the spirit and idea of the present invention. It should be understood that various alternative solutions to the embodiments of the present invention described herein can be adopted in the process of practicing the present invention.

Claims

1. A wear detection system for injection molds used in automobile parts processing, characterized in that: include: The injection mold data acquisition module is used to obtain the point cloud data of the injection mold at several sampling moments. The point cloud data includes several three-dimensional data points and corresponding grayscale values ​​in the mold coordinate system; An injection mold surface model building module is used to build an initial surface model of the injection mold according to the point cloud data of the injection mold, and to obtain the surface model of the injection mold at each sampling moment; The injection mold wear-prone area acquisition module is used to obtain the wear-prone index of each data point according to the coordinates and the angle formed by each data point in the initial surface model of the injection mold; and obtain the wear-prone area of ​​the injection mold according to the wear-prone index and gray value of each data point; The wear area and corresponding wear degree acquisition module is used to obtain the wear degree of each data point at each sampling moment and the wear direction of each data point at each sampling moment according to the difference in the surface model of the injection mold at each sampling moment; obtain the wear area of ​​the injection mold according to the wear degree and wear direction of the neighborhood data points of the data point; obtain the wear degree of each current wear area of ​​the injection mold according to the coordinates of the data points in the wear area of ​​the injection mold at each sampling moment and the distance between the wear area of ​​the injection mold and the easy-to-wear area.

2. The injection mold wear detection system for automobile parts processing according to claim 1, characterized in that: The step of constructing an initial surface model of the injection mold according to the point cloud data of the injection mold, and obtaining the surface model of the injection mold at each sampling moment, comprises: Any data point at the first sampling moment is recorded as the target data point, and a ray is drawn from the origin of the mold coordinate system to the target data point. The data point on the ray that is closest to the origin is recorded as a Class I data point. The Delaunay triangulation method is used to generate a triangulated network for all Class I data points, which is recorded as the initial surface model of the injection mold. For the point cloud at the i-th sampling moment, the data points with different three-dimensional coordinates from all point clouds at the first sampling moment are screened out and recorded as Class II data points at the i-th sampling moment. The initial surface model of the injection mold and all Class II data points at the i-th sampling moment constitute the surface model of the injection mold at the i-th sampling moment.

3. The injection mold wear detection system for automobile parts processing according to claim 1, characterized in that: The wear index of each data point satisfies the expression: ； In the formula, represents the wear index of the cth data point; Indicates the number of angles between all adjacent triangles in the triangle formed by the cth data point and all adjacent data points; , Represents the ath and bth included angles of adjacent triangles formed by the cth data point and all adjacent data points; Represents the number of adjacent data points of the cth data point; Represents the distance between the hth adjacent data point of the cth data point and the adjacent plane of the cth data point; Represents the normalization function.

4. The injection mold wear detection system for automobile parts processing according to claim 1, characterized in that: The method of obtaining the easily-worn area of ​​the injection mold comprises: Taking any triangle as the target triangle, based on the wear index and gray value of the data points contained in the target triangle, the possibility of the target triangle belonging to the wear-prone area is calculated; If the possibility that the target triangle belongs to the easy-to-wear area is greater than a first threshold, the target triangle is set as an easy-to-wear area; if the possibility that the adjacent triangles of the easy-to-wear area belong to the easy-to-wear area is greater than the first threshold, the adjacent triangles of the easy-to-wear area are included in the easy-to-wear area, until the possibility that all adjacent triangles of the easy-to-wear area belong to the easy-to-wear area is no greater than the first threshold.

5. The injection mold wear detection system for automobile parts processing according to claim 4, characterized in that: The probability that the target triangle belongs to the wear-prone area satisfies the expression: ； In the formula, Indicates the probability that the target triangle belongs to the wear-prone area; Represents the mean value of the wear index of the data points contained in the target triangle; Indicates the range of grayscale values ​​of the data points contained in the target triangle; Indicates the number of data points contained in the target triangle; Represents the normalization function.

6. The injection mold wear detection system for automobile parts processing according to claim 1, characterized in that: The step of obtaining the degree of wear of each data point at each sampling moment includes: The kth data point of the surface model at the i-th sampling time is recorded as , the straight line connected to the origin of the mold coordinate system is recorded as ; Get Among the data points from the 1st sampling time to the i-1th sampling time, The data point closest to the time point is recorded as the data point before wear of the kth data point of the surface model at the i-th sampling moment, recorded as ; Get The sampling time of and The Euclidean distance of The time distance between the sampling time of the ith sampling time and the ith sampling time is compared, recorded as The degree of wear at the i-th sampling moment.

7. The injection mold wear detection system for automobile parts processing according to claim 6, characterized in that: The obtaining of the wear direction of each data point at each sampling moment includes: like The distance from the origin is greater than The distance from the origin is The wear direction at the i-th sampling moment is recorded as the positive wear direction. The distance from the origin is less than The distance from the origin is The wear direction at the i-th sampling moment is recorded as the negative wear direction.

8. The injection mold wear detection system for automobile parts processing according to claim 6, characterized in that: The step of obtaining the wear area of ​​the injection mold comprises: Get Among the data points from the 1st sampling moment to the i-1th sampling moment, the data points with wear degree greater than 0 are recorded as ; Get the relevant data points of The intersection point with the initial surface model of the injection mold is denoted as The initial coordinates of the data point; calculate the accumulated wear of each data point according to the coordinates of the data point and related data points; obtain The accumulated amount of wear; Any data point with a wear accumulation greater than 0 is taken as a wear area. If there are data points with a wear accumulation greater than 0 in the neighborhood of the wear area, they are merged into the wear area to obtain several wear areas at the i-th sampling moment; the neighborhood of the wear area is a spherical space within a preset radius range of any data point in the wear area.

9. The injection mold wear detection system for automobile parts processing according to claim 8, characterized in that: The accumulated wear of the data points satisfies the expression: ； In the formula, express The accumulated amount of wear; express The number of relevant data points; represents the direction function, if The sth relevant data point of is in the positive wear direction, then The value of is 1, if The sth relevant data point of is in the negative wear direction, then The value of is -1; express The wear degree of the sth related data point.

10. The injection mold wear detection system for automobile parts processing according to claim 1, characterized in that: The wear degree of each wear area of ​​the injection mold currently satisfies the expression: ； In the formula, Indicates the wear degree of the current v-th wear area; Indicates the ordinal number of the last sampling moment; represents the average accumulated wear of all data points in the vth wear area at the Ith moment; represents the number of data points belonging to the easy-to-wear area in the vth wear area at the Ith moment; represents the number of data points of the vth wear area at the Ith moment; represents the number of relevant data points of the u-th data point in the v-th wear region at the i-th moment; , It represents the distance between the w+1th and wth related data points, and the distance between the wth and w-1th related data points of the uth data point in the vth wear area at the Ith moment.

Images