Surface flatness detection method and system for die pressing feeding and discharging machine

Abstract

The invention relates to the technical field of flatness detection, in particular to a surface flatness detection method and system for a mold pressing feeding and discharging machine. The method comprises the following steps: acquiring an initial defect index according to the difference of distances from point cloud data of blocks on the surface of the molding press to a reference plane, and correcting the initial defect index by combining the difference of overall distribution trends of the point cloud data of the blocks and the point cloud data of the surface of the molding press and the position distribution of the point cloud data of the blocks to obtain a corrected defect index; acquiring a fitting coefficient according to the dispersion degree of the correction defect index of each block in the preset direction, the variation trend of the overall distribution trend of the point cloud data of each block in the preset direction and the correction defect index, and performing plane fitting on the point cloud data by using the fitting coefficient, and detecting the flatness of the surface of the molding press according to the distance from the point cloud data to the fitting plane. According to the invention, the accuracy of flatness detection of the loading and unloading surfaces of the molding press is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of flatness detection, and specifically relates to a method and system for detecting the surface flatness of a die pressing loading and unloading machine. Background Art

[0002] A die press is a forming device specifically used for materials such as plastics, rubbers, ceramics, and glasses. By applying pressure and heat, it cures plastic materials to form the required shapes and sizes, and is widely used in fields such as automobiles, aviation, and construction. The flatness of the loading and unloading surface of the die press is a key factor affecting the quality of die-pressed products. Detecting the flatness of the loading and unloading surface of the die press is of great significance for ensuring product quality, improving production efficiency, and reducing equipment wear.

[0003] Existing methods usually use a laser beam to scan the loading and unloading surface of the die press, and construct a three-dimensional height map of the surface through the intensity or position change of the reflected light, so as to analyze the flatness of the loading and unloading surface. However, during the working process of the die press, vibrations may occur, resulting in fluctuations in the point cloud data generated by the vibrations on the loading and unloading surface. At the same time, the unevenness defects of the loading and unloading surface itself will also cause fluctuations in the point cloud data, resulting in the misidentification of the point cloud data generated by the normal vibrations of the device as unevenness defects, and thus the accuracy of detecting the flatness of the loading and unloading surface of the die press is relatively low. Summary of the Invention

[0004] In order to solve the technical problem that the point cloud data of normal vibrations during the operation of the die press reduces the accuracy of detecting the flatness of the loading and unloading surface of the die press, the purpose of the present invention is to provide a method and system for detecting the surface flatness of a die pressing loading and unloading machine. The specific technical solutions adopted are as follows: In the first aspect, an embodiment of the present invention provides a method for detecting the surface flatness of a die pressing loading and unloading machine, the method comprising: Obtaining point cloud data of the die press surface; dividing the point cloud data into a plurality of sub-blocks; Obtaining a reference plane for each sub-block, and obtaining an initial defect index for each sub-block according to the difference between the distances from the point cloud data of each sub-block to the reference plane; According to the difference between the overall distribution trend of the point cloud data of each sub-block and the overall distribution trend of the point cloud data of the die press surface, and the position distribution of the point cloud data of each sub-block, correcting the initial defect index to obtain a corrected defect index for each sub-block; According to the dispersion degree of the corrected defect indices of all sub-blocks in each preset direction for each sub-block, the change trend of the overall distribution trend of the point cloud data of all sub-blocks in all preset directions for each sub-block, and the corrected defect index of each sub-block, obtaining a fitting coefficient for the point cloud data of each sub-block; Perform plane fitting on the point cloud data on the surface of the molding press based on the fitting coefficients, and detect the flatness of the surface of the molding press according to the distance from the point cloud data on the surface of the molding press to the fitted plane.

[0005] Further, obtain the reference plane for each block, and obtain the initial defect index for each block according to the difference between the distances from the point cloud data of each block to the reference plane, including: Divide the point cloud data of each block into normal point cloud data and abnormal point cloud data; Perform plane fitting on the normal point cloud data of each block to obtain the reference plane for the corresponding block, and obtain the quartiles of the distances from each type of point cloud data of each block to the reference plane; Form an n-tuple from the quartiles corresponding to each type of point cloud data of each block, where n is the total number of quartiles corresponding to each type of point cloud data; Denote the distance between the n-tuples of the normal point cloud data and the abnormal point cloud data of each block as the distance difference index for each block; Obtain the interquartile range of the distances from the normal point cloud data of each block to the reference plane; According to the interquartile range and the distance difference index, obtain the initial defect index for each block; Both the interquartile range and the distance difference index have a positive correlation with the initial defect index.

[0006] Further, the obtaining of the corrected defect index for each block includes: Perform plane fitting on the point cloud data of each block respectively to obtain the second plane for the corresponding block, and perform plane fitting on the point cloud data on the surface of the molding press to obtain the overall plane; Obtain the similarity index between the normal vectors of the second plane and the overall plane for each block; Obtain the mean square error between all the point cloud data of each block and the corresponding point cloud data on the second plane, and denote it as the fitting error for each block; According to the similarity index and the fitting error, obtain the first correction coefficient for each block; The similarity index has a negative correlation with the first correction coefficient, and the fitting error has a positive correlation with the first correction coefficient; Use the sum value of the first correction coefficient and the constant 1 to perform weighted processing on the initial defect index of each block to obtain the corrected defect index for each block.

[0007] Further, the obtaining of the fitting coefficient of the point cloud data of each block includes: Obtain the dispersion index and the concentration index of the corrected defect indices of all blocks in each preset direction for each block; According to the dispersion index and the concentration index, obtain the normal vibration index for each block in each preset direction; Obtain the direction trend index of each block according to the change trend of the overall distribution trend of the point cloud data of all blocks in all preset directions for each block; According to the normal vibration index and the direction trend index, obtain the second correction coefficient of each block; the normal vibration index and the second correction coefficient are negatively correlated, and the direction trend index and the second correction coefficient are positively correlated; use the second correction coefficient to perform weighted processing on the correction defect index of each block to obtain the final defect index of each block; perform negative correlation and normalization processing on the final defect index to obtain the fitting coefficient of each point cloud data of each block.

[0008] Further, the obtaining the direction trend index of each block includes: Obtain the angles between the normal vectors of the second planes of each block and each coordinate axis of the coordinate system where the point cloud data is located, and record them as the analysis angles of each block in each dimension; one dimension corresponds to one of the coordinate axes; arrange the analysis angles of all blocks in each dimension in each preset direction for each block in sequence to obtain the angle sequence of each block in each preset direction in each dimension; Obtain the first-order difference sequence of the angle sequence, and record the proportion of the maximum value of the number of consecutive positive numbers and the number of consecutive negative numbers in the first-order difference sequence to the total number of elements in the angle sequence as the local trend index of each block in each preset direction in each dimension; select the maximum value among the local trend indices of each block in each preset direction in all dimensions as the overall trend index of each block in each preset direction; Take the sum of the overall trend indices of each block in all preset directions as the direction trend index of each block.

[0009] Further, the plane fitting the point cloud data on the surface of the molding press based on the fitting coefficient, and detecting the flatness of the surface of the molding press according to the distance from the point cloud data on the surface of the molding press to the fitted plane includes: Based on the fitting coefficient, use the weighted least squares method to perform plane fitting on the point cloud data on the surface of the molding press to obtain the final reference plane; Obtain the distance from the centroid of the second plane of each block to the final reference plane, and record it as the judgment distance of each block; Judge whether the judgment distances of all blocks on the surface of the molding press are all less than a preset threshold. If so, the flatness of the surface of the molding press is qualified; if not, the flatness of the surface of the molding press is unqualified.

[0010] Further, the dividing the point cloud data of each block into normal point cloud data and abnormal point cloud data includes: Perform plane fitting on the point cloud data of each block to obtain the first plane corresponding to each block; Obtain the distances between all the point cloud data of each block and the first plane, use the maximum inter-class variance method for the distances to obtain the division threshold, and respectively record the point cloud data with distances less than the division threshold as normal point cloud data, and the point cloud data with distances greater than or equal to the division threshold as abnormal point cloud data.

[0011] Further, the similarity index is the cosine similarity.

[0012] Further, the obtaining the discrete index and the concentration index of the corrected defect index of all blocks in each preset direction for each block includes: The discrete index is the variance of the corrected defect index of all blocks in each preset direction for each block, and the concentration index is the mean value of the corrected defect index of all blocks in each preset direction for each block.

[0013] In a second aspect, another embodiment of the present invention provides a surface flatness detection system for a molding loading and unloading machine, and the system includes: A data acquisition module, configured to obtain point cloud data of the surface of the molding press; divide the point cloud data into multiple blocks; A defect analysis module, configured to obtain a reference plane for each block, and obtain an initial defect index for each block according to the differences between the distances from the point cloud data of each block to the reference plane; A defect correction module, configured to correct the initial defect index according to the difference between the overall distribution trend of the point cloud data of each block and the overall distribution trend of the point cloud data of the surface of the molding press, and the position distribution of the point cloud data of each block, to obtain a corrected defect index for each block; A fitting coefficient analysis module, configured to obtain a fitting coefficient of the point cloud data of each block according to the degree of dispersion of the corrected defect index of all blocks in each preset direction for each block, the change trend of the overall distribution trend of the point cloud data of all blocks in all preset directions for each block, and the corrected defect index of each block; A flatness detection module, configured to perform plane fitting on the point cloud data of the surface of the molding press based on the fitting coefficient, and detect the flatness of the surface of the molding press according to the distances from the point cloud data of the surface of the molding press to the plane obtained by fitting.

[0014] The present invention has the following beneficial effects: In the embodiment of the present invention, the reference plane presents the overall distribution trend of the point cloud data with normal performance in the block, and the difference between the distances from the point cloud data of the block to the reference plane presents the degree of unevenness defect of the block, so as to obtain the initial defect index; the point cloud data of the unevenness defect of the device itself shows the characteristics of strong bending and jumping changes, while the fluctuations of the point cloud data generated by the vibration of the device have high consistency. The difference between the overall distribution trend of the point cloud data of the block and the point cloud data of the surface of the die press presents the fluctuation consistency of the point cloud data of the block. Combining the position distribution of the point cloud data of the block to correct the initial defect index to obtain the corrected defect index, so as to reduce the dependence of the analysis of the unevenness defect on the point cloud data generated by the vibration of the device; in actual situations, there may be a defect of the device itself where the surrounding is significantly bent but the middle is relatively flat. The direction of the overall distribution trend of the point cloud data of the block in a certain direction of the bending defect presents a trend change characteristic, and the unevenness defect characteristics of the block are relatively discrete, while the characteristics of the point cloud data generated by the vibration due to the consistent vibration intensity of the device are opposite to the above two characteristics; according to the dispersion degree of the corrected defect indexes of all blocks in the preset direction, and the change trend of the overall distribution trend of the point cloud data of all blocks in the preset direction, adaptively adjust the corrected defect index to obtain the fitting coefficient, avoiding misjudgment of large-range defects easily caused by only analyzing the unevenness defect of a single block, reducing the probability that the point cloud data generated by the vibration of the device is misidentified as the unevenness defect of the device itself, so that the fitting plane constructed based on the fitting coefficient can accurately represent the overall flatness level of the surface of the die press, and improving the accuracy of detecting the flatness of the surface of the die press according to the distance from the point cloud data to the fitting plane. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0016] Figure 1 It is a flowchart of the steps of a method for detecting the surface flatness of a die press loading and unloading machine provided by an embodiment of the present invention; Figure 2 It is a flowchart of the steps of a method for obtaining a fitting coefficient provided by an embodiment of the present invention; Figure 3 It is a system structure diagram of a system for detecting the surface flatness of a die press loading and unloading machine provided by an embodiment of the present invention; Figure 4 It is a schematic diagram of a computer device of a device for detecting the surface flatness of a die press loading and unloading machine provided by an embodiment of the present invention. Detailed implementation manners

[0017] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following combines the accompanying drawings and preferred embodiments to detail the specific implementation manners, structures, features, and effects of a surface flatness detection method and system for a die pressing loading and unloading machine according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0019] The following specifically describes the specific solution of a surface flatness detection method and system for a die pressing loading and unloading machine provided by the present invention with reference to the accompanying drawings.

[0020] Embodiment 1: The present invention proposes a surface flatness detection method for a die pressing loading and unloading machine. Please refer to Figure 1 , which shows a flowchart of the steps of a surface flatness detection method for a die pressing loading and unloading machine provided by an embodiment of the present invention. The method includes: Step S1: Obtain the point cloud data of the surface of the die press; divide the point cloud data into multiple blocks.

[0021] The loading and unloading of the die press usually refer to the upper and lower dies. During the process of the die press producing products, the workpiece is first placed at the designated position for unloading, and then the loading is pressed into the unloading to stamp out the required shape. Before the loading of the die press is pressed into the unloading, a laser scanner is used to perform laser scanning on the loading surface of the die press in the working state to obtain the point cloud data of the surface of the die press. This solution is applicable to the flatness detection of the loading and unloading surfaces of die presses that produce products with a flat surface.

[0022] The point cloud data belongs to three-dimensional coordinates. In this embodiment, the point cloud data at the lower left corner of the loading surface of the die press is selected as the coordinate origin, the long side direction of the surface of the die press or the workbench is the X-axis, the direction perpendicular to the X-axis direction and parallel to the ground is selected as the Y-axis, and the normal direction from the ground upward to the surface of the die press is the Z-axis to construct a three-dimensional coordinate system; the point cloud data of the surface of the die press is in this three-dimensional coordinate system.

[0023] It should be noted that in other embodiments of the present invention, lidar and ultrasonic sensors can also be used to obtain the point cloud data, and other origins and directions can also be selected to construct the three-dimensional coordinate system. The acquisition of the point cloud data is a well-known technology and will not be limited here.

[0024] The feeding surface area of the molding press is relatively large. To ensure the accuracy of flatness detection, the point cloud data of the molding press surface is divided into blocks. The specific method is as follows: The plane formed by the X-axis and the Y-axis in the three-dimensional coordinate system is evenly divided into square areas with a side length of 10 cm; for the point cloud data of the molding press surface, the point cloud data whose positions on both the X-axis and the Y-axis are within the same square area forms a block.

[0025] In other embodiments, the point cloud data of the molding press surface can also be imported into Integrated Computer Engineering and Manufacturing (ICEM) software, and the structured mesh division is performed using the block division strategy in the ICEM software to divide the point cloud data of the molding press surface into multiple blocks.

[0026] Step S2: Obtain the reference plane of each block, and based on the differences between the distances from the point cloud data of each block to the reference plane, obtain the initial defect index of each block.

[0027] Since the molding press is in a working state during the acquisition of point cloud data, vibrations will occur during the process of the feeding of the molding press being pressed into the discharging. The point cloud data generated by the vibrations of the device will cause unevenness defects on the feeding surface. At the same time, the unevenness existing on the feeding surface itself will also be manifested as unevenness defects. Then, the unevenness defects generated by the vibrations of the device will affect the flatness detection of the feeding surface. Therefore, it is necessary to distinguish whether the flatness abnormality of the feeding surface is a defect caused by the vibrations during the operation of the device or a defect manifested by the unevenness of the device itself.

[0028] If there are unevenness problems on the feeding surface of the molding press, such as bending or local deformation during the downward pressing process, etc., it will cause poor consistency in the point cloud data of the feeding surface. Then, the distances from the point cloud data at the uneven position to the overall plane formed by the point cloud data of the device surface are far and inconsistent. Therefore, if the distances from the point cloud data of the block to the reference plane are farther and more inconsistent, the greater the possibility that the block is a defect of the unevenness of the device surface itself, and the greater the degree of unevenness defect of the block. The differences between the distances from the point cloud data of the block to the reference plane can measure the degree of unevenness defect of the block, thereby obtaining the initial defect index.

[0029] Preferably, in some possible implementation manners of the embodiments of the present invention, the method for obtaining the initial defect index includes: dividing the point cloud data of each block into normal point cloud data and abnormal point cloud data; performing plane fitting on the normal point cloud data of each block to obtain the reference plane of the corresponding block, and obtaining the quartiles of the distances from each type of point cloud data of each block to the reference plane; forming an n-tuple from the quartiles corresponding to each type of point cloud data of each block, where n is the total number of quartiles corresponding to each type of point cloud data; recording the distance between the n-tuples of the normal point cloud data and the abnormal point cloud data of each block as the distance difference index of each block; obtaining the interquartile range of the distances from the normal point cloud data of each block to the reference plane; and obtaining the initial defect index of each block according to the interquartile range and the distance difference index.

[0030] The point cloud data of the unevenness defect usually shows large fluctuations in the point cloud, making the point cloud data show large jump characteristics, that is, the point cloud data of the block is discontinuous, weakening the plane characteristics. Therefore, before obtaining the reference plane, it is necessary to divide the point cloud data of the block, select the point cloud data with normal flatness in the device to construct the reference plane, so as to improve the accuracy of the unevenness defect analysis.

[0031] The method for dividing the point cloud data of the block in this embodiment is: performing plane fitting on the point cloud data of each block to obtain the first plane of the corresponding block; obtaining the distances between all the point cloud data of each block and the first plane, using the maximum inter-class variance method for the distances to obtain the division threshold, and respectively recording the point cloud data with a distance less than the division threshold as normal point cloud data, and the point cloud data with a distance greater than or equal to the division threshold as abnormal point cloud data.

[0032] In other embodiments, based on the distances from the point cloud data of each block to the first plane, the K-means clustering algorithm is used to cluster the point cloud data of each block to obtain two clusters, where the value of K is equal to the constant 2; obtaining the mean of the distances from the point cloud data in each cluster to the first plane, and respectively recording the point cloud data in the cluster corresponding to the minimum mean as normal point cloud data, and the point cloud data in the cluster corresponding to the maximum mean as abnormal point cloud data.

[0033] It should be noted that in this embodiment, the least squares method is used to perform plane fitting on the point cloud data of the block; in other embodiments, the Chebyshev method, the principal component analysis method, and the Random Sample Consensus (RANSAC) algorithm can also be used for plane fitting. The least squares method, the maximum inter-class variance method, and the K-means clustering algorithm are all well-known technologies to those skilled in the art and will not be elaborated here.

[0034] The point cloud data of the defective positions with uneven feeding surfaces of the molding press itself exhibits characteristics of strong bending and jumping changes. Although the point cloud data generated by the device vibration shows strong unevenness characteristics, the surface vibration of the device is consistent, making the fluctuations of the point cloud data generated by the device vibration in all sub-blocks have high consistency. Therefore, the greater the inconsistency between the abnormal point cloud data and the normal point cloud data of the sub-block in terms of the distance to the reference plane, the greater the possibility that the sub-block exhibits the defect of unevenness of the device itself, and the greater the degree of unevenness defect. The n-tuple reflects the fluctuation of the distance of each type of point cloud data to the reference plane. The larger the distance difference index, the greater the inconsistency between the normal point cloud data and the abnormal point cloud data of the sub-block in terms of the distance to the reference plane, and the greater the possibility that the sub-block exhibits the defect of unevenness of the device itself.

[0035] The interquartile range of the distance from the normal point cloud data of the sub-block to the reference plane reflects the concentration of the normal point cloud data distribution. If the interquartile range is smaller, the normal point cloud distribution of the sub-block is more concentrated, and the greater the possibility that the point cloud data of the sub-block is generated by the device vibration. On the contrary, the greater the possibility that the point cloud data of the sub-block is the defect of unevenness of the device itself, and the more obvious the unevenness defect.

[0036] In summary, both the interquartile range and the distance difference index have a positive correlation with the initial defect index. In the embodiments of the present invention, the product of the interquartile range of the distance from the normal point cloud data of each sub-block to the reference plane and the distance difference index is used as the initial defect index of each sub-block. In the embodiments of the present invention, the correlation between the interquartile range, the distance difference index, and the initial defect index can also be constructed through other basic mathematical operations, which will not be limited and elaborated here.

[0037] It should be noted that there are two types of point cloud data in this embodiment, namely normal point cloud data and abnormal point cloud data. The quartiles include the upper quartile and the lower quartile. Then, the n-tuple composed of the quartiles corresponding to each type of point cloud data of each sub-block is a binary tuple, where .

[0038] Step S3: According to the difference between the overall distribution trend of the point cloud data of each sub-block and the overall distribution trend of the point cloud data on the surface of the molding press, as well as the position distribution of the point cloud data of each sub-block, correct the initial defect index to obtain the corrected defect index of each sub-block.

[0039] Analyzing the initial defect index depends on the existence of point cloud data generated by device vibration in the sub-block. If the point cloud data of the sub-block are all defects of unevenness of the device itself, it is easy to make the division of normal point cloud data and abnormal point cloud data in step S2 inaccurate, resulting in a large error in the initial defect index. Therefore, it is necessary to correct the initial defect index.

[0040] Compared with the unevenness caused by the vibration of the device, the unevenness characteristics of the unevenness defect of the device itself are more obvious, making the position distribution of the point cloud data of the divided unevenness defect far from the overall distribution trend, and at the same time, the greater the difference between the overall distribution trend of the divided point cloud data and the overall distribution trend of the point cloud data on the surface of the molding press. Therefore, according to the difference between the overall distribution trend of the divided point cloud data and the overall distribution trend of the point cloud data on the surface of the molding press, and the position distribution of the divided point cloud data, the initial defect index is corrected to further improve the accuracy of the analysis of the divided unevenness defect.

[0041] Preferably, in some possible implementation manners of the embodiments of the present invention, the method for obtaining the corrected defect index includes: performing plane fitting on the point cloud data of each divided block to obtain a second plane corresponding to the divided block, and performing plane fitting on the point cloud data on the surface of the molding press to obtain an overall plane; obtaining a similarity index between the normal vectors of the second plane of each divided block and the overall plane; obtaining the mean square error between all the point cloud data of each divided block and the corresponding point cloud data on the second plane, which is denoted as the fitting error of each divided block; obtaining a first correction coefficient for each divided block according to the similarity index and the fitting error; both the similarity index and the fitting error are positively correlated with the first correction coefficient; using the sum value of the first correction coefficient and the constant 1 to perform weighted processing on the initial defect index of each divided block to obtain the corrected defect index of each divided block.

[0042] Since inaccurate division of the point cloud data of the divided block will cause errors in the initial defect index, the second plane obtained by performing plane fitting on all the point cloud data of the divided block in this embodiment is used for subsequent analysis. In this embodiment, the least squares method is used to perform plane fitting on the point cloud data of each divided block and the point cloud data on the surface of the molding press respectively.

[0043] For the point cloud data of each divided block, if the number of point cloud data belonging to the unevenness defect of the device itself is larger, the deformation presented by the point cloud data of the divided block is more obvious, then the fitting error when performing plane fitting on the point cloud data of the divided block is larger, and the fitting error presents the position distribution of the point cloud data of the divided block; at the same time, the second plane of the divided block is more affected by the unevenness defect on the surface of the device, and the local unevenness characteristics of the divided block are more obvious, making the difference between the direction of the second plane of the divided block and the direction of the surface of the device larger. Therefore, if the fitting error when performing plane fitting on the point cloud data of the divided block is larger, and the difference between the direction of the second plane of the divided block and the direction of the surface of the device is larger, then the possibility that the point cloud data of the divided block belongs to the unevenness defect of the device surface itself is larger, and the possibility of belonging to the point cloud data generated by the vibration of the device is smaller.

[0044] The normal vector of the second plane reflects the overall distribution trend of the segmented point cloud data, and the normal vector of the overall plane reflects the overall distribution trend of the point cloud data on the surface of the molding press. If the similarity index between the normal vector of the segmented second plane and the normal vector of the overall plane is smaller, it indicates that the direction difference between the overall distribution trend of the segmented point cloud data and the overall distribution trend of the device surface is larger, and the greater the possibility that the segmented point cloud data belongs to the unevenness defect of the device surface itself. If the fitting error is larger, it indicates that the unevenness characteristics of the segmented point cloud data are more obvious, and the greater the possibility that the segmented point cloud data belongs to the unevenness defect of the device surface itself; furthermore, the smaller the initial defect index of the segment is than the actual defect degree, the initial defect index of the segment should be adjusted upward, and the first correction coefficient should be larger. Therefore, the similarity index and the first correction coefficient are negatively correlated, and the fitting error and the first correction coefficient are positively correlated. In the embodiment of the present invention, a negative correlation mapping is performed on the similarity index between the normal vector of the second plane of each segment and the normal vector of the overall plane, and the product of the mapping result and the fitting error is normalized to obtain the first correction coefficient of each segment.

[0045] In the embodiment of the present invention, the correlation relationship between the similarity index, the fitting error, and the first correction coefficient can also be constructed through other basic mathematical operations, which will not be limited and elaborated here.

[0046] The fitting error in this embodiment is the mean square error between all the point cloud data of the segment and the corresponding point cloud data on the second plane, which is equal to the mean of the squares of the distances between all the point cloud data of the segment and the corresponding point cloud data on the second plane; in other embodiments, the mean square error can also be replaced by the root mean square error and the mean absolute error, etc., which will not be limited here.

[0047] In this embodiment, the cosine similarity between the normal vector of the second plane of the segment and the normal vector of the overall plane is used as the similarity index; in another embodiment, the Pearson correlation coefficient between the normal vector of the second plane of the segment and the normal vector of the overall plane is used as the similarity index; in other embodiments, the Euclidean distance or Manhattan distance between the normal vector of the second plane of the segment and the normal vector of the overall plane can also be negatively correlated and normalized to obtain the similarity index.

[0048] In a specific implementation manner of the embodiment of the present invention, the corrected defect index is expressed by the formula: In the formula, is the corrected defect index of the k-th segment on the surface of the molding press; is the first correction coefficient of the k-th segment on the surface of the molding press; is the initial defect index of the k-th segment on the surface of the molding press; is the fitting error of the k-th segment on the surface of the molding press; is the normal vector of the second plane of the k-th block on the surface of the molding press; is the normal vector of the overall plane; is the similarity index between the normal vectors of the second plane of the k-th block on the surface of the molding press and the overall plane; cos is the cosine function; Norm is the normalization function.

[0049] It should be noted that when the correction defect index is larger, the greater the possibility that the k-th block belongs to the defect of the uneven surface of the device itself, and the initial defect index of the k-th block is less than the actual defect degree. Using to perform weighted processing to increase the correction defect index ; when the correction defect index is smaller, the greater the possibility that the k-th block belongs to the point cloud data generated by the vibration of the device, and the initial defect index is closer to the actual defect degree. In this embodiment, by subtracting from the constant 1, a negative correlation mapping of is realized. Other embodiments can select other negative correlation mapping methods, which will not be exemplified here.

[0050] Step S4: According to the dispersion degree of the correction defect indexes of all blocks in each preset direction for each block, the change trend of the overall distribution trend of the point cloud data of all blocks in all preset directions for each block, and the correction defect index of each block, obtain the fitting coefficient of the point cloud data of each block.

[0051] Since the initial defect index and the correction defect index consider the uneven characteristics of the point cloud data of a single block and only analyze the local deformation generated during the working process of the feeding surface of the molding press, in actual situations, there may be defects where the surrounding is significantly curved but the middle is relatively flat, such as the bending generated during the large-scale molding pressing process, and the flat position in the middle belongs to the uneven defect but shows strong planar characteristics, resulting in misjudgment of large-scale defects when only analyzing the uneven defects of a single block.

[0052] The deformation caused by extrusion reflects the bending defect of the device through all sub-blocks in a certain direction; if there is a bending defect on the surface of the device, the direction of the overall distribution trend of the point cloud data of the sub-blocks in a certain direction on the surface of the device shows a trend change characteristic; if the bending defect on the surface of the device is caused by the vibration of the device, since the vibration intensity of all sub-blocks is the same, the distance from the point cloud data generated by the vibration of the device in different sub-blocks to its overall distribution trend remains the same, then the direction of the overall distribution trend of the point cloud data of all sub-blocks in a certain direction of the sub-blocks remains basically the same. At the same time, due to the strong unevenness characteristics around the bending defect generated during the large-scale pressing process but the strong planar characteristics at the flat middle position, that is, the discreteness degree of the corrected defect index of the sub-blocks in a certain direction is more obvious.

[0053] Therefore, the discreteness degree of the corrected defect index of all sub-blocks of each sub-block in each preset direction, and the change trend of the overall distribution trend of the point cloud data of all sub-blocks of each sub-block in all preset directions can be used to further adjust the corrected defect index, and improve the accuracy of the unevenness defect of the sub-block performance again, so as to improve the accuracy of the fitting coefficient.

[0054] Please refer to Figure 2 , which shows the step flow chart of a method for obtaining a fitting coefficient provided by an embodiment of the present invention. The method includes: Step S410: Obtain the discrete index and the concentration index of the corrected defect index of all sub-blocks of each sub-block in each preset direction; according to the discrete index and the concentration index, obtain the normal vibration index of each sub-block in each preset direction.

[0055] If the discrete index is smaller, the corrected defect indexes of all sub-blocks of each sub-block in each preset direction are closer, indicating that the possibility that the sub-blocks in each preset direction of each sub-block are defective due to the vibration of the device is greater, then the normal vibration index is greater; on the contrary, it indicates that the possibility that the sub-blocks in each preset direction of each sub-block are bent and defective due to extrusion is greater, then the normal vibration index is smaller.

[0056] It should be noted that variance, standard deviation, range, etc. can all reflect the discreteness degree of a set of data. In this embodiment, the variance of the corrected defect index of all sub-blocks of each sub-block in each preset direction is selected as the discrete index. In other embodiments, the variance can be replaced with indexes such as standard deviation and range to obtain the discrete index.

[0057] The centralized index reflects the overall level of the unevenness defect degree of all blocks in each preset direction for each block; since the unevenness defect degree caused by the device vibration is less than the unevenness defect degree of the device itself, the corrected defect index of the block at the uneven position caused by the device vibration is smaller. If the centralized index is smaller, the unevenness defect degree of all blocks in each preset direction for each block is smaller, indicating that the possibility that the block in each preset direction is defective due to the device vibration is greater, and then the normal vibration index is greater.

[0058] It should be noted that the mean, median, and mode can all reflect the centralized distribution trend of a set of data. In this embodiment, the mean of the corrected defect indexes of all blocks in each preset direction for each block is selected as the centralized index. In other embodiments, the mean can be replaced with indexes such as the median or mode to obtain the centralized index.

[0059] Therefore, the discrete index has a negative correlation with the centralized index and the normal vibration index. In the embodiment of the present invention, the product of the discrete index and the centralized index of each block in each preset direction is subjected to a negative correlation mapping to obtain the normal vibration index of each block in each preset direction. In this embodiment, an exponential function with the natural constant as the base is selected for the negative correlation mapping. The specific method is: first take the opposite number of the product of the discrete index and the centralized index of each block in each preset direction, and use this opposite number as the exponent of the exponential function with the natural constant as the base, so as to achieve the negative correlation mapping.

[0060] In this embodiment, the preset directions include: the directions parallel to the X-axis and Y-axis of the three-dimensional coordinate system in step S1, that is, the directions of the row and column where each block is located. The implementer can set it according to the specific situation.

[0061] Step S420: Obtain the direction trend index of each block according to the change trend of the overall distribution trend of the point cloud data of all blocks in all preset directions for each block.

[0062] Preferably, in some possible implementation manners of the embodiments of the present invention, the method for obtaining the direction trend index includes: obtaining the included angles between the normal vectors of the second planes of each sub-block and each coordinate axis of the coordinate system where the point cloud data is located, and recording them as the analysis angles of each sub-block in each dimension; one dimension corresponds to one coordinate axis; arranging the analysis angles of all sub-blocks in each dimension in each preset direction of each sub-block in sequence to obtain the angle sequence of each sub-block in each preset direction in each dimension; obtaining the first-order difference sequence of the angle sequence, and recording the proportion of the maximum value of the number of consecutive positive numbers and the number of consecutive negative numbers in the first-order difference sequence to the total number of elements in the angle sequence as the local trend index of each sub-block in each preset direction in each dimension; selecting the maximum value of the local trend indexes of each sub-block in each preset direction in all dimensions as the overall trend index of each sub-block in each preset direction; and taking the sum of the overall trend indexes of each sub-block in all preset directions as the direction trend index of each sub-block.

[0063] Due to the vibration of the device, the direction of the overall distribution trend of the point cloud data of all sub-blocks in a certain direction is basically maintained stable, while the bending generated on the feeding surface during the molding process causes the direction of the overall distribution trend of the point cloud data of the sub-blocks in a certain direction to show a trend change. For the convenience of analysis, the analysis angles of all sub-blocks in each dimension in each preset direction of each sub-block are arranged in the order of the sub-blocks in the preset direction to obtain the angle sequence.

[0064] If the uneven defect presents a bending feature, the direction of the overall change trend of the point cloud data of the sub-blocks in the preset direction shows a trend change, that is, the analysis angles in the angle sequence show a certain degree of monotonic change, so that the number of consecutive positive numbers and the number of consecutive negative numbers in the first-order difference sequence of the angle sequence are more. If the uneven defect is caused by the vibration generated during the operation of the device and there is no obvious direction feature in the fitting direction, making the analysis angles in the angle sequence show randomness, then there will be no large-range consecutive positive numbers or consecutive negative numbers in the first-order difference sequence of the analysis angles.

[0065] Both the continuous positive numbers and the continuous negative numbers in the first-order difference sequence of the angle sequence can reflect the monotonicity change of the analysis angle. The more the number of continuous positive numbers and the number of continuous negative numbers, the more obvious the monotonic change of the analysis angle. In this embodiment, the proportion of the maximum value of the number of continuous positive numbers and the number of continuous negative numbers in the first-order difference sequence of the angle sequence to the total number of elements in the angle sequence is used as the local trend index. If the local trend index is larger, it indicates that the trend change of the direction of the second plane of each block in the preset direction is more obvious, and the greater the possibility that each block has a bending defect in each preset direction; on the contrary, the greater the possibility that the uneven defect of each block in each preset direction is caused by the device vibration. The maximum value of the local trend indexes of each block in each preset direction in all dimensions is selected as the overall trend index of each block in each preset direction.

[0066] Considering the trend change situations in multiple preset directions and the bending defects of each block in multiple preset directions to improve the accuracy of the bending defects of the block, the sum of the overall trend indexes of each block in all preset directions is used as the direction trend index of each block; if the direction trend index is larger, the greater the possibility that each block belongs to the bending defect of the device itself, and on the contrary, the greater the possibility of belonging to the vibration defect of the device.

[0067] It should be noted that the coordinate system where the point cloud data is located is defined in step S1; in this embodiment, there are three dimensions, and one coordinate axis of the three-dimensional coordinate system where the point cloud data is located corresponds to one dimension.

[0068] Step S430: Obtain the second correction coefficient of each block according to the normal vibration index and the direction trend index; the normal vibration index and the second correction coefficient are negatively correlated, and the direction trend index and the second correction coefficient are positively correlated; use the second correction coefficient to perform weighted processing on the correction defect index of each block to obtain the final defect index of each block; perform negative correlation and normalization processing on the final defect index to obtain the fitting coefficient of each point cloud data of each block.

[0069] If the normal vibration index is smaller and the direction trend index is larger, the greater the possibility that the block belongs to the bending defect of the device itself, the more accurate the correction defect index, and the smaller the adjustment degree of the correction defect index; on the contrary, the greater the possibility that the block belongs to the defect caused by the device vibration, the greater the error existing in the correction defect index, and the greater the adjustment degree of the correction defect index. The larger the second correction index, the smaller the adjustment degree of the correction defect index. Therefore, the normal vibration index and the second correction coefficient are negatively correlated, and the direction trend index and the second correction coefficient are positively correlated.

[0070] In this solution, by fitting the point cloud data on the surface of the die press, the final reference plane for measuring flatness is constructed. The final reference plane should be constructed using the point cloud data with normal flatness in the device. The greater the final defect index of the point cloud data, the greater the possibility that it belongs to the uneven defect of the device itself, and the greater the measurement error generated. To ensure the accuracy of the construction of the final reference plane, when fitting the plane for the point cloud data with a greater final defect index, a smaller weight needs to be assigned, and when fitting the plane for the point cloud data with a smaller final defect index, a greater weight needs to be assigned to improve the accuracy of the plane fitting. Therefore, it is necessary to perform negative correlation and normalization processing on the final defect index to obtain the fitting coefficient.

[0071] In a specific implementation manner of the embodiment of the present invention, the fitting coefficient of the point cloud data is expressed by the formula: In the formula, is the fitting coefficient of each point cloud data of each sub-block; is the overall trend index of the k-th sub-block on the surface of the die press in the v-th preset direction; V is the total number of preset directions, and in this embodiment, V takes the empirical value 2; is the direction trend index of the k-th sub-block on the surface of the die press; is the normal vibration index of the k-th sub-block on the surface of the die press; is the second correction coefficient of the k-th sub-block on the surface of the die press; is the corrected defect index of the k-th sub-block on the surface of the die press; is the final defect index of the k-th sub-block on the surface of the die press; a is a preset positive number, taking the empirical value 0.1, and its function is to prevent the denominator from being zero and causing the fraction to be meaningless; Norm is the normalization function.

[0072] Step S5: Based on the fitting coefficient, perform plane fitting on the point cloud data on the surface of the die press, and detect the flatness of the surface of the die press according to the distance from the point cloud data on the surface of the die press to the fitted plane.

[0073] Take the fitting coefficient as the weight when performing plane fitting for each point cloud data, and use the weighted least squares method to perform plane fitting on the point cloud data on the surface of the die press to obtain the final reference plane. Among them, the weighted least squares method is a well-known technology to those skilled in the art and will not be elaborated here.

[0074] Obtain the distance from the centroid of the second plane of each block to the final reference plane, which is denoted as the judgment distance of each block; determine whether the judgment distances of all blocks on the surface of the press are less than the preset threshold. If so, the flatness of the press surface is qualified; if not, the flatness of the press surface is unqualified, and relevant parameters of the press need to be adjusted to make the judgment distance greater than the preset threshold less than the preset threshold, or install a new feeding with qualified flatness to maintain the normal flatness of the press surface.

[0075] The centroid of the second plane of each block reflects the spatial position of the second plane, and the distance from the centroid to the final reference plane can reflect the overall level of the distance from the point cloud data of each block to the final reference plane. Then the judgment distance can measure the distance between the block and the surface of the device, and is used to detect the flatness of the press surface.

[0076] It should be noted that in this embodiment, the preset threshold takes an empirical value of 0.15 mm, and the implementer can set it by himself according to the specific situation. In other embodiments, the distance from the center of gravity or the center of the convex hull of the second plane of the block to the final reference plane can be selected as the judgment distance. The detection methods for the flatness of the surfaces of the feeding and discharging of the press are the same.

[0077] This solution adaptively sets the weight when fitting the plane of the point cloud data according to the uneven defect degree of the block where the point cloud data is located, so that the final reference plane represents the overall flatness level of the press surface, thereby improving the accuracy of the flatness detection of the press surface.

[0078] So far, the present invention is completed.

[0079] Embodiment 2: The present invention provides a surface flatness detection system for a press loading and unloading machine. Please refer to Figure 3 , which shows the system structure diagram of a surface flatness detection system for a press loading and unloading machine provided by an embodiment of the present invention. The system includes: A data acquisition module 610, configured to obtain point cloud data of the press surface; divide the point cloud data into multiple blocks; A defect analysis module 620, configured to obtain a reference plane for each block, and obtain an initial defect index for each block according to the difference between the distances from the point cloud data of each block to the reference plane; A defect correction module 630, configured to correct the initial defect index according to the difference between the overall distribution trend of the point cloud data of each block and the overall distribution trend of the point cloud data of the press surface, and the position distribution of the point cloud data of each block, so as to obtain a corrected defect index for each block; The fitting coefficient analysis module 640 is configured to obtain the fitting coefficient of the point cloud data of each block according to the dispersion degree of the corrected defect indexes of all blocks in each preset direction of each block, the change trend of the overall distribution trend of the point cloud data of all blocks in all preset directions of each block, and the corrected defect index of each block. The flatness detection module 650 is configured to perform plane fitting on the point cloud data of the surface of the molding press based on the fitting coefficient, and detect the flatness of the surface of the molding press according to the distance from the point cloud data of the surface of the molding press to the fitted plane.

[0080] It should be noted that: for the device provided in the above embodiment, only the division of the above functional modules is used for illustration. In actual applications, the above functions can be allocated to different functional modules according to needs, that is, the internal structure of the computer device is divided into different functional modules to complete all or part of the functions described above. In addition, an embodiment of a surface flatness detection system for a molding loading and unloading machine and an embodiment of a surface flatness detection method for a molding loading and unloading machine provided in the above embodiment belong to the same concept. For the specific implementation process, please refer to the method embodiment, which will not be elaborated here.

[0081] Embodiment 3: Figure 4 The schematic diagram of a computer device of a surface flatness detection device for a molding loading and unloading machine provided by an embodiment of the present invention. Exemplarily, as Figure 4 shown, the computer device includes: a memory 701, a processor 702, and a computer program 703 stored in the memory 701 and running on the processor 702. When the processor 702 executes the computer program 703, the computer device can execute any one of the surface flatness detection methods for a molding loading and unloading machine introduced above.

[0082] In addition, an embodiment of the present application also protects a device, which may include a memory and a processor. The memory stores executable program code, and the processor is configured to call and execute the executable program code to execute a surface flatness detection method for a molding loading and unloading machine provided by an embodiment of the present application.

[0083] This embodiment can divide the functions of the device according to the above method examples. For example, it can correspond to each functional module, or integrate two or more functions into one processing module. The above integrated module can be implemented in the form of hardware. It should be noted that the division of modules in this embodiment is illustrative, only a logical function division, and there may be other division methods in actual implementation.

[0084] It should be understood that the device provided in this embodiment is used to execute the above-mentioned method for detecting the surface flatness of a molding loading and unloading machine, so the same effects as the above-mentioned implementation method can be achieved.

[0085] In the case of adopting an integrated unit, the device may include a processing module and a storage module. Among them, when the device is applied to a device, the processing module can be used to control and manage the actions of the device. The storage module can be used to support the device to execute mutual program codes, etc.

[0086] Among them, the processing module can be a processor or a controller, which can implement or execute various exemplary logic blocks, modules, and circuits included in the disclosure of the present application. The processor can also be a combination that realizes computing functions, such as including a combination of one or more microprocessors, a combination of digital signal processing (DSP) and a microprocessor, etc. The storage module can be a memory.

[0087] It should be noted that: the above-mentioned sequence of embodiments of the present invention is only for description and does not represent the advantages or disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0088] Each embodiment in this specification is described in a progressive manner. The same or similar parts among the embodiments can be referred to each other, and the key points of each embodiment are the differences from other embodiments.

[0089] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for detecting the surface flatness of a molding loading and unloading machine, characterized in that, The method includes: Obtaining point cloud data on the surface of a molding press; dividing the point cloud data into multiple blocks; Obtaining a reference plane for each block, and obtaining an initial defect index for each block according to the difference between the distances from the point cloud data of each block to the reference plane; Correcting the initial defect index according to the difference between the overall distribution trend of the point cloud data of each block and the overall distribution trend of the point cloud data on the surface of the molding press, and the position distribution of the point cloud data of each block, to obtain a corrected defect index for each block; Obtaining a fitting coefficient for the point cloud data of each block according to the degree of dispersion of the corrected defect indices of all blocks in each preset direction for each block, the change trend of the overall distribution trend of the point cloud data of all blocks in all preset directions for each block, and the corrected defect index of each block; Performing plane fitting on the point cloud data on the surface of the molding press based on the fitting coefficient, and detecting the flatness of the surface of the molding press according to the distance from the point cloud data on the surface of the molding press to the plane obtained by fitting.

2. The surface flatness detection method for a molding loading and unloading machine according to claim 1, characterized in that The obtaining of the reference plane for each block, and obtaining the initial defect index for each block according to the difference between the distances from the point cloud data of each block to the reference plane, includes: Dividing the point cloud data of each block into normal point cloud data and abnormal point cloud data; Performing plane fitting on the normal point cloud data of each block to obtain a reference plane for the corresponding block, and obtaining the quartiles of the distances from each type of point cloud data of each block to the reference plane; Forming an n-tuple from the quartiles corresponding to each type of point cloud data of each block, where n is the total number of quartiles corresponding to each type of point cloud data; Denoting the distance between the n-tuples of the normal point cloud data and the abnormal point cloud data of each block as the distance difference index of each block; Obtaining the interquartile range of the distances from the normal point cloud data of each block to the reference plane; obtaining the initial defect index for each block according to the interquartile range and the distance difference index; both the interquartile range and the distance difference index have a positive correlation with the initial defect index.

3. A surface flatness detection method for a molding loading and unloading machine according to claim 1, characterized in that, The obtaining of the corrected defect index for each block includes: Performing plane fitting on the point cloud data of each block respectively to obtain a second plane for the corresponding block, and performing plane fitting on the point cloud data on the surface of the molding press to obtain an overall plane; obtaining a similarity index between the normal vectors of the second plane and the overall plane of each block; Obtaining the mean square error between all the point cloud data of each block and the corresponding point cloud data on the second plane, and denoting it as the fitting error of each block; Obtaining a first correction coefficient for each block according to the similarity index and the fitting error; the similarity index has a negative correlation with the first correction coefficient, and the fitting error has a positive correlation with the first correction coefficient; Using the sum value of the first correction coefficient and the constant 1 to perform weighted processing on the initial defect index of each block to obtain the corrected defect index of each block.

4. A method for detecting the surface flatness of a molding loading and unloading machine according to claim 3, characterized in that, The obtaining of the fitting coefficient for the point cloud data of each block includes: Obtain the discrete index and the centralized index of the corrected defect index of all blocks in each preset direction for each block; according to the discrete index and the centralized index, obtain the normal vibration index of each block in each preset direction; According to the change trend of the overall distribution trend of the point cloud data of all blocks in all preset directions for each block, obtain the direction trend index of each block; According to the normal vibration index and the direction trend index, obtain the second correction coefficient of each block; the normal vibration index and the second correction coefficient are negatively correlated, and the direction trend index and the second correction coefficient are positively correlated; use the second correction coefficient to perform weighted processing on the corrected defect index of each block to obtain the final defect index of each block; perform negative correlation and normalization processing on the final defect index to obtain the fitting coefficient of each point cloud data of each block.

5. A method for detecting the surface flatness of a molding loading and unloading machine according to claim 4, characterized in that, The obtaining of the direction trend index of each block includes: Obtain the angles between the normal vectors of the second planes of each block and each coordinate axis of the coordinate system where the point cloud data is located, and record them as the analysis angles of each block in each dimension; one dimension corresponds to one of the coordinate axes; arrange the analysis angles of all blocks in each dimension in each preset direction for each block in order to obtain the angle sequence of each block in each preset direction in each dimension; Obtain the first-order difference sequence of the angle sequence, and record the proportion of the maximum value of the number of consecutive positive numbers and the number of consecutive negative numbers in the first-order difference sequence to the total number of elements in the angle sequence as the local trend index of each block in each preset direction in each dimension; select the maximum value of the local trend indexes of each block in each preset direction in all dimensions as the overall trend index of each block in each preset direction; Take the sum of the overall trend indexes of each block in all preset directions as the direction trend index of each block.

6. A method for detecting the surface flatness of a molding loading and unloading machine according to claim 3, characterized in that, The plane fitting of the point cloud data on the surface of the molding press based on the fitting coefficient, and detecting the flatness of the surface of the molding press according to the distance from the point cloud data on the surface of the molding press to the fitted plane, includes: Based on the fitting coefficient, use the weighted least squares method to perform plane fitting on the point cloud data on the surface of the molding press to obtain the final reference plane; Obtain the distance from the centroid of the second plane of each block to the final reference plane, and record it as the judgment distance of each block; Judge whether the judgment distances of all blocks on the surface of the molding press are all less than a preset threshold. If so, the flatness of the surface of the molding press is qualified; if not, the flatness of the surface of the molding press is unqualified.

7. A method for detecting the surface flatness of a molding loading and unloading machine according to claim 2, characterized in that, The dividing of the point cloud data of each block into normal point cloud data and abnormal point cloud data includes: Perform plane fitting on the point cloud data of each block to obtain the first plane corresponding to each block; Obtain the distances between all the point cloud data of each block and the first plane, use the maximum inter-class variance method for the distances to obtain the division threshold, and respectively record the point cloud data with a distance less than the division threshold as normal point cloud data, and the point cloud data with a distance greater than or equal to the division threshold as abnormal point cloud data.

8. A method for detecting the surface flatness of a molding loading and unloading machine according to claim 3, characterized in that, The similarity index is the cosine similarity.

9. A method for detecting the surface flatness of a molding loading and unloading machine according to claim 4, characterized in that, Obtaining the discrete index and the concentration index of the corrected defect index of all blocks in each preset direction for each block includes: The discrete index is the variance of the corrected defect index of all blocks in each preset direction for each block, and the concentration index is the mean of the corrected defect index of all blocks in each preset direction for each block.

10. A surface flatness detection system for a molding loading and unloading machine, characterized in that, The system includes: A data acquisition module, configured to acquire point cloud data on the surface of a molding press; divide the point cloud data into multiple blocks; A defect analysis module, configured to obtain a reference plane for each block, and obtain an initial defect index for each block according to the difference between the distances from the point cloud data of each block to the reference plane; A defect correction module, configured to correct the initial defect index according to the difference between the overall distribution trend of the point cloud data of each block and the overall distribution trend of the point cloud data on the surface of the molding press, and the position distribution of the point cloud data of each block, so as to obtain a corrected defect index for each block; A fitting coefficient analysis module, configured to obtain a fitting coefficient of the point cloud data of each block according to the degree of dispersion of the corrected defect index of all blocks in each preset direction for each block, the change trend of the overall distribution trend of the point cloud data of all blocks in all preset directions for each block, and the corrected defect index of each block; A flatness detection module, configured to perform plane fitting on the point cloud data on the surface of the molding press based on the fitting coefficient, and detect the flatness of the surface of the molding press according to the distance from the point cloud data on the surface of the molding press to the fitted plane.

Images