A wheel cover detection method and system based on three-dimensional modeling

Abstract

The present invention relates to the field of automatic detection, and specifically to a wheel cover detection method and system based on three-dimensional modeling, which constructs a three-dimensional model of the wheel cover by blue light scanning. Based on the shape characteristics of the wheel cover, the present application first extracts the center of the wheel cover, and then performs a rough inspection based on the center of the wheel cover and the rough inspection template, so as to verify the depth information of the wheel cover as a whole. After passing the rough inspection, the fine inspection process is entered. In the fine inspection process, the wheel cover is divided into multiple inspection areas on the horizontal plane, and surface defect inspection is performed on each area separately. In addition, the shape and size of the wheel cover are also detected based on the outer ring area. For the buckle structure with a complex structure, the present application performs detection by extracting the corresponding cross-sectional structure. The present application can realize a fully automatic detection process for the wheel cover structure, which has higher detection efficiency and detection accuracy compared to manual detection.

Description

Technical Field

[0001] The invention relates to the field of automatic detection, and in particular to a wheel cover detection method and system based on three-dimensional modeling. Background Art

[0002] A wheel cover (also called a wheel decorative cover or hub cover) is a component on a car wheel. It not only serves to beautify the car, but also protects the wheel hub from damage.

[0003] Wheel cover inspection needs to include size inspection, appearance inspection, and connection component defect inspection, etc. The existing inspection methods generally use manual inspection or instrument inspection in a special fixture. The inspection efficiency is low and the accuracy is not enough. Summary of the invention

[0004] In view of this, an object of the present invention is to provide a wheel cover detection method and system based on three-dimensional modeling to solve the problems in the background technology.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] A wheel cover detection method based on three-dimensional modeling of the present invention comprises the following steps:

[0007] Perform blue light scanning at multiple angles on the wheel cover to be inspected to obtain a three-dimensional model of the wheel cover;

[0008] Extracting features from the three-dimensional model to obtain the center of the wheel cover;

[0009] Performing a rough inspection on the wheel cover to be inspected based on the wheel cover center and a pre-constructed rough inspection template to obtain a first inspection result;

[0010] When the first detection result is normal, aligning the pre-constructed mask with the front and back sides of the three-dimensional model based on the center of the wheel cover, and extracting multiple areas to be detected of the three-dimensional model based on the aligned masks, wherein the areas to be detected include an outer ring area, an inner ring area, a banner area, and a buckle area;

[0011] Extracting the size features and shape features of the outer ring area, extracting the depth information of the area to be detected, and extracting the cross-sectional contour features of the buckle area along a preset cross-sectional line;

[0012] Based on the features of the pre-built wheel cover standard model, the size features of the outer ring area, the roundness features of the outer ring area, the depth information of the area to be inspected, and the cross-sectional contour features, the wheel cover to be inspected is subjected to shape detection, size detection, surface defect detection, and buckle detection to obtain a second detection result.

[0013] In one embodiment of the present application, shape detection, size detection, surface defect detection and buckle detection are performed on the wheel cover to be detected based on the features of the pre-constructed wheel cover standard model, the size features of the outer ring area, the roundness features of the outer ring area, the depth information of the area to be detected, and the cross-sectional profile features, to obtain a second detection result, including:

[0014] Performing roundness detection on the wheel cover to be detected based on the shape characteristics of the outer ring area to obtain a shape detection result;

[0015] Performing outer ring size detection on the wheel cover to be detected based on the pre-established outer ring standard size and the size of the outer ring area to obtain a size detection result;

[0016] Performing surface defect detection on the wheel cover to be detected based on the pre-constructed depth reference information and the depth information of the area to be detected to obtain a surface detection result;

[0017] The cross-sectional profile feature is compared with a pre-constructed standard buckle cross-sectional profile for similarity to obtain a buckle detection result.

[0018] In one embodiment of the present application, feature extraction is performed on the three-dimensional model to obtain the wheel cover center, including:

[0019] Extracting a front two-dimensional contour and a back two-dimensional contour of the three-dimensional model;

[0020] Filtering out the outer contour in the front two-dimensional contour and the outer contour in the back two-dimensional contour based on the contour size of the pre-built configuration;

[0021] Extract the first outer contour center point in the front two-dimensional contour and the second outer contour center point in the back two-dimensional contour, and construct a central axis based on the first outer contour center point and the second outer contour center point, and use the central axis as the wheel cover center.

[0022] In one embodiment of the present application, a rough inspection is performed on the wheel cover to be inspected based on the center of the wheel cover and a pre-constructed rough inspection template to obtain a first inspection result, including:

[0023] Extracting a front image and a back image of the three-dimensional model;

[0024] Performing grayscale conversion on the front image and the back image respectively to obtain a front grayscale image and a back grayscale image;

[0025] Get the depth values ​​of multiple points on the front and back of the 3D model , and based on the depth values ​​of multiple points The grayscale values ​​of the corresponding positions of the front grayscale image and the back grayscale image Reassign the value to get the depth image , wherein the depth image The depth value is linearly proportional or inversely proportional to the grayscale value;

[0026] Based on the wheel cover center, the depth image With the rough inspection template Overlap and calculate the overlapping depth image With rough inspection template The difference image , and calculate the difference image The sum of the gray values ​​of all pixels in ;

[0027] When the sum of the grayscale values ​​is greater than a preset threshold, the depth image is converted based on the center of the wheel cover. Rotate target angle , and return to calculate the difference image of the overlapping image The sum of the grayscale values ​​of all pixels in the depth image is calculated until the sum of the grayscale values ​​is less than or equal to a preset threshold; when the sum of the grayscale values ​​is less than or equal to the preset threshold, it is determined that the wheel cover to be detected has passed the rough inspection; after the depth image is rotated 360°, when the sum of the grayscale values ​​cannot be less than or equal to the preset threshold, it is determined that the wheel cover to be detected has not passed the rough inspection.

[0028] In one embodiment of the present application, extracting the size features and shape features of the outer ring area includes:

[0029] Extracting a front two-dimensional contour of the three-dimensional model;

[0030] Performing ellipse fitting on the front two-dimensional contour to obtain an ellipse contour;

[0031] Extract multiple inner diameter lines evenly distributed around the ellipse contour , wherein both ends of the inner diameter line intersect with the elliptical contour, and the inner diameter line passes through the center of the wheel cover, wherein, is a positive integer;

[0032] Calculate multiple evenly spaced inner diameter lines The average and variance , and the average value as the size feature of the outer ring area, and the variance as a shape feature.

[0033] In one embodiment of the present application, extracting the depth information of the area to be detected includes:

[0034] Rotating the three-dimensional model to a standard posture based on the wheel cover center and the positioning feature so that the posture of the detection area is consistent with the preset posture, wherein the positioning feature is a valve notch;

[0035] The detection area at the preset position is divided into multiple sector areas based on multiple annular dividing lines and inner diameter lines. ;

[0036] For each sector area Sampling the depth value to obtain multiple sampling points Depth value ;

[0037] Calculate the area of ​​each sector Multiple sampling points The average depth value and depth value variance are used to obtain the depth information.

[0038] In one embodiment of the present application, surface defect detection is performed on the wheel cover to be detected based on the pre-constructed depth reference information and the depth information of the area to be detected to obtain a surface detection result, including:

[0039] Based on sector area The position of the fan-shaped area is determined from the depth reference information A depth reference range and a variance reference range of the depth reference information, wherein the depth reference information includes a depth reference range and a variance reference range of a fan-shaped area at a plurality of positions;

[0040] Each sector area The average depth value and the depth value variance are compared with the corresponding depth reference range and variance reference range respectively;

[0041] In the sector area When the average depth value and the depth value variance fall into the corresponding depth reference range and variance reference range respectively, through the fan-shaped area Surface detection; in the sector area When the average depth value does not fall into the corresponding depth reference range, the fan-shaped area is determined There are abnormal depth defects in the fan-shaped area When the depth value variance does not fall into the corresponding variance reference range, the fan-shaped area is determined There are abnormal flatness defects.

[0042] In one embodiment of the present application, the cross-sectional profile feature is compared with a pre-constructed standard buckle cross-sectional profile for similarity to obtain a buckle detection result, including:

[0043] Based on the preset matching coordinates, the cross-sectional profile feature Pre-built standard snap-fit ​​profiles Perform registration;

[0044] Calculate the overlap between the profile feature after registration and the profile of the standard buckle , wherein the overlap degree The mathematical expression is:

[0045]

[0046] In the formula, Standard buckle profile The area of is the overlap area between the cross-sectional profile feature after registration and the cross-sectional profile of the standard buckle;

[0047] In the overlap When the overlap degree is greater than the preset threshold, the buckle detection is passed; otherwise, it is determined that the buckle has abnormal defects.

[0048] In one embodiment of the present application, the roundness of the wheel cover to be inspected is detected based on the shape feature of the outer ring area to obtain a shape detection result, and the outer ring size of the wheel cover to be inspected is detected based on the pre-constructed outer ring standard size and the size of the outer ring area to obtain a size detection result, including:

[0049] The variance The variance is compared with the preset threshold and the When the value is less than or equal to the preset variance threshold, the roundness of the wheel cover to be tested is determined to be normal; otherwise, the roundness of the wheel cover to be tested is determined to be abnormal;

[0050] The average value The average value is compared with the preset inner diameter reference range and the When the inner diameter falls within the preset reference range, the size of the wheel cover to be tested is determined to be normal; otherwise, the size of the wheel cover to be tested is determined to be abnormal.

[0051] The present application also provides a wheel cover detection system based on three-dimensional modeling, comprising:

[0052] A modeling module is used to perform blue light scanning at multiple angles on the wheel cover to be inspected to obtain a three-dimensional model of the wheel cover;

[0053] A center extraction module, used for extracting features from the three-dimensional model to obtain the center of the wheel cover;

[0054] A first detection module, configured to perform a rough inspection on the wheel cover to be inspected based on the center of the wheel cover and a pre-built rough inspection template to obtain a first inspection result;

[0055] A region division module, configured to align a pre-constructed mask with the front and back sides of the three-dimensional model based on the center of the wheel cover when the first detection result is normal, and extract a plurality of regions to be detected of the three-dimensional model based on the aligned masks, wherein the regions to be detected include an outer ring region, an inner ring region, a banner region, and a buckle region;

[0056] A feature extraction module, used to extract the size features and shape features of the outer ring area, extract the depth information of the area to be detected, and extract the cross-sectional profile features of the buckle area along a preset cross-sectional line;

[0057] The second detection module is used to perform shape detection, size detection, surface defect detection and buckle detection on the wheel cover to be inspected based on the features of a pre-built wheel cover standard model, the size features of the outer ring area, the roundness features of the outer ring area, the depth information of the area to be inspected, and the cross-sectional contour features to obtain a second detection result.

[0058] The beneficial effects of the present invention are as follows: a wheel cover detection method and system based on three-dimensional modeling of the present invention constructs a three-dimensional model of the wheel cover by means of blue light scanning. Based on the shape characteristics of the wheel cover, the present application first extracts the center of the wheel cover, and then performs a rough inspection based on the center of the wheel cover and the rough inspection template, so as to verify the depth information of the wheel cover as a whole. After passing the rough inspection, the fine inspection process is entered. In the fine inspection process, the wheel cover is divided into multiple inspection areas on the horizontal plane, and surface defect inspection is performed on each area separately. In addition, the shape and size of the wheel cover are also detected based on the outer ring area. For the buckle structure with a complex structure, the present application performs inspection by extracting the corresponding cross-sectional structure. The present application can realize a fully automatic inspection process for the wheel cover structure, which has higher inspection efficiency and inspection accuracy compared to manual inspection. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] The present invention will be further described below in conjunction with the accompanying drawings and embodiments:

[0060] Figure 1 is a schematic diagram of a three-dimensional model structure of a wheel cover shown in an embodiment of the present application;

[0061] Figure 2 is a schematic diagram of the principle of blue light scanning shown in an embodiment of the present application;

[0062] Figure 3 is a flow chart of a wheel cover detection method based on three-dimensional modeling shown in an embodiment of the present application;

[0063] Figure 4 This is a schematic diagram of the rough inspection and adjustment process in this application;

[0064] Figure 5It is a schematic diagram of each area of ​​the wheel cover in this application;

[0065] Figure 6 is a schematic diagram of a sector-shaped area in an embodiment of the present application;

[0066] Figure 7 A schematic diagram comparing the buckle profile features in this application with the cross-sectional profile of a standard buckle;

[0067] Figure 8 It is a structural diagram of a wheel cover detection system based on three-dimensional modeling shown in one embodiment of the present application. DETAILED DESCRIPTION

[0068] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0069] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show the layers related to the present invention rather than being drawn according to the number, shape and size of the layers in actual implementation. In actual implementation, the type, quantity and proportion of each layer may be changed arbitrarily, and the layer layout may also be more complicated.

[0070] In the following description, numerous details are discussed to provide a more thorough explanation of embodiments of the present invention; however, it is apparent to one skilled in the art that embodiments of the present invention may be practiced without these specific details.

[0071] Figure 1 is a schematic diagram of a three-dimensional model structure of a wheel cover shown in an embodiment of the present application. Figure 1 The wheel cover shown in the figure is automatically inspected based on machine vision. Figure 1 As shown, the wheel cover structure is generally centrally symmetrical, including a front side and a back side. There are multiple snap-fit ​​structures on the back side, and the wheel cover is fixed to the wheel hub through the snap-fit ​​structures.

[0072] Figure 2 is a schematic diagram of the principle of blue light scanning shown in an embodiment of the present application, such as Figure 2 As shown, the present application utilizes a blue light-based 3D scanner to perform three-dimensional modeling. A 3D scanner is a set of equipment that can obtain the three-dimensional morphology of an object's surface.

[0073] 3D scanners can scan and measure the three-dimensional shape of an object while moving, so they have the characteristics of real-time data acquisition, dynamic data acquisition, convenient and flexible handheld operation, etc. Generally, they include two (or more) industrial cameras (including lenses) and a laser transmitter as their core components. The industrial cameras and lasers are fixed together by metal parts. Other accessories include data cables, power cables, indicator lights, light sources, control circuit boards, etc.

[0074] Tracker By tracking a polyhedral target on a spherical scanner, Track is able to determine the position of the scanner (in space).

[0075] The scanner laser emits multiple laser lines to the surface of the object, and then the binocular camera obtains the image data of the laser lines on the surface of the object. After image processing and binocular stereo vision solution, the three-dimensional point cloud data of the position of the laser lines on the surface of the object can be obtained.

[0076] When the scanner laser line continuously passes over the surface of the object, all the laser point clouds form a three-dimensional construction of the object's surface.

[0077] The distance between the scanner camera lens and the focus point is the scanner reference distance. When working, it is more appropriate to keep the distance between the scanner and the object at the reference distance. The depth of field is the range of clear imaging of the camera. The depth of field of the camera ensures that the scanner can perform 3D scanning at the depth of field distance.

[0078] Finally, the scanned laser point cloud is reconstructed in three dimensions through relevant tools to obtain a three-dimensional model of the wheel cover.

[0079] Figure 3 is a flow chart of a wheel cover detection method based on three-dimensional modeling shown in an embodiment of the present application, such as Figure 3 As shown, a wheel cover detection method based on three-dimensional modeling in this embodiment may include the following steps:

[0080] S310, performing blue light scanning at multiple angles on the wheel cover to be inspected to obtain a three-dimensional model of the wheel cover;

[0081] Blue light scanning modeling has the advantages of high precision and low error. It is suitable for application scenarios that require very fine details to be captured. Since there are some buckles and edge mechanisms in the wheel cover that require high-precision detection, this application chooses to scan the structural details of the wheel cover with blue light and build the corresponding three-dimensional model.

[0082] In this application, a handheld blue light scanner can be used to scan the wheel cover. The scanner and the wheel cover can also be fixed in corresponding fixtures, and then the scanner can be adjusted by setting a fixed scanning position and scanning angle, and a three-dimensional model can be constructed after multiple scans. Multiple multi-angle scans are conducive to collecting each structural detail of the wheel cover, providing more accurate data details in subsequent inspections.

[0083] S320, extracting features from the three-dimensional model to obtain a wheel cover center;

[0084] Since the wheel cover is generally circular in shape, the most critical position is the center of the wheel cover. The wheel cover in this application has thickness information, so the extracted center is actually the central axis. The specific extraction process is as follows:

[0085] S321, extracting a front two-dimensional contour and a back two-dimensional contour of the three-dimensional model;

[0086] The present application first extracts the front two-dimensional contour and the back two-dimensional contour of the three-dimensional model respectively. When extracting the contour, the front image and the back image of the three-dimensional model need to be gray-scale converted respectively, and then high-pass filtered to retain high-frequency information for easy contour extraction. After filtering, binarization is performed and the contour is extracted.

[0087] The contour at this time includes various contour lines, such as the contour of the hollow hole, the outer contour of the wheel cover, the buckle contour, the threaded hole contour, etc., so further screening is required.

[0088] S322, filtering out the outer contour in the front two-dimensional contour and the outer contour in the back two-dimensional contour based on the contour size configured in advance;

[0089] Since the size of the outer contour of the wheel cover is much larger than the size of other contours, the present application directly uses a size screening method to screen out the outer contour in the front two-dimensional contour and the outer contour in the back two-dimensional contour.

[0090] S323, extracting the first outer contour center point in the front two-dimensional contour and the second outer contour center point in the back two-dimensional contour, and constructing a central axis based on the first outer contour center point and the second outer contour center point, and taking the central axis as the wheel cover center.

[0091] When extracting the center, you need to first perform ellipse fitting on the two-dimensional contour to obtain the contour ellipse. Then extract the center from the fitted ellipse to obtain the center points of the front and back sides. Connect the center points of the front and back sides with a straight line to obtain the central axis as the center of the three-dimensional model.

[0092] S330, performing a rough inspection on the wheel cover to be inspected based on the wheel cover center and a pre-constructed rough inspection template to obtain a first inspection result;

[0093] The rough inspection process in this application mainly focuses on the approximate depth of the front and back of the wheel cover, and whether the wheel hub has obvious defects such as burrs and bulges. This application verifies the wheel cover by pre-building a standard model. During the rough inspection process, the rough inspection template is constructed by extracting the contour and depth features of the standard model.

[0094] The specific process of rough inspection is as follows:

[0095] S331, extracting a front image and a back image of the three-dimensional model;

[0096] S332, performing grayscale conversion on the front image and the back image respectively to obtain a front grayscale image and a back grayscale image;

[0097] The purpose of grayscale conversion is to remove irrelevant information in the front image and the back image, so as to facilitate the subsequent replacement of pixel information in the grayscale image with depth information.

[0098] S333, obtaining depth values ​​of multiple points on the front and back of the three-dimensional model , and based on the depth values ​​of multiple points The grayscale values ​​of the corresponding positions of the front grayscale image and the back grayscale image Reassign the value to get the depth image , wherein the depth image The depth value is linearly proportional or inversely proportional to the grayscale value;

[0099] Among them, this application adopts pixel sampling, or uses the depth value of all pixels To replace the gray value of the corresponding position The depth image after information replacement still expresses depth information through grayscale. For example, the pixel with a larger grayscale value has a larger depth value. The mathematical expression of replacement is:

[0100]

[0101] In the formula, represents the normalization function, It is the front grayscale image or the back grayscale image after information replacement.

[0102] The grayscale image after information replacement contains a lot of depth information and the distribution information of depth information. Therefore, by comparing the depth image with the rough inspection template, some obvious defects can be detected, such as obvious bulges, burrs with depth information on the edge, etc.

[0103] S334, based on the wheel cover center, the depth image With the rough inspection template Overlap and calculate the overlapping depth image With rough inspection template The difference image , and calculate the difference image The sum of the gray values ​​of all pixels in ;

[0104] The sum of the grayscale values ​​of all pixels in the difference image can reflect the depth image. With rough inspection template The difference between And the difference image The mathematical expression for the sum of the grayscale values ​​of all pixels in is:

[0105]

[0106]

[0107] in, For the depth image The gray value of the pixel in To find the difference image The sum of the grayscale values ​​of all pixels in the image. From the above formula, we can see that this application takes the depth image With rough inspection template The absolute value of the difference between the grayscale values ​​(depth values) of each corresponding pixel point. Therefore, the difference in the grayscale values ​​(depth values) of the two will be accumulated in middle.

[0108] Since this application uses the wheel cover center to measure the depth image With the rough inspection template To overlap, multiple rotations are required to find the angle with the highest degree of overlap for the final difference judgment.

[0109] Figure 4 This is a schematic diagram of the rough inspection and adjustment process in this application, such as Figure 4 As shown, the present application adjusts the depth image through the following process angle.

[0110] S335, in the gray value sum When the depth image is greater than a preset threshold, the depth image is converted based on the wheel cover center. Rotate target angle , and return to calculate the difference image of the overlapping image The sum of the grayscale values ​​of all pixels in the image is calculated until the sum of the grayscale values ​​is less than or equal to a preset threshold;

[0111] When the sum of the grayscale values ​​is less than or equal to a preset threshold, it is determined that the wheel cover to be detected has passed the rough inspection; when the sum of the grayscale values ​​cannot be less than or equal to the preset threshold after the depth image is rotated 360°, it is determined that the wheel cover to be detected has not passed the rough inspection.

[0112] In this application, if the total gray value If it is greater than the preset threshold, the depth image is rotated clockwise or counterclockwise. Target Angle , and again calculate the difference image of the overlapping images The sum of the grayscale values ​​of all pixels in to update During this process, once If the depth information of the two wheels is less than or equal to the preset threshold, the wheel cover will stop rotating. This means that the difference in depth information between the two wheels is small and the wheel cover can pass the rough inspection. If the difference in depth information between the two wheels cannot be reduced even after a full 360° rotation, it means that there is a large defect in the wheel cover. At this time, the obvious defect detection information is output and sent to the inspector. Since it is an obvious defect, the inspector can quickly detect the defect location with the naked eye.

[0113] If it passes the rough inspection, it means that there are no obvious defects on the wheel cover, so it will enter the subsequent detailed inspection process.

[0114] S340, when the first detection result is normal, aligning a pre-constructed mask with the front and back sides of the three-dimensional model based on the center of the wheel cover, and extracting a plurality of to-be-detected areas of the three-dimensional model based on the aligned masks, wherein the to-be-detected areas include an outer ring area, an inner ring area, a banner area, and a buckle area;

[0115] Since detailed detection needs to be performed, the present application implements a partition detection strategy. Figure 5 It is a schematic diagram of each area of ​​the wheel cover in this application, such as Figure 5 As shown, based on the differences in the various regions of the wheel cover itself, the present application divides the front and back of the wheel cover into an outer ring region, an inner ring region, a banner region, and a buckle region. In this embodiment, a mask is used to quickly extract the above regions. During the extraction, the three-dimensional model is selected based on the positioning features of the wheel cover itself, such as the gap for accommodating the air nozzle, and after adjusting to the target angle, it is extracted based on the center covering mask.

[0116] S350, extracting size features and shape features of the outer ring area, extracting depth information of the area to be detected, and extracting cross-sectional contour features of the buckle area along a preset cross-sectional line;

[0117] After partitioning, this application extracts different features for detection based on its structural characteristics. For example, the outer ring area is the area that reflects the overall size and shape of the wheel cover, so this application extracts size features and shape features in the outer ring area. Due to the complexity of the structure of the buckle area, it is difficult to detect it through depth information, so a cross-sectional view is extracted. All areas except the buckle area can be detected for surface defects through depth information, so depth information is extracted.

[0118] Specifically, the process of extracting the size features and shape features of the outer ring area includes:

[0119] S3501, extracting the front two-dimensional contour of the three-dimensional model;

[0120] Since the hub structure on the front of the wheel cover is relatively simple and can reflect the appearance surface, the front two-dimensional contour can be extracted to extract size features and shape features.

[0121] S3502, performing ellipse fitting on the front two-dimensional contour to obtain an ellipse contour;

[0122] S3503, extracting multiple inner diameter lines evenly distributed around the ellipse contour , wherein both ends of the inner diameter line intersect with the elliptical contour, and the inner diameter line passes through the center of the wheel cover, wherein, is a positive integer;

[0123] S3504, calculate multiple uniformly distributed inner diameter lines The average and variance , and the average value as the size feature of the outer ring area, and the variance as a shape feature.

[0124] For multiple inner diameter wires Calculate the average value as the inner diameter of the wheel cover. Then The variance is calculated as the basis of its roundness. The smaller the variance, the higher the roundness.

[0125] Extracting the depth information of the area to be detected includes:

[0126] S3511, rotating the three-dimensional model to a standard posture based on the wheel cover center and the positioning feature, so that the posture of the detection area is consistent with the preset posture, wherein the positioning feature is a valve notch;

[0127] Depend on Figure 1 It can be seen that there is a valve notch in the wheel cover that does not satisfy the central symmetry. The valve notch is used as a positioning feature to rotate the three-dimensional model to the standard posture.

[0128] S3512, dividing the detection area in the preset posture into multiple sector areas based on multiple annular dividing lines and inner diameter lines ;

[0129] After adjusting the posture, each detection area can be divided into multiple sector areas using multiple circular dividing lines and inner diameter lines. The purpose of this is to keep the depth information of all points in each area consistent as much as possible, so as to facilitate subsequent verification using reference data.

[0130] Figure 6 is a schematic diagram of a sector-shaped area in an embodiment of the present application, and the sector-shaped areas are divided as follows Figure 6 shown.

[0131] S3513, for each sector area Sampling the depth value to obtain multiple sampling points Depth value ;

[0132] S3514, calculate each sector area Multiple sampling points The average depth value and depth value variance are used to obtain the depth information.

[0133] The depth information is composed of the average depth value and depth value variance of multiple sampling points. The average depth value reflects the overall level of the depth value of the fan-shaped area, and the depth value variance reflects the flatness of the fan-shaped area.

[0134] S360, based on the features of the pre-built wheel cover standard model, the size features of the outer ring area, the roundness features of the outer ring area, the depth information of the area to be inspected, and the cross-sectional contour features, the wheel cover to be inspected is subjected to shape detection, size detection, surface defect detection, and buckle detection to obtain a second detection result.

[0135] In this application, the aspects of fine detection based on the above features mainly include:

[0136] (1) performing roundness detection on the wheel cover to be detected based on the shape characteristics of the outer ring area to obtain a shape detection result;

[0137] Specifically: the variance The variance is compared with the preset variance threshold and the When the value is less than or equal to the preset variance threshold, the roundness of the wheel cover to be tested is determined to be normal; otherwise, the roundness of the wheel cover to be tested is determined to be abnormal;

[0138] If the variance of the lengths of multiple inner diameter lines is within the preset variance tolerance range, it means that the roundness of the wheel cover is normal.

[0139] (2) performing outer ring size detection on the wheel cover to be detected based on the pre-established outer ring standard size and the size of the outer ring area to obtain a size detection result;

[0140] Specifically: the average value The average value is compared with the preset inner diameter reference range and the When the inner diameter falls within the preset reference range, the size of the wheel cover to be tested is determined to be normal; otherwise, the size of the wheel cover to be tested is determined to be abnormal.

[0141] If the average value of the lengths of multiple inner diameter lines is within the preset length tolerance range, it means that the inner diameter size of the wheel cover is normal.

[0142] (3) performing surface defect detection on the wheel cover to be inspected based on the pre-constructed depth reference information and the depth information of the area to be inspected to obtain a surface inspection result;

[0143] Specifically: Based on the fan-shaped area The position of the fan-shaped area is determined from the depth reference information A depth reference range and a variance reference range of the depth reference information, wherein the depth reference information includes a depth reference range and a variance reference range of a fan-shaped area at a plurality of positions;

[0144] In this application, the front and back of the standard model are divided into multiple sector-shaped areas in advance in the same way, and the depth reference range and variance reference range of each sector-shaped area are constructed by depth acquisition to obtain depth reference information. Since the three-dimensional model is adjusted to the standard posture in the previous article, the position of the sector-shaped area can be directly read to determine the corresponding depth reference range and variance reference range.

[0145] Each sector area The average depth value and the depth value variance are compared with the corresponding depth reference range and variance reference range respectively;

[0146] In the sector area When the average depth value and the depth value variance fall into the corresponding depth reference range and variance reference range respectively, through the fan-shaped area Surface detection; in the sector area When the average depth value does not fall into the corresponding depth reference range, the fan-shaped area is determined There are abnormal depth defects in the fan-shaped area When the depth value variance does not fall into the corresponding variance reference range, the fan-shaped area is determined There are abnormal flatness defects.

[0147] If a defect is detected, a mark is generated in the corresponding sector area, and the abnormality type and abnormality mark are sent to relevant personnel together with the image to help them quickly locate the abnormal area.

[0148] (4) Compare the similarity between the cross-sectional profile feature and the pre-constructed standard buckle cross-sectional profile to obtain the buckle detection result.

[0149] Finally, for the more complex buckle area, the present application extracts the corresponding cross-sectional profile features in the previous text, and compares the cross-sectional profile features with the pre-built standard buckle cross-sectional profile for similarity to complete the buckle verification.

[0150] Figure 7 Schematic diagram comparing the buckle profile features in this application with the cross-sectional profile of a standard buckle. The buckle structure is as follows Figure 7 shown.

[0151] Based on the preset matching coordinates, the cross-sectional profile feature Pre-built standard snap-fit ​​profiles Perform registration;

[0152] Since the posture is adjusted in advance in this application, the cross-sectional profile feature can be fixed at a fixed position in the same coordinate system. Pre-built standard snap-fit ​​profiles Perform registration.

[0153] Calculate the overlap between the profile feature after registration and the profile of the standard buckle , wherein the overlap degree The mathematical expression is:

[0154]

[0155] In the formula, Standard buckle profile The area of is the overlap area between the cross-sectional profile feature after registration and the cross-sectional profile of the standard buckle;

[0156] In the overlap When the overlap degree is greater than the preset threshold, the buckle detection is passed; otherwise, it is determined that the buckle has abnormal defects.

[0157] After registration, the similarity between the two is calculated by calculating the overlap. If the overlap is greater than the preset overlap threshold, it means that the shapes of the two are basically the same, the buckle can achieve the corresponding function, and the test is passed. If the overlap is low, there may be problems with the position and shape, and the buckle may not be able to achieve the corresponding function. At this time, the position of the abnormal buckle is marked and sent to the relevant personnel.

[0158] Finally, all abnormal information is packaged and uploaded to the cloud for storage. If there is no abnormal information, it will enter the next wheel cover inspection process to achieve automatic inspection.

[0159] The present invention provides a wheel cover detection method based on three-dimensional modeling, which constructs a three-dimensional model of the wheel cover by means of blue light scanning. Based on the shape characteristics of the wheel cover, the present application first extracts the center of the wheel cover, and then performs a rough inspection based on the center of the wheel cover and the rough inspection template, so as to verify the depth information of the wheel cover as a whole. After passing the rough inspection, the fine inspection process is entered. In the fine inspection process, the wheel cover is divided into multiple inspection areas on the horizontal plane, and surface defect inspection is performed on each area separately. In addition, the shape and size of the wheel cover are also detected based on the outer ring area. For the buckle structure with a complex structure, the present application performs inspection by extracting the corresponding cross-sectional structure. The present application can realize a fully automatic inspection process for the wheel cover structure, which has higher inspection efficiency and inspection accuracy compared to manual inspection.

[0160] like Figure 8 As shown, the present application also provides a wheel cover detection system based on three-dimensional modeling, comprising:

[0161] A modeling module is used to perform blue light scanning at multiple angles on the wheel cover to be inspected to obtain a three-dimensional model of the wheel cover;

[0162] A center extraction module, used for extracting features from the three-dimensional model to obtain the center of the wheel cover;

[0163] A first detection module, configured to perform a rough inspection on the wheel cover to be inspected based on the center of the wheel cover and a pre-built rough inspection template to obtain a first inspection result;

[0164] A region division module, configured to align a pre-constructed mask with the front and back sides of the three-dimensional model based on the center of the wheel cover when the first detection result is normal, and extract a plurality of regions to be detected of the three-dimensional model based on the aligned masks, wherein the regions to be detected include an outer ring region, an inner ring region, a banner region, and a buckle region;

[0165] A feature extraction module, used to extract the size features and shape features of the outer ring area, extract the depth information of the area to be detected, and extract the cross-sectional profile features of the buckle area along a preset cross-sectional line;

[0166] The second detection module is used to perform shape detection, size detection, surface defect detection and buckle detection on the wheel cover to be inspected based on the features of a pre-built wheel cover standard model, the size features of the outer ring area, the roundness features of the outer ring area, the depth information of the area to be inspected, and the cross-sectional contour features to obtain a second detection result.

[0167] The present invention provides a wheel cover detection system based on three-dimensional modeling, which constructs a three-dimensional model of the wheel cover by means of blue light scanning. Based on the shape characteristics of the wheel cover, the present application first extracts the center of the wheel cover, and then performs a rough inspection based on the center of the wheel cover and the rough inspection template, so as to verify the depth information of the wheel cover as a whole. After passing the rough inspection, the fine inspection process is entered. In the fine inspection process, the wheel cover is divided into multiple inspection areas on the horizontal plane, and surface defect inspection is performed on each area separately. In addition, the shape and size of the wheel cover are also detected based on the outer ring area. For the buckle structure with a complex structure, the present application performs inspection by extracting the corresponding cross-sectional structure. The present application can realize a fully automatic inspection process for the wheel cover structure, which has higher inspection efficiency and inspection accuracy compared to manual inspection.

[0168] This embodiment also provides an electronic terminal, including: a processor and a memory;

[0169] The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the terminal executes any one of the methods in this embodiment.

[0170] The computer-readable storage medium in this embodiment can be understood by ordinary technicians in this field: all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to the computer program. The aforementioned computer program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk and other media that can store program codes.

[0171] The electronic terminal provided in this embodiment includes a processor, a memory, a transceiver and a communication interface. The memory and the communication interface are connected to the processor and the transceiver and complete communication with each other. The memory is used to store computer programs, the communication interface is used to communicate, and the processor and the transceiver are used to run computer programs so that the electronic terminal executes each step of the above method.

[0172] In this embodiment, the memory may include a random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0173] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0174] In the above-mentioned embodiments, although the present invention has been described in conjunction with the specific embodiments of the present invention, many replacements, modifications and variations of these embodiments will be apparent to those of ordinary skill in the art based on the foregoing description. The embodiments of the present invention are intended to cover all such replacements, modifications and variations falling within the broad scope of the appended claims.

[0175] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.

Claims

1. A wheel cover detection method based on three-dimensional modeling, characterized in that: Includes steps: Perform blue light scanning at multiple angles on the wheel cover to be inspected to obtain a three-dimensional model of the wheel cover; Extracting features from the three-dimensional model to obtain the wheel cover center includes: extracting a front two-dimensional contour and a back two-dimensional contour of the three-dimensional model; screening out an outer contour in the front two-dimensional contour and an outer contour in the back two-dimensional contour based on a pre-built and configured contour size; extracting a first outer contour center point in the front two-dimensional contour and a second outer contour center point in the back two-dimensional contour, and constructing a central axis based on the first outer contour center point and the second outer contour center point, and using the central axis as the wheel cover center; Based on the wheel cover center and the pre-built rough inspection template, the wheel cover to be inspected is roughly inspected to obtain a first inspection result, including: extracting the front image and the back image of the three-dimensional model; performing grayscale conversion on the front image and the back image respectively to obtain a front grayscale image and a back grayscale image; obtaining depth values ​​of multiple points on the front and back of the three-dimensional model , and based on the depth values ​​of multiple points The grayscale values ​​of the corresponding positions of the front grayscale image and the back grayscale image Reassign the value to get the depth image , wherein the depth image The depth value is linearly proportional or inversely proportional to the gray value; based on the center of the wheel cover, the depth image With the rough inspection template Overlap and calculate the overlapping depth image With rough inspection template The difference image , and calculate the difference image The sum of the grayscale values ​​of all pixels in the wheel cover; when the sum of the grayscale values ​​is greater than a preset threshold, the depth image is Rotate target angle , and return to calculate the difference image of the overlapping image The sum of the grayscale values ​​of all pixels in the depth image is calculated until the sum of the grayscale values ​​is less than or equal to a preset threshold; when the sum of the grayscale values ​​is less than or equal to the preset threshold, it is determined that the wheel cover to be detected has passed the rough inspection; after the depth image is rotated 360°, when the sum of the grayscale values ​​cannot be less than or equal to the preset threshold, it is determined that the wheel cover to be detected has not passed the rough inspection; When the first detection result is normal, aligning the pre-constructed mask with the front and back sides of the three-dimensional model based on the center of the wheel cover, and extracting multiple areas to be detected of the three-dimensional model based on the aligned masks, wherein the areas to be detected include an outer ring area, an inner ring area, a banner area, and a buckle area; Extracting the size features and shape features of the outer ring area, extracting the depth information of the area to be detected, and extracting the cross-sectional contour features of the buckle area along a preset cross-sectional line; Based on the features of a pre-constructed wheel cover standard model, the size features of the outer ring area, the roundness features of the outer ring area, the depth information of the area to be detected, and the cross-sectional contour features, the wheel cover to be detected is subjected to shape detection, size detection, surface defect detection, and buckle detection to obtain a second detection result, including: performing roundness detection on the wheel cover to be detected based on the shape features of the outer ring area to obtain a shape detection result; performing outer ring size detection on the wheel cover to be detected based on the pre-constructed outer ring standard size and the size of the outer ring area to obtain a size detection result; performing surface defect detection on the wheel cover to be detected based on the pre-constructed depth reference information and the depth information of the area to be detected to obtain a surface detection result; and performing a similarity comparison between the cross-sectional contour features and the pre-constructed standard buckle cross-sectional contour to obtain a buckle detection result.

2. The wheel cover detection method based on three-dimensional modeling according to claim 1, characterized in that: Extracting the size features and shape features of the outer ring area includes: Extracting a front two-dimensional contour of the three-dimensional model; Performing ellipse fitting on the front two-dimensional contour to obtain an ellipse contour; Extract multiple inner diameter lines evenly distributed around the ellipse contour , wherein both ends of the inner diameter line intersect with the elliptical contour, and the inner diameter line passes through the center of the wheel cover, wherein, is a positive integer; Calculate multiple evenly spaced inner diameter lines The average and variance , and the average value as the size feature of the outer ring area, and the variance as a shape feature.

3. The wheel cover detection method based on three-dimensional modeling according to claim 1, characterized in that: Extracting the depth information of the area to be detected includes: Rotating the three-dimensional model to a standard posture based on the wheel cover center and the positioning feature so that the posture of the detection area is consistent with the preset posture, wherein the positioning feature is a valve notch; The detection area at the preset position is divided into multiple sector areas based on multiple annular dividing lines and inner diameter lines. ; For each sector area Sampling the depth value to obtain multiple sampling points Depth value ; Calculate the area of ​​each sector Multiple sampling points The average depth value and depth value variance are used to obtain the depth information.

4. The wheel cover detection method based on three-dimensional modeling according to claim 3 is characterized in that: Performing surface defect detection on the wheel cover to be detected based on the pre-built depth reference information and the depth information of the area to be detected to obtain a surface detection result, including: Based on sector area The position of the fan-shaped area is determined from the depth reference information A depth reference range and a variance reference range of the depth reference information, wherein the depth reference information includes a depth reference range and a variance reference range of a fan-shaped area at a plurality of positions; Each sector area The average depth value and the depth value variance are compared with the corresponding depth reference range and variance reference range respectively; In the sector area When the average depth value and the depth value variance fall into the corresponding depth reference range and variance reference range respectively, through the fan-shaped area Surface detection; in the sector area When the average depth value does not fall into the corresponding depth reference range, the fan-shaped area is determined There are abnormal depth defects in the fan-shaped area When the depth value variance does not fall into the corresponding variance reference range, the fan-shaped area is determined There are abnormal flatness defects.

5. The wheel cover detection method based on three-dimensional modeling according to claim 1, characterized in that: The cross-sectional profile feature is compared with a pre-built standard buckle cross-sectional profile for similarity, and a buckle detection result is obtained, including: Based on the preset matching coordinates, the cross-sectional profile feature Pre-built standard snap-fit ​​profiles Perform registration; Calculate the overlap between the profile feature after registration and the profile of the standard buckle , wherein the overlap degree The mathematical expression is: In the formula, Standard buckle profile The area of is the overlap area between the cross-sectional profile feature after registration and the cross-sectional profile of the standard buckle; In the overlap When the overlap degree is greater than the preset threshold, the buckle detection is passed; otherwise, it is determined that the buckle has abnormal defects.

6. The wheel cover detection method based on three-dimensional modeling according to claim 2, characterized in that: The method includes: performing roundness detection on the wheel cover to be detected based on the shape feature of the outer ring area to obtain a shape detection result; and performing outer ring size detection on the wheel cover to be detected based on a pre-constructed outer ring standard size and a size of the outer ring area to obtain a size detection result, including: The variance The variance is compared with the preset threshold and the When the value is less than or equal to the preset variance threshold, the roundness of the wheel cover to be tested is determined to be normal; otherwise, the roundness of the wheel cover to be tested is determined to be abnormal; The average value The average value is compared with the preset inner diameter reference range and the When the inner diameter falls within the preset reference range, the size of the wheel cover to be tested is determined to be normal; otherwise, the size of the wheel cover to be tested is determined to be abnormal.

7. A wheel cover detection system based on three-dimensional modeling, using a wheel cover detection method based on three-dimensional modeling as claimed in any one of claims 1 to 6, characterized in that: include: A modeling module is used to perform blue light scanning at multiple angles on the wheel cover to be inspected to obtain a three-dimensional model of the wheel cover; The center extraction module is used to extract features from the three-dimensional model to obtain the wheel cover center, including: extracting the front two-dimensional contour and the back two-dimensional contour of the three-dimensional model; screening out the outer contour in the front two-dimensional contour and the outer contour in the back two-dimensional contour based on the contour size configured in advance; extracting the first outer contour center point in the front two-dimensional contour and the second outer contour center point in the back two-dimensional contour, and constructing a central axis based on the first outer contour center point and the second outer contour center point, and taking the central axis as the wheel cover center; The first detection module is used to perform a rough inspection on the wheel cover to be inspected based on the center of the wheel cover and a pre-built rough inspection template to obtain a first detection result, including: extracting the front image and the back image of the three-dimensional model; performing grayscale conversion on the front image and the back image respectively to obtain a front grayscale image and a back grayscale image; obtaining depth values ​​of multiple points on the front and back of the three-dimensional model , and based on the depth values ​​of multiple points The grayscale values ​​of the corresponding positions of the front grayscale image and the back grayscale image Reassign the value to get the depth image , wherein the depth image The depth value is linearly proportional or inversely proportional to the gray value; based on the center of the wheel cover, the depth image With the rough inspection template Overlap and calculate the overlapping depth image With rough inspection template The difference image , and calculate the difference image The sum of the grayscale values ​​of all pixels in the wheel cover; when the sum of the grayscale values ​​is greater than a preset threshold, the depth image is Rotate target angle , and return to calculate the difference image of the overlapping image The sum of the grayscale values ​​of all pixels in the depth image is calculated until the sum of the grayscale values ​​is less than or equal to a preset threshold; when the sum of the grayscale values ​​is less than or equal to the preset threshold, it is determined that the wheel cover to be detected has passed the rough inspection; after the depth image is rotated 360°, when the sum of the grayscale values ​​cannot be less than or equal to the preset threshold, it is determined that the wheel cover to be detected has not passed the rough inspection; A region division module, configured to align a pre-constructed mask with the front and back sides of the three-dimensional model based on the center of the wheel cover when the first detection result is normal, and extract a plurality of regions to be detected of the three-dimensional model based on the aligned masks, wherein the regions to be detected include an outer ring region, an inner ring region, a banner region, and a buckle region; A feature extraction module, used to extract the size features and shape features of the outer ring area, extract the depth information of the area to be detected, and extract the cross-sectional profile features of the buckle area along a preset cross-sectional line; The second detection module is used to perform shape detection, size detection, surface defect detection and buckle detection on the wheel cover to be detected based on the features of a pre-constructed wheel cover standard model, the size features of the outer ring area, the roundness features of the outer ring area, the depth information of the area to be detected, and the cross-sectional contour features to obtain a second detection result, including: performing roundness detection on the wheel cover to be detected based on the shape features of the outer ring area to obtain a shape detection result; performing outer ring size detection on the wheel cover to be detected based on the pre-constructed outer ring standard size and the size of the outer ring area to obtain a size detection result; performing surface defect detection on the wheel cover to be detected based on the pre-constructed depth reference information and the depth information of the area to be detected to obtain a surface detection result; and performing a similarity comparison between the cross-sectional contour features and the pre-constructed standard buckle cross-sectional contour to obtain a buckle detection result.

Images