Oil seal oil return line height detection method

Abstract

The invention relates to the technical field of oil seal detection, in particular to an oil seal oil return line height detection method, which comprises the following steps: acquiring all point cloud data on an oil seal inner ring; constructing a cylinder corresponding to the oil seal, and dividing the bottom surface of the cylinder into a plurality of fan-shaped blocks; calculating an axis distance; calculating a deformation characterization value of each fan-shaped block; determining the local concentration degree of each point cloud data; determining a discrimination coefficient of each point cloud data, and obtaining the point cloud data corresponding to the oil return line; clustering all the point cloud data corresponding to the oil return line, and determining an adjustment coefficient of each point cloud data corresponding to the oil return line; and determining the filtering scale of each point cloud data corresponding to the oil return line, performing filtering by combining a filtering algorithm, performing three-dimensional reconstruction on all the filtered point cloud data, and measuring the height of the oil return line of the oil seal. According to the invention, the appropriate filtering scale of the filtering algorithm is dynamically selected, the filtering effect is enhanced, and the accuracy of oil seal oil return line height detection is improved.

Description

Technical Field

[0001] This application relates to the technical field of oil seal detection, and specifically relates to a method for detecting the height of the oil return line of an oil seal. Background Art

[0002] The oil return line on the oil seal is located at the main lip part of its inner side wall in a threaded structure. When the shaft rotates, the pumping effect generated by the oil return line effectively prevents the loss of lubricating oil and the intrusion of external pollutants, thus ensuring the normal operation of the machine and extending the service life of the machine. The height of the oil return line thread affects the sealing performance of the oil seal. Insufficient height will cause sealing failure or poor grease backflow, while excessive height will increase frictional losses.

[0003] During the production process of oil seals, traditional methods for detecting the height of the oil return line generally use contact measurement methods such as product slicing and silicone printing. These methods are not only cumbersome and inefficient in operation, but also can only perform sampling inspections, unable to meet the full inspection requirements of the oil seal production line. Ultra-high-speed profilometers can meet the full inspection requirements for mass production, but they need to perform filtering and denoising processing on the 3D point cloud data on the oil return line of the oil seal. When performing K-nearest neighbor mean filtering, generally, the point cloud data within a fixed filtering scale is directly selected as the neighborhood point set, which is extremely likely to damage the directional texture features of the point cloud data on the threaded oil return line and is difficult to retain the edge detail features of the threaded oil return line, resulting in large errors in the detection of the oil return line height. Summary of the Invention

[0004] In order to solve the above technical problems, a method for detecting the height of the oil return line of an oil seal is provided to solve the existing problems.

[0005] The solution of this application to solve the technical problems is to provide a method for detecting the height of the oil return line of an oil seal, including the following steps: Collect all the point cloud data on the inner ring of the oil seal; Perform three-dimensional fitting on all the point cloud data to construct a cylinder corresponding to the oil seal. Divide the bottom surface of the cylinder into multiple sector blocks, and divide each point cloud data into each sector block according to the distribution of the polar angle in the polar coordinates of each point cloud data in each sector block; Based on the distance between each point cloud data and the cylinder axis, combined with the radius of the bottom surface of the cylinder, calculate the axis distance; calculate the deformation characterization value of each sector block through the discrete situation and mutation characteristics of the axis distance of the point cloud data between each sector block and the other sector blocks; Analyze the distribution density of all the point cloud data in the neighborhood of each point cloud data to determine the local concentration of each point cloud data; combine the axis distance and the deformation characterization value to determine the discrimination coefficient of each point cloud data, screen the point cloud data, and obtain the point cloud data corresponding to the oil return line; Cluster all the point cloud data corresponding to the oil return line, analyze the deviation between the point cloud data corresponding to the oil return line in each clustering cluster and the remaining point cloud data in its neighborhood, as well as the distribution of the point cloud data in its neighborhood, and determine the adjustment coefficient of each point cloud data corresponding to the oil return line in combination with the deformation characterization value; Based on the adjustment coefficient, determine the filtering scale of each point cloud data corresponding to the oil return line, perform filtering in combination with the filtering algorithm, perform three-dimensional reconstruction on all the filtered point cloud data, and measure the oil return line height of the oil seal.

[0006] Preferably, the dividing of each point cloud data into each sector block includes: Taking the center of the bottom surface of the cylinder as the origin and the axis of the cylinder as the z-axis, construct a three-dimensional coordinate system; map all the point cloud data on the cylinder to the xOy plane of the three-dimensional coordinate system, and calculate the polar coordinates of each point cloud data; According to the sector block area where the polar angle is distributed in the polar coordinates of each point cloud data, divide all the point cloud data into different sector blocks.

[0007] Preferably, the calculation process of the axis distance is: calculate the closest distance from each point cloud data to the axis of the cylinder; take the ratio between the closest distance and the radius of the bottom surface of the cylinder as the axis distance of each point cloud data.

[0008] Preferably, the calculating of the deformation characterization value of each sector block includes: Calculate the dispersion degree of the axis distances of all the point cloud data in each sector block; calculate the ratio between the dispersion degree of each sector block and the average value of the dispersion degrees of all the sector blocks, and denote it as the relative dispersion ratio; Perform mutation point detection on the dispersion degrees in all the sector blocks to obtain the mutation probability of each sector block; The deformation characterization value is the normalized result of the product of the relative dispersion ratio and the mutation probability.

[0009] Preferably, the determining of the local concentration degree of each point cloud data includes: Taking each point cloud data as the center, denote the neighborhood range with a preset radius as the local neighborhood; Calculate the range of values of the position coordinates corresponding to all the point cloud data in the local neighborhood on the x-axis, y-axis, and z-axis respectively; calculate the average value of the range of values corresponding to all the axes; The local concentration degree is the ratio between the average value and the diameter of the local neighborhood.

[0010] Preferably, the determining of the discrimination coefficient of each point cloud data includes: taking the deformation characterization value of the sector block to which each point cloud data belongs as the weight, perform weighted summation on the reciprocal of the local concentration degree and the reciprocal of the axis distance, and use it as the discrimination coefficient of each point cloud data.

[0011] Preferably, obtaining the point cloud data corresponding to the oil return line includes: obtaining the segmentation threshold of the discrimination coefficient of all the point cloud data; and designating the point cloud data with the discrimination coefficient greater than the segmentation threshold as the point cloud data corresponding to the oil return line.

[0012] Preferably, determining the adjustment coefficient of each point cloud data corresponding to the oil return line includes: By clustering all the point cloud data corresponding to the oil return line, obtaining the local density of each point cloud data corresponding to the oil return line, wherein the truncation distance of the clustering algorithm is a preset value; Selecting, from the position coordinates of the remaining all point cloud data corresponding to the oil return line within the neighborhood corresponding to the truncation distance of any point cloud data corresponding to the oil return line and belonging to the same clustering cluster as the any point cloud data, the mean value as the neighborhood center coordinate; Calculating the distance between the position coordinate of the any point cloud data and the neighborhood center coordinate, and designating it as the center deviation degree; Calculating the product value between the local density of each point cloud data corresponding to the oil return line and the deformation characterization value of the sector block to which it belongs; The adjustment coefficient is the normalized result of the ratio of the product value to the center deviation degree.

[0013] Preferably, for the th point cloud data corresponding to the oil return line, the filtering scale is calculated by the formula: , where is a preset value, is the preset initial filtering scale, and is the adjustment coefficient of the th point cloud data corresponding to the oil return line.

[0014] Preferably, measuring the height of the oil return line of the oil seal includes: according to the three-dimensional cylinder after three-dimensional reconstruction, calculating the minimum distance from each point cloud data corresponding to the oil return line on the three-dimensional cylinder to the axis of the three-dimensional cylinder, and taking the mean value of the difference between the radius of the inner circle of the oil seal and the minimum distance of all the point cloud data corresponding to the oil return line as the height of the oil return line of the oil seal.

[0015] This application has at least the following beneficial effects: In this application, all point cloud data is fitted to construct a cylinder, and multiple sector blocks are divided on the bottom surface of the cylinder. Based on the distribution of the polar angles of different point cloud data in the sectors of the polar coordinates, all the point cloud data within each sector block is obtained. Based on the distances from each point cloud data to the cylinder axis, combined with the radius of the bottom surface of the cylinder, the axis distance is calculated, and the deformation characterization value of each sector block is calculated. The beneficial effect is that it considers classifying the point cloud data within the same range of polar angles into one category, so as to analyze the distance between the point cloud data at the position of the sector block and the axis of the cylinder, and to evaluate whether there is a deformation phenomenon at the position of the sector block. Secondly, through the distribution density of all the point cloud data within the neighborhood of each point cloud data, the local concentration degree of each point cloud data is determined, and then the discrimination coefficient of each point cloud data is determined to obtain the point cloud data corresponding to the oil return line. The beneficial effect is that it considers the distribution of the point cloud data on the sector blocks in the deformation area. When the point cloud data is distributed on the sector blocks in the deformation area, at this time, it is impossible to distinguish whether the point cloud data is distributed on the oil return line through the inter-axis distance, and it can be judged by the local concentration degree. When the point cloud data is not in the deformation area, at this time, it can be distinguished by the inter-axis distance, improving the recognition accuracy of the point cloud data inside the oil seal inner ring. By separating the point cloud data distributed on the oil return line, the height of the oil return line can be measured subsequently; further, by clustering all the point cloud data corresponding to the oil return line, the beneficial effect is that the point cloud data on the oil return line belonging to the same circle is classified into one category. Secondly, the adjustment coefficient of each point cloud data corresponding to the oil return line is determined; based on the adjustment coefficient, the filtering scale of each point cloud data corresponding to the oil return line is determined, and filtering is performed in combination with the filtering algorithm. Three-dimensional reconstruction is performed on all the filtered point cloud data to measure the height of the oil return line of the oil seal. The beneficial effect is that when filtering each point cloud data corresponding to the oil return line, a suitable filtering scale of the filtering algorithm is dynamically selected, enhancing the filtering effect while retaining the spiral texture and edge detail features of the oil return line of the oil seal inner ring, reducing the detection error of the oil return line height, and improving the accuracy of detecting the height of the oil return line of the oil seal. Brief Description of the Drawings

[0016] The following further elaborates on a method for detecting the height of the oil return line of an oil seal according to this application with reference to the accompanying drawings.

[0017] Figure 1 It is a flowchart of the steps of a method for detecting the height of the oil return line of an oil seal provided by an embodiment of this application; Figure 2 It is a flowchart of the steps of a method for obtaining the deformation characterization value of each sector block provided by an embodiment of this application; Figure 3 It is a schematic diagram of the division of sector blocks in the cylinder corresponding to the inner ring of the oil seal provided by an embodiment of this application; Figure 4Schematic diagram of projecting point cloud data provided by an embodiment of the present application onto the xOy plane. Detailed implementation manners

[0018] In order to make the objectives, technical solutions and advantages of the present application more clearly understood, the following further elaborates in detail on a method for detecting the height of the oil return line of an oil seal proposed in the present application in combination with the accompanying drawings and implementation examples. It should be understood that the specific implementation examples described herein are only used to explain the present application and are not used to limit the present application.

[0019] 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 this application belongs.

[0020] Please refer to Figure 1 , which shows a flowchart of the steps of a method for detecting the height of the oil return line of an oil seal provided by an embodiment of the present application. The method includes the following steps: Step 1: Collect all the point cloud data on the inner ring of the oil seal.

[0021] An oil seal is a sealing element used in mechanical equipment to prevent lubricating oil leakage and entry of external contaminants. It is usually installed between stationary parts, rotating shafts or reciprocating moving parts. The oil return line is a special structure on the inclined surface of the lip of the oil seal, which forms a pumping effect during rotation to pump the lubricating oil flowing into the lip part back to one side of the oil seal, playing a sealing role.

[0022] Therefore, use an automatic pneumatic fixture to fix the oil seal, and use a contour measuring instrument with a rotary scanner probe. Insert the scanner probe into the inner ring of the oil seal and perform a 360° rotary scan to collect all the point cloud data on the inner ring of the oil seal.

[0023] So far, all the point cloud data on the inner ring of the oil seal has been obtained.

[0024] Step 2: Perform three-dimensional fitting on all the point cloud data, construct a cylinder corresponding to the oil seal, divide the bottom surface of the cylinder into multiple sector blocks, and divide each point cloud data into each sector block according to the distribution of the polar angle in the polar coordinates of each point cloud data in each sector block; based on the distance between each point cloud data and the cylinder axis, combined with the radius of the bottom surface of the cylinder, calculate the axis distance; calculate the deformation characterization value of each sector block through the discrete situation and mutation characteristics of the axis distance of the point cloud data between each sector block and the other sector blocks.

[0025] The inner ring of the oil seal usually contacts the rotating shaft, and its outer shape is generally cylindrical, aiming to form a tight seal with the rotating shaft. Although there may be oil return lines on the surface of the inner ring of the oil seal, these oil return lines are usually local optimizations of the cylinder surface. Overall, the inner ring of the oil seal still maintains the basic shape of the cylinder, that is, the three-dimensional geometric shape of the inner ring of the oil seal is approximately cylindrical as a whole.

[0026] Based on the above analysis, fitting is performed on all point cloud data, specifically as follows: Perform cylinder fitting on all point cloud data to obtain the cylinder axis; In this embodiment, the Random Sample Consensus (RANSAC) algorithm is used for cylinder fitting. By loading all point cloud data into the PCL library, a SACSegmentation object is created, and the model type is set to cylinder SACMODEL_CYLINDER, and the method type is RANSAC for cylinder fitting. Among them, the RANSAC algorithm is a well-known technology and will not be elaborated here.

[0027] Secondly, the non-protruding part on the inner surface of the oil seal is used as the reference plane. The inner surface of the oil seal mainly consists of a smooth reference plane and an oil return line with obvious thread-like protrusions. Among them, the point cloud data on the reference plane is far from the cylinder axis, while the oil return line with thread-like protrusions has a certain height. Therefore, the point cloud data on the oil return line is closer to the cylinder axis. Therefore, analyze the distance from each point cloud data to the cylinder axis, specifically as follows: Calculate the closest distance from each point cloud data to the cylinder axis; Take the ratio between the closest distance and the radius of the cylinder bottom surface as the axis distance of each point cloud data; Furthermore, due to factors such as uneven clamping force of the pneumatic fixture and radial runout of the scanner probe, the point cloud data of the inner surface of the oil seal is deformed and cannot maintain a uniform circular contour. Compared with the distance from the point cloud data on the reference plane to the cylinder axis, there will be a situation where the distance from the point cloud data of the oil return line to the cylinder axis is greater, making it difficult to distinguish the point cloud data corresponding to the reference plane and the oil return line points on the deformed inner surface of the oil seal through the axis distance. Based on this, analyze the situation where the position of the point cloud data on the inner surface of the oil seal is deformed, and calculate the deformation characterization value. The step flow chart of the method for obtaining the deformation characterization value of each sector block provided in the embodiment of the present application is as Figure 2 shown, specifically including: Taking the center of the bottom surface of the cylinder as the origin and the cylinder axis as the z-axis, a three-dimensional coordinate system is constructed; Map all point cloud data on the cylinder to the xOy plane of the three-dimensional coordinate system, and calculate the polar coordinates of each point cloud data; Divide the bottom surface of the cylinder into multiple sector blocks, and divide all point cloud data into different sector blocks according to the sector blocks where the polar angles in the polar coordinates of each point cloud data are distributed; In this embodiment, the circumference of the bottom surface of the cylinder is 360°, which is divided into 24 intervals, that is, each interval is 15° as an interval, and each interval corresponds to a sector block, so that the bottom surface of the cylinder is divided into 24 sector blocks. The sector block corresponding to the interval where the polar angle is located in the polar coordinates of all the point cloud data on the cylinder is used to divide all the point cloud data into different sector blocks. As other embodiments, the implementer can set it by himself according to the actual situation.

[0028] It should be noted that for the convenience of understanding, assuming that the bottom surface of the cylinder is divided into 8 sector blocks as an example, the schematic diagram of the division of the sector blocks corresponding to the oil seal inner ring in the cylinder provided in this embodiment is as Figure 3 shown; the schematic diagram of the point cloud data projected onto the xOy plane provided in this embodiment is as Figure 4 shown, where 1 represents sector block 1, 2 is sector block 2, A is the point where the point cloud data is projected onto the xOy plane. If the polar angle in the polar coordinates of the th point cloud data is 10°, then the polar angle of the point cloud data is distributed in the corresponding Figure 3 sector block 2, then the th point cloud data belongs to sector block 2.

[0029] Calculate the degree of dispersion of the axial distances of all the point cloud data in each sector block; In this embodiment, the degree of dispersion is measured by calculating the standard deviation of the axial distances of all the point cloud data in each sector block. As other embodiments, the implementer can adopt other methods in the prior art, such as variance, coefficient of variation, etc. This embodiment does not make special restrictions on this.

[0030] Perform change point detection on the degree of dispersion in all the sector blocks to obtain the change probability of each sector block; In this embodiment, after arranging the degree of dispersion in all the sector blocks in ascending order according to the polar angle range of the corresponding interval, change point detection is performed through the Bayesian change point detection algorithm. Among them, the Bayesian change point detection algorithm is a well-known technology and will not be elaborated here.

[0031] Calculate the ratio of the degree of dispersion of each sector block to the mean value of the degree of dispersion of all the sector blocks, which is denoted as the relative dispersion ratio; Take the normalized result of the product of the relative dispersion ratio and the change probability as the deformation characterization value of each sector block; In this embodiment, taking the deformation characterization value of the th sector block as an example, its calculation formula is: Among them, is the deformation characterization value of the th sector block, is the The degree of discreteness of the fan-shaped blocks, is the mean value of the discrete degree of all the fan-shaped blocks, For the The mutation probability of a sector block, is a normalization function. In this embodiment, a sigmoid function is used for normalization. The sigmoid function is a well-known technology and will not be described in detail here. As other implementation methods, the implementer can adopt other methods of the prior art, such as the tanh function, etc. This embodiment does not impose any special restrictions on this.

[0032] It should be noted that, the larger the relative discrete ratio is, the greater the unevenness of the point cloud data of the inner ring of the oil seal in the sector block compared to the overall point cloud data when there is a distance difference between the reference plane and the oil return line itself; the larger the mutation probability is, the more abrupt the difference in unevenness of the inner ring of the oil seal in the sector block is; the larger the obtained deformation characterization value is, the more likely it is that the inner ring of the oil seal in the sector block is in the position area where the oil seal is deformed due to factors such as uneven clamping force of the fixture.

[0033] At this point, the deformation characterization value of each sector block is obtained.

[0034] Step 3: Analyze the distribution density of all point cloud data in the neighborhood of each point cloud data to determine the local concentration of each point cloud data; combine the axis distance and deformation characterization value to determine the discrimination coefficient of each point cloud data, screen the point cloud data, and obtain the point cloud data corresponding to the oil return line.

[0035] The reference surface of the inner ring of the oil seal belongs to the remaining part of the inner ring surface of the oil seal except the raised oil return line. The distance between the point cloud data in the neighborhood is small, the coordinate values ​​are similar, and the local distribution concentration is high. The oil return line is usually distributed in a regular spiral shape with rich details and textures. When it is subjected to uneven clamping force of the pneumatic clamp and radial runout of the scanner probe, the three-dimensional shape characteristics of the oil return line point cloud are more easily destroyed, and the distribution concentration of the point cloud data in the neighborhood is low. Therefore, the distribution concentration of each point cloud data in the neighborhood is analyzed, and the local concentration is calculated, specifically: Taking each point cloud data as the center, the neighborhood range of the preset radius is recorded as the local neighborhood; In this embodiment, the preset radius is 0.1 mm. As for other implementations, the implementer can set it according to the actual situation.

[0036] Calculate the range of the position coordinates corresponding to all point cloud data in the local neighborhood on the x-axis, y-axis and z-axis respectively; For each point cloud data, the average value of the range corresponding to all axes in the local neighborhood is calculated; and the ratio between the average value and the diameter of the local neighborhood is used as the local concentration of each point cloud data; It should be noted that, for the convenience of understanding, taking three point cloud data as an example, it is assumed that there are three point cloud data in the local neighborhood, denoted as A, B, and C respectively, and their position coordinates are A , B , C . Among them, the maximum values of each axis in the position coordinates of the three point cloud data A, B, and C are respectively , , , and the minimum values are respectively , , . Therefore, the range of the x-axis is ; the range of the y-axis is ; the range of the z-axis is ; secondly, since the preset radius is 0.1, the diameter of the local neighborhood is 0.2.

[0037] It should be noted that the smaller the local concentration degree, the higher the distribution concentration degree of the point cloud data in the local neighborhood, and the more likely it belongs to the point cloud data at the reference plane in the inner circle of the oil seal.

[0038] Furthermore, for the point cloud data in the fan-shaped block area of the inner circle of the oil seal with a large deformation degree, the point cloud data is distinguished by the local concentration degree to avoid the situation where the distance difference between the oil return line and the point cloud data on the reference plane is larger; for the point cloud data in the fan-shaped block area of the inner circle of the oil seal with a small deformation degree, the distance difference between the point cloud data on the reference plane and the oil return line is significant, and the point cloud data is distinguished by the difference feature of the axis distance. Therefore, based on the axis distance and the local concentration degree, combined with the deformation characterization value, a discrimination coefficient is determined to distinguish the possibility that the point cloud data is the corresponding point cloud data on the reference plane or the oil return line. Specifically: Taking the deformation characterization value of the fan-shaped block to which each point cloud data belongs as the weight, the weighted sum of the reciprocal of the local concentration degree and the reciprocal of the axis distance is used as the discrimination coefficient of each point cloud data; In this embodiment, taking the discrimination coefficient of the th point cloud data as an example, its calculation formula is: Among them, is the discrimination coefficient of the th point cloud data, is the deformation characterization value of the nth fan-shaped block to which the th point cloud data belongs, is the axis distance of the th point cloud data, is the local concentration degree of the th point cloud data.

[0039] It should be noted that since the above-mentioned deformation characterization value has been normalized by the sigmoid function the value range of the deformation characterization value is .

[0040] It should be noted that when the deformation characterization value is larger, it indicates that there is a deformation phenomenon in the position area of the sector block to which each point cloud data belongs. At this time, it is judged by the local concentration degree. The larger the reciprocal of the local concentration degree, the more likely the point cloud data is distributed on the oil return line of the inner circle of the oil seal; when the deformation characterization value is smaller, it indicates that there is no deformation in the position area of the sector block to which each point cloud data belongs. At this time, it is judged by the axis distance. The larger the reciprocal of the axis distance, the closer the point cloud data is to the cylindrical axis, and the more likely the point cloud data is distributed on the oil return line of the inner circle of the oil seal, and the larger the discrimination coefficient.

[0041] Furthermore, based on the discrimination coefficient, all point cloud data are screened to obtain the point cloud data corresponding to the oil return line, specifically: The threshold segmentation algorithm is used to obtain the segmentation threshold of the discrimination coefficient of all point cloud data; The point cloud data with the discrimination coefficient greater than the segmentation threshold is recorded as the point cloud data corresponding to the oil return line; In this embodiment, the Otsu threshold segmentation algorithm is used to obtain the segmentation threshold. Among them, the Otsu threshold segmentation algorithm is a well-known technology and will not be elaborated here. As other implementation manners, implementers can use other methods of the existing technology, such as the cross-validation method, etc. This embodiment does not make special restrictions on this.

[0042] So far, the point cloud data corresponding to the oil return line is obtained.

[0043] Step 4: Cluster all the point cloud data corresponding to the oil return line, analyze the deviation between each point cloud data corresponding to the oil return line in each cluster and the remaining point cloud data in its neighborhood, and the distribution of the point cloud data in its neighborhood. Combining the deformation characterization value, determine the adjustment coefficient of each point cloud data corresponding to the oil return line; based on the adjustment coefficient, determine the filtering scale of each point cloud data corresponding to the oil return line, combine the filtering algorithm for filtering, and perform three-dimensional reconstruction on all the filtered point cloud data to measure the height of the oil return line of the oil seal.

[0044] Furthermore, the oil return line on the surface of the inner circle of the oil seal is regularly spirally distributed. By clustering all the point cloud data corresponding to the oil return line, the point cloud data on the same circle of the oil return line is divided into one category, specifically: Cluster all the point cloud data corresponding to the oil return line to obtain multiple clusters and the local density of each point cloud data corresponding to the oil return line, where the truncation distance of the clustering algorithm is a preset value; In this embodiment, a density peak clustering algorithm (DPC) is used for clustering, wherein the DPC clustering algorithm is a well-known technology and will not be described in detail here. In order to prevent the point cloud data on similar oil return lines from being segmented into the same cluster cluster, the cutoff distance of the DPC clustering algorithm is set to half of the oil return line spacing, wherein the oil return line spacing refers to the spacing between the two corresponding spiral lines, and the spacing is obtained according to the oil seal production specification data. In this embodiment, the oil return line spacing is set to 2 mm. As other implementation methods, the implementer can set it according to the actual situation.

[0045] Secondly, in the process of collecting point cloud data on the inner ring of the oil seal, due to the interference of factors such as light, the oil return line structure of the inner ring of the oil seal will be incomplete, the point cloud data corresponding to the oil return line will be intermittent, and outlier noise will exist. Therefore, it is necessary to filter the point cloud data corresponding to the oil return line. While removing the noise, repair the missing point cloud and enhance the quality of the point cloud, retain the spiral texture and edge structure of the oil return line as much as possible, so that the distribution of the point cloud data corresponding to the oil return line on the inner ring of the oil seal has certain symmetrical characteristics and is evenly distributed on the oil return line.

[0046] Based on the above analysis, it is necessary to analyze the distribution of the point cloud data corresponding to the oil return line within the cluster and calculate the center deviation to evaluate the distribution uniformity of the point cloud data corresponding to the oil return line within the cluster. Specifically: From any point cloud data corresponding to the oil return line in the neighborhood corresponding to the cutoff distance, select the mean of the position coordinates of all other point cloud data corresponding to the oil return line in the same cluster as the point cloud data corresponding to the oil return line, and record it as the neighborhood center coordinates; In this embodiment, taking any point cloud data corresponding to the oil return line as the center, the Euclidean distance between the remaining point cloud data corresponding to the oil return line in each cluster and the any point cloud data is calculated, and all the remaining point cloud data corresponding to the oil return line whose Euclidean distance with any point cloud data corresponding to the oil return line is less than the cutoff distance are selected; therefore, the mean of the position coordinates of all the remaining point cloud data corresponding to the oil return line is calculated. For the convenience of understanding, it is assumed that there are three remaining point cloud data corresponding to the oil return line, which are recorded as A, B, and C, respectively, and their position coordinates are A, B, and C, respectively. , B , C , then the coordinates of the neighborhood center are .

[0047] Calculate the distance between the position coordinates of any point cloud data corresponding to the oil return line and the center coordinates of the neighborhood, and record it as the center deviation; In this embodiment, the Euclidean distance between the position coordinates of any point cloud data corresponding to the oil return line and the neighborhood center coordinates is calculated and denoted as the center deviation degree.

[0048] It should be noted that the larger the center deviation degree is, the more significant the uneven distribution of the remaining point cloud data in the local neighborhood of the point cloud data corresponding to the oil return line is.

[0049] Secondly, when filtering the point cloud data corresponding to the oil return line, the point cloud data corresponding to the oil return line has significant directional texture features. Directly using the K-nearest neighbor mean filter may lose the directional texture features of the oil return line and the edge detail features of the threaded oil return line. Therefore, different neighborhood scales need to be set according to the distribution density of different point cloud data corresponding to the oil return line and the deformation conditions of the positions where different point cloud data corresponding to the oil return line are located, so as to select an appropriate neighborhood scale for filtering, avoiding the blurring of the directional texture of the oil return line caused by too large a neighborhood scale and the uneven distribution of the corresponding point cloud data on the oil return line and poor filtering effect of noise points caused by too small a neighborhood scale.

[0050] Therefore, based on the local density and the center deviation degree, combined with the deformation characterization value, an adjustment coefficient is determined, specifically: Calculate the product value between the local density of each point cloud data corresponding to the oil return line and the deformation characterization value of the sector block to which it belongs, and take the normalized result of the ratio of the product value to the center deviation degree as the adjustment coefficient of each point cloud data corresponding to the oil return line; In this embodiment, the sigmoid function is used for normalization processing. The sigmoid function is a well-known technology and will not be elaborated here. As other implementation manners, implementers can adopt other methods of existing technologies, such as the tanh function, etc. This embodiment does not make special restrictions on this; secondly, when calculating the ratio, to avoid the denominator being 0, a preset value greater than 0 is added to the denominator. The preset value greater than 0 is taken as 0.1. As other implementation manners, implementers can set it according to the actual situation.

[0051] It should be noted that the larger the local density is, the more concentrated the point cloud distribution in the neighborhood of the point cloud data is. At this time, a larger filtering neighborhood scale should be adopted to expand more neighborhood points and enhance the filtering effect. The larger the deformation characterization value is, the larger the filtering neighborhood scale is given, which can smooth the deformed area of the inner ring of the oil seal and improve the accuracy of subsequent height detection; the center deviation degree reflects the symmetry characteristics of all point cloud data corresponding to the oil return line within the neighborhood range. The larger the center deviation degree is, the more uneven the distribution of the point cloud data corresponding to the oil return line in the neighborhood is, and the filtering neighborhood scale should be reduced to retain the directional texture features of the point cloud data corresponding to the oil return line and the edge detail features of the threaded oil return line.

[0052] Further, based on the adjustment coefficient, determine the filtering scale of the K-nearest neighbor mean filtering algorithm, specifically: The calculation formula for the filtering scale of each point cloud data corresponding to the oil return line is: Wherein, is the filtering scale of the th point cloud data corresponding to the oil return line, is a preset value, is the preset initial filtering scale, is the adjustment coefficient of the th point cloud data corresponding to the oil return line.

[0053] In this embodiment, the preset initial filtering scale takes the value of 0.1, and the preset value takes the value of 0.5. Among them, the preset value is to balance the filtering intensity and dynamic adjustment ability. The value range of the adjustment coefficient is . When is 0, the filtering scale is the smallest, taking the value of 0.05. When is 1, the filtering scale is the largest, taking the value of 0.15. Through the preset value , the filtering scale of the point cloud data corresponding to the oil return line is adjusted left and right based on the preset initial filtering scale. As other implementation manners, the implementer can set it according to the actual situation.

[0054] It should be noted that for the point cloud data other than the point cloud data corresponding to the oil return line, its filtering scale is set to the preset initial filtering scale; Taking the point cloud data of the inner circle of the oil seal as the center, construct a neighborhood with a radius of the filtering scale, and use the K-nearest neighbor mean filtering algorithm to filter all the point cloud data; It should be noted that the K-nearest neighbor mean filtering algorithm is a well-known technology and will not be elaborated here.

[0055] Perform three-dimensional reconstruction on all the filtered point cloud data to construct a three-dimensional cylinder of the inner circle of the oil seal; In this embodiment, a three-dimensional cylinder of the inner circle of the oil seal is constructed through the Poisson Reconstruction algorithm. Among them, the Poisson Reconstruction algorithm is a well-known technology and will not be elaborated here.

[0056] Obtain the radius of the inner circle of the oil seal; It should be noted that the radius of the inner circle of the oil seal is obtained through the oil seal production specification data.

[0057] Calculate the minimum distance from each point cloud data corresponding to the oil return line on the three-dimensional cylinder to the axis of the three-dimensional cylinder, and take the average value of the difference between the radius of the inner ring of the oil seal and the minimum distance of all point cloud data corresponding to the oil return line as the oil return line height of the oil seal; Select the maximum and minimum values of the difference between the radius of the inner ring of the oil seal and the minimum distance of all point cloud data corresponding to the oil return line, and use them as the maximum oil return line height and the minimum oil return line height of the oil seal respectively; and display the measurement result of the oil return line height of the oil seal on the terminal.

[0058] It should be understood that although Figure 1 each step in the flowchart of Figure 1 is shown in sequence according to the indication of the arrow, these steps are not necessarily executed in the order indicated by the arrow. Unless there is a clear description in this article, there is no strict order limit for the execution of these steps, and these steps can be executed in other orders. Moreover,

[0059] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0060] The above-described embodiments merely represent several implementation manners of the present application. Their descriptions are relatively specific and detailed, but should not be construed as limitations on the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made. Therefore, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present application without departing from the content of the technical solution of the present application all belong to the protection scope of the technical solution of the present application.

Claims

1. A method for detecting the height of the oil return line of an oil seal, characterized in that, The method includes the following steps: Collect all the point cloud data on the inner ring of the oil seal; Perform three-dimensional fitting on all the point cloud data to construct a cylinder corresponding to the oil seal, divide the bottom surface of the cylinder into multiple sector blocks, and divide each point cloud data into each sector block according to the distribution of the polar angles in the polar coordinates of each point cloud data in each sector block; Based on the distance between each point cloud data and the cylinder axis, combined with the radius of the bottom surface of the cylinder, calculate the axis distance; calculate the deformation characterization value of each sector block through the discrete situation and mutation characteristics of the axis distances of the point cloud data between each sector block and the other sector blocks; Analyze the distribution density of all the point cloud data in the neighborhood of each point cloud data to determine the local concentration degree of each point cloud data; combine the axis distance and the deformation characterization value to determine the discrimination coefficient of each point cloud data, screen the point cloud data, and obtain the point cloud data corresponding to the oil return line; Cluster all the point cloud data corresponding to the oil return line, analyze the deviation between each point cloud data corresponding to the oil return line in each cluster and the other point cloud data in its neighborhood, and the distribution of the point cloud data in its neighborhood, and combine the deformation characterization value to determine the adjustment coefficient of each point cloud data corresponding to the oil return line; Based on the adjustment coefficient, determine the filtering scale of each point cloud data corresponding to the oil return line, perform filtering in combination with the filtering algorithm, perform three-dimensional reconstruction on all the filtered point cloud data, and measure the height of the oil return line of the oil seal.

2. The height detection method of an oil seal oil return line according to claim 1, wherein The step of dividing each point cloud data into each sector block includes: Taking the center of the bottom surface of the cylinder as the origin and the cylinder axis as the z-axis, construct a three-dimensional coordinate system; map all the point cloud data on the cylinder to the xOy plane of the three-dimensional coordinate system, and calculate the polar coordinates of each point cloud data; According to the sector block area where the polar angle in the polar coordinates of each point cloud data is distributed, divide all the point cloud data into different sector blocks.

3. The height detection method of the oil seal oil return line according to claim 1, wherein The calculation process of the axis distance is: calculate the nearest distance from each point cloud data to the cylinder axis; take the ratio between the nearest distance and the radius of the bottom surface of the cylinder as the axis distance of each point cloud data.

4. The height detection method of an oil seal oil return line according to claim 1, characterized in that The calculation of the deformation characterization value of each sector block includes: Calculate the discrete degree of the axis distances of all the point cloud data in each sector block; calculate the ratio between the discrete degree of each sector block and the mean value of the discrete degrees of all the sector blocks, and record it as the relative discrete ratio; Perform mutation point detection on the discrete degrees of all the sector blocks to obtain the mutation probability of each sector block; The deformation characterization value is the normalized result of the product of the relative discrete ratio and the mutation probability.

5. The method for detecting the height of the oil return line of an oil seal according to claim 1, wherein The determination of the local concentration degree of each point cloud data includes: Taking each point cloud data as the center, record the neighborhood range with a preset radius as the local neighborhood; Calculate the range of values of the position coordinates corresponding to all the point cloud data in the local neighborhood on the x-axis, y-axis, and z-axis respectively; calculate the average value of the range of values corresponding to all the axes in the local neighborhood; The local concentration degree is the ratio between the average value and the diameter of the local neighborhood.

6. The height detection method of the oil seal oil return line according to claim 1, characterized in that Determining the discrimination coefficient of each point cloud data includes: using the deformation characterization value of the sector block to which each point cloud data belongs as a weight, and performing a weighted sum of the reciprocal of the local concentration and the reciprocal of the axis distance as the discrimination coefficient of each point cloud data.

7. The height detection method of an oil seal oil return line according to claim 1, characterized in that Obtaining the point cloud data corresponding to the oil return line includes: obtaining the segmentation threshold of the discrimination coefficient of all point cloud data; and designating the point cloud data with the discrimination coefficient greater than the segmentation threshold as the point cloud data corresponding to the oil return line.

8. The method for detecting the height of the oil return line of an oil seal according to claim 1, wherein Determining the adjustment coefficient of each point cloud data corresponding to the oil return line includes: By clustering all the point cloud data corresponding to the oil return line, obtaining the local density of each point cloud data corresponding to the oil return line, where the truncation distance of the clustering algorithm is a preset value; Selecting, from the neighborhood corresponding to the truncation distance of any point cloud data corresponding to the oil return line, the mean value of the position coordinates of all the remaining point cloud data corresponding to the oil return line within the same clustering cluster as the any point cloud data, and denoting it as the neighborhood center coordinate; Calculating the distance between the position coordinate of the any point cloud data and the neighborhood center coordinate, and denoting it as the center deviation; Calculating the product value between the local density of each point cloud data corresponding to the oil return line and the deformation characterization value of the sector block to which it belongs; The adjustment coefficient is the normalized result of the ratio of the product value to the center deviation.

9. The method for detecting the height of the oil return line of an oil seal according to claim 1, wherein, The filtering scale of the point cloud data corresponding to the oil return line is calculated by the formula: , where is a preset value, is the preset initial filtering scale, is the adjustment coefficient of the point cloud data corresponding to the oil return line.

10. The height detection method of an oil seal oil return line according to claim 1, characterized in that Measuring the oil return line height of the oil seal includes: according to the three-dimensional cylinder after three-dimensional reconstruction, calculating the minimum distance from each point cloud data corresponding to the oil return line on the three-dimensional cylinder to the axis of the three-dimensional cylinder, and taking the mean value of the difference between the radius of the inner circle of the oil seal and the minimum distance of all the point cloud data corresponding to the oil return line as the oil return line height of the oil seal.

Images