A method, device and system for detecting the shape and surface of automobile parts

Abstract

The present invention relates to the technical field of component surface detection, and specifically to a method, device and system for detecting the surface of automobile components. The present invention first obtains the reflective interference degree of the block area according to the corresponding curvature values ​​of all data points in the block area and the reflective score value; obtains the defect performance measurement of the block area according to the curvature value of the block area and other block areas; uses the reflective interference degree and defect performance measurement of the block area to adjust the Gaussian filter standard deviation to obtain the adjusted Gaussian filter standard deviation of each block area; performs three-dimensional Gaussian filtering on the initial point cloud data to obtain the point cloud filter data of the automobile component; and performs surface quality detection of the automobile component according to the point cloud filter data. The present invention adjusts the standard deviation of the Gaussian filter by the noise level and defect performance level of different areas in the initial point cloud data to improve the denoising effect and improve the accuracy of the surface quality detection of automobile components.

Description

Technical Field

[0001] The present invention relates to the technical field of component surface detection, and in particular to a method, device and system for detecting the surface of an automobile component. Background Art

[0002] Auto parts are the basic units that make up the whole car. Each auto part has a specific function and role to ensure the normal operation of the car. Auto parts are an important part of the whole car, which is related to ensuring the normal operation and driving safety of the car. In the automobile manufacturing industry that pursues high precision and consistency, the surface detection of auto parts is used to verify whether the surface shape of auto parts meets the design requirements. Usually, point cloud data of auto parts is collected for surface detection.

[0003] Since the initial point cloud data of automobile parts contains noise, the noise will lead to inaccurate surface quality inspection of automobile parts. Therefore, it is necessary to denoise the initial point cloud data of automobile parts. The existing technology uses three-dimensional Gaussian filtering to denoise the point cloud data. The existing technology sets the same Gaussian filter standard deviation for all data points based on empirical values, but does not fully consider the differences in noise performance and defect performance in different regions, which makes the Gaussian filter standard deviation unreasonable, resulting in poor denoising effect on the initial point cloud data, affecting the accuracy of surface quality inspection of automobile parts. Summary of the invention

[0004] In order to solve the technical problem that the prior art has poor denoising effect on initial point cloud data, which affects the accuracy of automobile parts surface quality detection, the purpose of the present invention is to provide an automobile parts surface detection method, device and system, and the technical scheme adopted is as follows:

[0005] A method for detecting the shape and surface of an automobile component, the method comprising:

[0006] Obtaining initial point cloud data and reflective score values ​​of automotive parts; each data point in the initial point cloud data has a corresponding curvature value;

[0007] The initial point cloud data of the automobile parts are divided into blocks to obtain each block area; the reflective interference degree of the block area is obtained according to the corresponding curvature values ​​and the reflective score values ​​of all data points in the block area; the defect performance metric of the block area is obtained according to the curvature values ​​of the block area and other block areas; the Gaussian filter standard deviation is adjusted by using the reflective interference degree and the defect performance metric of the block area to obtain the adjusted Gaussian filter standard deviation of each block area;

[0008] According to the adjusted Gaussian filter standard deviation of each of the block areas, the initial point cloud data is subjected to three-dimensional Gaussian filtering to obtain point cloud filter data of the automobile parts; and the surface quality inspection of the automobile parts is performed based on the point cloud filter data.

[0009] Furthermore, the method for obtaining the block area includes:

[0010] Using the k-means clustering method, all data points in the initial point cloud data of automobile parts are clustered according to the Euclidean distance between every two data points to obtain each block area.

[0011] Furthermore, the method for obtaining the reflection interference degree includes:

[0012] According to the fluctuation of the curvature values ​​corresponding to all data points in the block area, the non-smoothness index of the block area is obtained;

[0013] The non-smoothness index of the block area and the reflection score value are forwardly fused to obtain the reflection interference degree of the block area.

[0014] Furthermore, the method for obtaining the defect performance metric includes:

[0015] Obtaining a first defect index of the block area according to a difference in curvature values ​​between the block area and its adjacent block areas;

[0016] Obtaining a second defect index of the block area according to fluctuations in the curvature values ​​of all data points in the block area;

[0017] Obtaining a third defect index of the block area according to the mismatch degree of the block area with all other block areas;

[0018] The first defect indicator, the second defect indicator, and the third defect indicator are forwardly integrated to obtain a defect manifestation metric for a block area.

[0019] Furthermore, the method for obtaining the first defect indicator includes:

[0020] Calculate the mean of the curvature values ​​of all data points in the block area to obtain the roughness index of the block area;

[0021] The absolute value of the difference between the unevenness index of the block area and each of its adjacent block areas is calculated, and all the absolute values ​​of the difference are averaged to obtain the first defect index of the block area.

[0022] Furthermore, the method for obtaining the second defect indicator includes:

[0023] The variance of the curvature values ​​of all data points in the block area is calculated to obtain the second defect index of the block area.

[0024] Furthermore, the method for obtaining the third defect indicator includes:

[0025] Each block region except the block region is used as each reference region of the block region; the average Hausdorff distance between the block region and all the reference regions is calculated to obtain the third defect index of the block region.

[0026] Furthermore, the method for obtaining the adjusted Gaussian filter standard deviation includes:

[0027] For any block area, calculate the sum of the defect performance metric and the preset denominator adjustment factor, calculate the ratio of the reflective interference degree to the sum and perform normalization processing to obtain the adjustment parameter of the block area; calculate the product of the adjustment parameter and the preset adjustment value to obtain the target adjustment index of the block area;

[0028] The product of the target adjustment index and the Gaussian filter standard deviation is calculated to obtain the adjusted Gaussian filter standard deviation of the block area.

[0029] A surface detection device for automobile parts, comprising:

[0030] A data acquisition module is used to acquire initial point cloud data and reflective score values ​​of automotive parts; each data point in the initial point cloud data has a corresponding curvature value;

[0031] The Gaussian filter standard deviation adjustment module is used to divide the initial point cloud data of automobile parts into blocks to obtain each block area; obtain the reflective interference degree of the block area according to the corresponding curvature values ​​and the reflective score values ​​of all data points in the block area; obtain the defect performance metric of the block area according to the curvature values ​​of the block area and other block areas; adjust the Gaussian filter standard deviation using the reflective interference degree and the defect performance metric of the block area to obtain the adjusted Gaussian filter standard deviation of each block area;

[0032] The detection module is used to perform three-dimensional Gaussian filtering on the initial point cloud data according to the adjusted Gaussian filtering standard deviation of each of the block areas to obtain the point cloud filtering data of the automobile parts; and perform surface quality detection of the automobile parts according to the point cloud filtering data.

[0033] A system for detecting the shape and surface of automobile parts comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the method for detecting the shape and surface of automobile parts as described in any one of the above items are implemented.

[0034] The present invention has the following beneficial effects:

[0035] Considering that the noise performance degree and defect performance degree of different areas in the initial point cloud data are different, the initial point cloud data of automobile parts are first divided into blocks to obtain each block area, so as to realize the subsequent regional analysis of the initial point cloud data. The reflective noise interference degree of the block area is quantified by the reflective interference degree, and the defect performance degree of the block area is quantified by the defect performance metric. Considering that the area with greater reflective interference requires a larger Gaussian filter standard deviation to remove noise, it is also considered that the larger the defect performance metric of the block area, the more obvious the defect, and more sophisticated filtering may be required to retain the defect information, that is, a smaller Gaussian filter standard deviation is required. The reflective interference degree and defect performance metric of the block area are used to adjust the Gaussian filter standard deviation to obtain the adjusted Gaussian filter standard deviation of each block area. The present invention improves the denoising effect and the accuracy of the surface quality detection of automobile parts by adjusting the standard deviation of the Gaussian filter. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0037] Figure 1 A flow chart of a method for detecting the shape and surface of an automobile component provided by one embodiment of the present invention;

[0038] Figure 2 A flow chart of a method for obtaining the degree of reflection interference provided by an embodiment of the present invention;

[0039] Figure 3 A flow chart of a method for obtaining defect performance metrics provided by an embodiment of the present invention;

[0040] Figure 4 A structural diagram of an automobile component surface detection system provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0041] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the method, device and system for detecting the shape and surface of an automobile part proposed by the present invention, its specific implementation, structure, features and effects in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0042] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0043] The specific scheme of the automobile parts surface detection method, device and system provided by the present invention is described in detail below with reference to the accompanying drawings.

[0044] The present invention provides a method, device and system for detecting the shape and surface of automobile parts. Figure 1 , which shows a flow chart of a method for detecting the shape and surface of an automobile part provided by an embodiment of the present invention, the method comprising the following steps:

[0045] Step S1: Obtain initial point cloud data and reflective score values ​​of automobile parts; each data point in the initial point cloud data has a corresponding curvature value.

[0046] The initial point cloud data of the automobile parts and the curvature value corresponding to each data point in the initial point cloud data are obtained from the automobile parts detection system. Specifically, in order to perform surface detection on automobile parts, a suitable three-dimensional scanner is selected to scan the automobile parts to obtain the initial point cloud data of the automobile parts. It should be noted that the specific method of using a three-dimensional scanner to scan automobile parts and obtain the initial point cloud data of automobile parts is a technical means of the technicians in the field, and only a brief description is given here: the three-dimensional scanner is used to perform an all-round scan around the automobile parts to ensure that the data is captured from different angles. After the scanning is completed, the software will generate three-dimensional point cloud data of the automobile parts, and the three-dimensional point cloud data is the initial point cloud data of the automobile parts. Then, each data point in the initial point cloud data has a corresponding curvature value. It should be noted that the method for obtaining the curvature value is a technical means well known to the technicians in the field, and the curvature value can be obtained by the least squares method. The least squares method is a technical means well known to the technicians in the field, and it will not be repeated here.

[0047] Obtain the reflective score value of the automobile parts from the automobile parts detection system. Specifically, in one embodiment of the present invention, the engineer will score the reflective level of the automobile parts according to the material properties of the automobile parts, the surface processing technology of the parts, the surface coating information, the surface color, etc., and obtain the reflective score value of the automobile parts, wherein the scoring values ​​are: 0.1, 0.2, 0.3, ... 1, a total of 10 score types, corresponding to 10 reflective levels. In other embodiments of the present invention, the reflective intensity value recorded by the three-dimensional scanner at each data point during collection can also be used to calculate the mean of the reflective intensity values ​​of all data points in the initial data of the point cloud and perform normalization processing to obtain the reflective score value of the automobile parts. It should be noted that the normalization method used is: normalization is performed using the norm normalization function to limit the value range to between 0 and 1. Wherein normalization is a technical means well known to those skilled in the art, and the selection of the normalization function can be linear normalization or standard normalization, etc., and the specific normalization method is not limited here.

[0048] It should be noted that the initial point cloud data is set to a three-dimensional space coordinate system, the XY plane of the three-dimensional space coordinate is parallel to the ground, and the Z axis is perpendicular to the ground. It should be noted that in order to facilitate calculations, all indicator data involved in the calculations in the embodiments of the present invention are pre-processed to eliminate the dimension effect. The specific means of removing the dimension effect are technical means well known to those skilled in the art and are not limited here.

[0049] Due to the noise interference in the initial point cloud data, the initial point cloud data is denoised by three-dimensional Gaussian filtering. In three-dimensional Gaussian filtering, the setting of the Gaussian filter standard deviation has a significant impact on the filtering effect, that is, when the Gaussian filter standard deviation is set too large, although more noise can be removed, the detailed information of the parts will also be smoothed out, resulting in inaccurate surface quality detection. When the Gaussian filter standard deviation is set too small, although more detailed information can be retained, the noise suppression ability is insufficient and the denoising effect is not good. The prior art is to set the same Gaussian filter standard deviation for the initial point cloud data based on empirical values, which can quickly perform preliminary denoising on the point cloud data. Since the noise performance and defect performance of different regions in the initial point cloud data are significantly different, it is often difficult to achieve an ideal denoising effect by using a unified Gaussian filter standard deviation. The present invention adjusts the standard deviation of the Gaussian filter according to the noise level and defect performance of different regions in the initial point cloud data to improve the denoising effect and the accuracy of the surface quality detection of automotive parts.

[0050] Step S2: Divide the initial point cloud data of automobile parts into blocks to obtain each block area; obtain the reflective interference degree of the block area according to the corresponding curvature values ​​and reflective score values ​​of all data points in the block area; obtain the defect performance metric of the block area according to the curvature values ​​of the block area and other block areas; adjust the Gaussian filter standard deviation using the reflective interference degree and defect performance metric of the block area to obtain the adjusted Gaussian filter standard deviation of each block area.

[0051] Considering that the noise performance and defect performance levels in different areas of the initial point cloud data are different, the initial point cloud data of automotive parts are first divided into blocks to obtain each block area, so as to realize the subsequent regional analysis of the initial point cloud data. The reflective noise interference level of the block area is quantified by the reflective interference degree, and the defect performance level of the block area is quantified by the defect performance metric. Considering that the area with greater reflective interference requires a larger Gaussian filter standard deviation to remove noise, it is also considered that the larger the defect performance metric of the block area, the more obvious the defect, and more sophisticated filtering may be required to retain the defect information, that is, a smaller Gaussian filter standard deviation is required. The reflective interference level and defect performance metric of the block area are used to adjust the Gaussian filter standard deviation to obtain the adjusted Gaussian filter standard deviation of each block area.

[0052] In order to process the point cloud data of automobile parts in blocks, preferably, in one embodiment of the present invention, the method for obtaining the block area includes:

[0053] Using the k-means clustering method, all data points in the initial point cloud data of automobile parts are clustered according to the Euclidean distance between every two data points to obtain each block area.

[0054] It should be noted that the k-means clustering algorithm is a technical means well known to those skilled in the art, and will not be described in detail here. Only a brief description is given of a brief process of using the k-means clustering algorithm in one embodiment of the present invention to cluster all data points in the initial point cloud data of automobile parts according to the Euclidean distance between every two data points to obtain each block area:

[0055] The optimal k value is obtained by the elbow method, and the Euclidean distance between every two data points in the initial point cloud data is used as the distance measurement value of the two data points, which is substituted into the clustering process to cluster all data points in the initial point cloud data of automobile parts to obtain the clustering results, i.e., each block area. In other embodiments of the present invention, the optimal k value of the k-means clustering algorithm can be determined by other methods such as the silhouette coefficient method, which is not limited here. It should be noted that the Euclidean distance is a technical means well known to those skilled in the art and will not be elaborated here.

[0056] For automotive parts with highly reflective features, the main source of noise in the initial point cloud data is reflective interference. When scanning using 3D scanning technology, that is, scanning using LiDAR, the laser beam will illuminate the uneven areas of the automotive parts at different incident angles, which will cause the incident angles of the laser beam to vary significantly. According to the law of reflection of light, the intensity and direction of the reflected light will change with the change of the incident angle. Therefore, when the reflected light is received by the receiver of the scanning device, its intensity distribution may be very uneven. This uneven reflected light will cause the reflective intensity value of the data point to fluctuate, and then generate reflective interference noise in the point cloud data. To quantify the degree of reflective noise interference, please refer to Figure 2 , which shows a flow chart of a method for obtaining the reflective interference degree in one embodiment of the present invention. Preferably, in one embodiment of the present invention, the method for obtaining the reflective interference degree of the block area includes:

[0057] Step S201: Obtain a non-smoothness index of the block area according to fluctuations of curvature values ​​corresponding to all data points in the block area.

[0058] Considering that the curvature value reflects the curvature degree of the object surface, the curvature value of the rough area of ​​the automobile parts will change significantly; for the smooth area of ​​the automobile parts, the curvature value remains basically unchanged; according to the fluctuation of the curvature value corresponding to all data points in the block area, the roughness index of the block area is obtained. The roughness index quantifies the roughness degree of the block area. The larger the roughness index, the less smooth the block area.

[0059] In one embodiment of the present invention, the standard deviation of the curvature values ​​corresponding to all data points in the block region is calculated to obtain the non-smoothness index of the block region, wherein the standard deviation reflects the degree of data fluctuation.

[0060] Step S202: forwardly fuse the non-smoothness index and the reflection score value of the block area to obtain the reflection interference degree of the block area.

[0061] Considering that the greater the degree of reflection and the greater the degree of roughness in the block area, the easier it is to generate reflection interference and thus generate noise in the point cloud data, the reflection interference degree of the block area is obtained. The reflection interference degree more comprehensively reflects the degree of reflection interference in the block area. The greater the reflection interference degree, the greater the degree of reflection noise interference in the block area.

[0062] It should be noted that forward fusion is an existing technology well known to those skilled in the art, and forward fusion can adopt simple product, arithmetic mean or other suitable fusion methods. In one embodiment of the present invention, the product of the unsmoothness index of the block area and the reflection score value is calculated to obtain the reflection interference degree of the block area.

[0063] Considering that there are normal structural areas and defective areas in automotive parts, the normal structural area is the part of the automotive parts that is precisely manufactured according to the design requirements. The size and surface characteristics meet the design requirements and show a high degree of regularity. The defective area is caused by errors in the processing process. The surface of the defective area often has obvious uneven features and the defective area has sporadic and irregular characteristics. In order to measure the defect performance of the block area, refer to Figure 3 , which shows a flow chart of a method for obtaining a defect performance metric in one embodiment of the present invention. Preferably, in one embodiment of the present invention, the method for obtaining a defect performance metric includes:

[0064] Step S211: obtaining a first defect index of a block region according to a difference in curvature values ​​between the block region and its adjacent block regions.

[0065] According to the difference in curvature values ​​between the block region and its adjacent block regions, a first defect index of the block region is obtained. The first defect index is used to preliminarily quantify the possibility of defects in the block region.

[0066] In one embodiment of the present invention, the mean value of the curvature values ​​of all data points in the block area is calculated to obtain the unevenness index of the block area; the absolute value of the difference between the unevenness index of the block area and each adjacent block area is calculated, and all the absolute values ​​of the difference are averaged to obtain the first defect index of the block area. The absolute value of the difference is used to reflect the difference.

[0067] Regarding the above steps, when there are defects in auto parts, the regular features of the auto parts are destroyed, that is, the defective area becomes uneven and the shape becomes irregular. This change will cause a significant difference in the curvature value between the defective area and the adjacent block area. The overall unevenness of the block area is quantified by the unevenness index. The larger the unevenness index, the greater the overall unevenness of the block area. The first defect index reflects the difference in unevenness between the block area and its adjacent block areas. The larger the difference, the higher the possibility that the block area has defects.

[0068] Step S212: obtaining a second defect index of the block area according to the fluctuation of the curvature values ​​of all data points in the block area.

[0069] Considering that the large fluctuation of the curvature value reflects the existence of uneven features inside the block area, the second defect index is used to quantify the uneven features of the defect area.

[0070] In one embodiment of the present invention, the variance of the curvature values ​​of all data points in the block area is calculated to obtain the second defect index of the block area, wherein the variance is used to reflect the fluctuation of the curvature value.

[0071] Step S213: obtaining a third defect index of the block region according to the mismatch degree between the block region and all other block regions.

[0072] The larger the third defect index is, the more accidental the shape characteristics of the block area are in the overall automotive parts, and the higher the possibility of defects in the block area is.

[0073] In one embodiment of the present invention, in the initial point cloud data, each block area except the block area is used as each reference area of ​​the block area; the average of the Hausdorff distances between the block area and all reference areas is calculated to obtain a third defect index of the block area. It should be noted that the Hausdorff distance is a technology well known to those skilled in the art, and the method for obtaining the Hausdorff distance is a technical means well known to those skilled in the art, which will not be described in detail here. The Hausdorff distance is a metric for measuring the shape difference between two sets, and can be used to evaluate the degree of mismatch between a block area and other block areas.

[0074] For the above steps, the Hausdorff distance between the block area and each reference area reflects the degree of mismatch between the block area and the reference area. The third defect index of the block area reflects the degree of difference in shape between the current block area and all other block areas. The greater the difference, the more likely it is that the block area has significant shape defects.

[0075] Step S214: forwardly fuse the first defect index, the second defect index, and the third defect index to obtain the defect performance metric of the block area.

[0076] The defect manifestation metric is used to comprehensively evaluate the defect manifestation degree of the block area.

[0077] It should be noted that forward fusion is an existing technology well known to those skilled in the art, and forward fusion can adopt simple product, arithmetic mean or other suitable fusion methods. In one embodiment of the present invention, the mean of the first defect index, the second defect index and the third defect index is calculated to obtain the defect performance metric of the block area.

[0078] Considering that the area with greater reflective interference requires a larger Gaussian filter standard deviation to remove noise, and also considering that the larger the defect performance metric of the block area, the more obvious the defect, and more sophisticated filtering may be required to retain the defect information, that is, a smaller Gaussian filter standard deviation is required. The Gaussian filter standard deviation is adjusted using the reflective interference and defect performance metrics of the block area to obtain the adjusted Gaussian filter standard deviation of each block area. Preferably, in one embodiment of the present invention, the method for obtaining the adjusted Gaussian filter standard deviation includes:

[0079] For any block area, calculate the sum of the defect performance metric and the preset denominator adjustment factor, calculate the ratio of the reflective interference degree to the sum and perform normalization processing to obtain the adjustment parameter of the block area; calculate the product of the adjustment parameter and the preset adjustment value to obtain the target adjustment index of the block area; calculate the product of the target adjustment index and the Gaussian filter standard deviation to obtain the adjusted Gaussian filter standard deviation of the block area. In one embodiment of the present invention, the preset adjustment value is 1.1, and the implementer can set it according to the implementation scenario. The preset denominator adjustment factor is used to prevent the denominator from being 0. In one embodiment of the present invention, the preset denominator adjustment factor is set to 0.01. The initial Gaussian filter standard deviation is set based on the empirical value. In one embodiment of the present invention, the Gaussian filter standard deviation is set to 1.5, and the implementer can set it according to the implementation scenario. It should be noted that the normalization method is: normalization is performed using the norm normalization function to limit the value range to between 0 and 1. Normalization is a technical means well known to those skilled in the art, and the normalization function may be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0080] For the above steps, the adjustment parameters reflect the comprehensive impact of the reflective interference and defect performance metrics on the Gaussian filter standard deviation adjustment. The preset adjustment value is a constant determined based on experience or experiment, which is used to control the degree of adjustment; the product of the adjustment parameter and the preset adjustment value is calculated to obtain the target adjustment index of the block area; the product of the target adjustment index and the Gaussian filter standard deviation is calculated to obtain the adjusted Gaussian filter standard deviation of the block area. The adjusted Gaussian filter standard deviation is used to filter the initial point cloud data of the block area.

[0081] Step S3: Perform three-dimensional Gaussian filtering on the initial point cloud data according to the adjusted Gaussian filtering standard deviation of each block area to obtain point cloud filtering data of automobile parts; and perform surface quality inspection of automobile parts according to the point cloud filtering data.

[0082] By adjusting the standard deviation of Gaussian filtering, the denoising effect can be improved and the accuracy of surface quality inspection of automobile parts can be improved.

[0083] In order to perform three-dimensional Gaussian filtering on the initial point cloud data, preferably, in one embodiment of the present invention, the method for obtaining the point cloud filtering data includes:

[0084] In the initial point cloud data, three-dimensional Gaussian filtering is performed on each block area according to the adjusted Gaussian filtering standard deviation of each block area to obtain point cloud filtering data of automobile parts. It should be noted that the three-dimensional Gaussian filtering method is a technical means well known to those skilled in the art and will not be described in detail here.

[0085] In one embodiment of the present invention, a method for detecting the surface quality of automobile parts includes:

[0086] Methods for surface quality inspection of automotive parts include: using professional software, such as TerraSolid software, to stitch the point cloud filter data of automotive parts to form a complete three-dimensional model of automotive parts. Compare the scanned three-dimensional model with the designed CAD model to analyze the surface deviation. The comparison software can display color-coded deviation maps to intuitively show which areas are beyond the tolerance range; based on the comparison results, the surface quality inspection results of automotive parts are generated, including dimensional deviation, surface flatness, geometric shape accuracy, etc., to provide a basis for subsequent production adjustments or quality control, and complete the surface inspection of automotive parts.

[0087] The present invention also proposes a device for detecting the shape and surface of automobile parts. Figure 4 , which shows a structural diagram of an automobile component surface detection device provided by an embodiment of the present invention. The device includes: a data acquisition module 101, a Gaussian filter standard deviation adjustment module 102 and a detection module 103.

[0088] The data acquisition module is used to obtain the initial point cloud data and reflective score values ​​of automotive parts; each data point in the initial point cloud data has a corresponding curvature value;

[0089] The Gaussian filter standard deviation adjustment module is used to divide the initial point cloud data of automobile parts into blocks to obtain each block area; obtain the reflective interference degree of the block area according to the corresponding curvature values ​​and reflective score values ​​of all data points in the block area; obtain the defect performance measurement of the block area according to the curvature value of the block area and other block areas; adjust the Gaussian filter standard deviation using the reflective interference degree and defect performance measurement of the block area to obtain the adjusted Gaussian filter standard deviation of each block area;

[0090] The detection module is used to perform three-dimensional Gaussian filtering on the initial point cloud data according to the adjusted Gaussian filtering standard deviation of each block area to obtain the point cloud filtering data of the automobile parts; and perform surface quality detection of the automobile parts according to the point cloud filtering data.

[0091] It should be noted that: the device provided in the above embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, 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, the automotive component surface detection device and the automotive component surface detection method embodiment provided in the above embodiment belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0092] A system for detecting the shape and surface of automobile parts comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of any one of the above-mentioned methods for detecting the shape and surface of automobile parts are implemented.

[0093] In summary, the embodiments of the present invention provide a method, device and system for detecting the shape and surface of automobile parts. First, the initial point cloud data of automobile parts are divided into blocks to obtain each block area; the reflective interference degree of the block area is obtained according to the corresponding curvature values ​​and reflective score values ​​of all data points in the block area; the defect performance metric of the block area is obtained according to the curvature value of the block area and the adjacent block area; the Gaussian filter standard deviation is adjusted by using the reflective interference degree and defect performance metric of the block area to obtain the adjusted Gaussian filter standard deviation of each block area; according to the Gaussian filter standard deviation of each block area, the initial point cloud data is subjected to three-dimensional Gaussian filtering to obtain the point cloud filter data of automobile parts; according to the point cloud filter data, the shape and surface quality detection of automobile parts is performed. In the embodiments of the present invention, the standard deviation of the Gaussian filter is adjusted according to the noise level and defect performance level of different areas in the initial point cloud data to improve the denoising effect and improve the accuracy of the shape and surface quality detection of automobile parts.

[0094] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying 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.

[0095] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for detecting the shape and surface of automobile parts, characterized in that: The method comprises: Obtaining initial point cloud data and reflective score values ​​of automotive parts; each data point in the initial point cloud data has a corresponding curvature value; The initial point cloud data of the automobile parts are divided into blocks to obtain each block area; the reflective interference degree of the block area is obtained according to the corresponding curvature values ​​and the reflective score values ​​of all data points in the block area; the defect performance metric of the block area is obtained according to the curvature values ​​of the block area and other block areas; the Gaussian filter standard deviation is adjusted by using the reflective interference degree and the defect performance metric of the block area to obtain the adjusted Gaussian filter standard deviation of each block area; According to the adjusted Gaussian filter standard deviation of each of the block areas, the initial point cloud data is subjected to three-dimensional Gaussian filtering to obtain point cloud filter data of the automobile parts; and the shape and surface quality inspection of the automobile parts is performed according to the point cloud filter data; The method for obtaining the adjusted Gaussian filter standard deviation includes: For any block area, calculate the sum of the defect performance metric and the preset denominator adjustment factor, calculate the ratio of the reflective interference degree to the sum and perform normalization processing to obtain the adjustment parameter of the block area; calculate the product of the adjustment parameter and the preset adjustment value to obtain the target adjustment index of the block area; The product of the target adjustment index and the Gaussian filter standard deviation is calculated to obtain the adjusted Gaussian filter standard deviation of the block area.

2. The method for detecting the shape and surface of automobile parts according to claim 1, characterized in that: The method for obtaining the block area includes: Using the k-means clustering method, all data points in the initial point cloud data of automobile parts are clustered according to the Euclidean distance between every two data points to obtain each block area.

3. The method for detecting the shape and surface of automobile parts according to claim 1, characterized in that: The method for obtaining the reflection interference degree includes: According to the fluctuation of the curvature values ​​corresponding to all data points in the block area, the non-smoothness index of the block area is obtained; The non-smoothness index of the block area and the reflection score value are forwardly fused to obtain the reflection interference degree of the block area.

4. The method for detecting the shape and surface of automobile parts according to claim 1, characterized in that: The method for obtaining the defect performance metric includes: Obtaining a first defect index of the block area according to a difference in curvature values ​​between the block area and its adjacent block areas; Obtaining a second defect index of the block area according to fluctuations in the curvature values ​​of all data points in the block area; According to the mismatch degree of the block area with all other block areas, a third defect index of the block area is obtained; The first defect indicator, the second defect indicator, and the third defect indicator are forwardly integrated to obtain a defect manifestation metric for a block area.

5. A method for detecting the shape and surface of automobile parts according to claim 4, characterized in that: The method for obtaining the first defect indicator includes: Calculate the mean of the curvature values ​​of all data points in the block area to obtain the roughness index of the block area; The absolute value of the difference between the unevenness index of the block area and each of its adjacent block areas is calculated, and all the absolute values ​​of the difference are averaged to obtain the first defect index of the block area.

6. A method for detecting the shape and surface of automobile parts according to claim 4, characterized in that: The method for obtaining the second defect indicator includes: The variance of the curvature values ​​of all data points in the block area is calculated to obtain the second defect index of the block area.

7. The method for detecting the shape and surface of automobile parts according to claim 4, characterized in that: The method for obtaining the third defect indicator includes: Each block region except the block region is used as each reference region of the block region; the average Hausdorff distance between the block region and all the reference regions is calculated to obtain the third defect index of the block region.

8. An automobile parts surface detection device, characterized in that: include: The data acquisition module is used to obtain the initial point cloud data and reflective score values ​​of automotive parts; Each data point in the initial point cloud data has a corresponding curvature value; Gaussian filter standard deviation adjustment module, used to divide the initial point cloud data of automobile parts into blocks and obtain each block area; Obtaining the reflective interference degree of the block area according to the curvature values ​​and the reflective score values ​​corresponding to all data points in the block area; Obtaining a defect performance metric of the block area according to curvature values ​​of the block area and other block areas; Using the reflective interference degree and the defect performance metric of the block area, adjusting the Gaussian filter standard deviation to obtain the adjusted Gaussian filter standard deviation of each block area; The method for obtaining the adjusted Gaussian filter standard deviation includes: for any block area, calculating the sum of the defect performance metric and the preset denominator adjustment factor, calculating the ratio of the reflective interference degree to the sum and performing normalization processing to obtain the adjustment parameter of the block area; calculating the product of the adjustment parameter and the preset adjustment value to obtain the target adjustment index of the block area; calculating the product of the target adjustment index and the Gaussian filter standard deviation to obtain the adjusted Gaussian filter standard deviation of the block area; The detection module is used to perform three-dimensional Gaussian filtering on the initial point cloud data according to the adjusted Gaussian filtering standard deviation of each of the block areas to obtain the point cloud filtering data of the automobile parts; and perform surface quality detection of the automobile parts according to the point cloud filtering data.

9. An automobile parts surface detection system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method for detecting the shape and surface of an automobile part as claimed in any one of claims 1 to 7 are implemented.

Images