CN115239625A - Method, device and equipment for detecting defects of top cover welding point cloud and storage medium - Google Patents
Method, device and equipment for detecting defects of top cover welding point cloud and storage medium Download PDFInfo
- Publication number
- CN115239625A CN115239625A CN202210703097.7A CN202210703097A CN115239625A CN 115239625 A CN115239625 A CN 115239625A CN 202210703097 A CN202210703097 A CN 202210703097A CN 115239625 A CN115239625 A CN 115239625A
- Authority
- CN
- China
- Prior art keywords
- point
- point cloud
- defect
- data
- top cover
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10028—Range image; Depth image; 3D point clouds
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Abstract
The invention relates to the technical field of machine vision, and provides a method, a device, equipment and a storage medium for detecting cloud defects of a top cover welding spot. Therefore, the method avoids the problems of insufficient manual visual detection precision and high labor cost, the point cloud of the surface to be detected of the top cover welding is obtained by using a machine visual method, the defect pits and the defect bumps are identified in the point cloud, the detection efficiency and the defect detection precision of the surface to be detected are improved, and the technical problems of low efficiency and high labor cost of manually detecting the defects of the top cover welding at present are solved.
Description
Technical Field
The invention relates to the technical field of machine vision, in particular to a method, a device and equipment for detecting cloud defects of welding spots of a top cover and a computer-readable storage medium.
Background
Along with the increasing market demand of new energy vehicles, the demand of power batteries as the heart of the new energy vehicles is naturally and rapidly increased, so that the research on the quality detection technology of the welding seam of the top cover of the battery core of the power batteries of the vehicles is very important, and the existing quality detection method completely depends on the traditional manual visual inspection mode. However, observe the welding seam through visual inspection personnel when detecting car power battery electricity core top cap welding seam quality, judge whether accord with the delivery standard through individual experience, such inspection mode makes the quality testing level vary from person to person, and is big to artifical dependency, and the quality testing level just can't be guaranteed to inefficiency to the cost of labor is high, and a welding equipment needs 2 unequally visual inspection personnel, because of the welding environment is abominable, visual inspection personnel's loss also can cause the continuation of cost to increase. Therefore, how to solve the problems of low accuracy and high cost caused by manual visual inspection of the quality of the welding line of the top cover of the power battery cell is a technical problem which needs to be solved at present.
Disclosure of Invention
The invention mainly aims to provide a method, a device and equipment for detecting cloud defects of welding spots of a top cover and a computer-readable storage medium, and aims to solve the technical problems of low quality efficiency and high cost of manual detection of welding spots of the top cover in the prior art.
In order to achieve the above object, the present invention provides a method for detecting cloud defects of a welding spot of a top cover, the method comprising: acquiring initial three-dimensional point cloud data of a surface to be detected based on a preset extraction method; acquiring initial three-dimensional point cloud data of a surface to be detected based on a preset extraction method; optimizing the initial three-dimensional point cloud data based on a preset optimization method to generate down-sampled point cloud data; processing each down-sampled point cloud data to obtain a point cloud normal vector and a point cloud curvature corresponding to each down-sampled point cloud data; extracting defect point cloud data to form a defect point cloud set based on a region growing algorithm, the point cloud normal vector and the point cloud curvature; and reporting the defect coordinates corresponding to the defect point cloud set.
Further, the optimizing the initial data based on a preset optimization method to generate down-sampled point cloud data includes: optimizing the initial three-dimensional point cloud data based on a voxel filtering algorithm to obtain point cloud gravity center data; and optimizing the point cloud gravity center data based on a statistical filtering algorithm to obtain the downsampled point cloud data.
Further, extracting defect point cloud data to form a defect point cloud set based on the region growing algorithm, the point cloud normal vector and the point cloud curvature, including: forming a plurality of downsampled point cloud data into a downsampled point cloud set, wherein each point in the downsampled point cloud set is a labeled point or an unlabeled point, the labeled point has a corresponding index value and is not a null value, and the index value of the unlabeled point is a null value; selecting any point in the down-sampling point cloud set as a point K, performing neighbor range search on the point K to select a random point n, and judging whether the point n is the label-free point; if the point n is the label-free point, performing normal vector discrimination and curvature discrimination on the point n based on the point cloud normal vector and the point cloud curvature; and determining whether the point n is defect point cloud data or not based on the normal vector judgment result and the curvature judgment result of the point n until all defect point cloud data in the down-sampling point cloud set are extracted and used as the defect point cloud set.
Further, if the point n is the label-free point, performing normal vector discrimination and curvature discrimination on the point n based on the point cloud normal vector and the point cloud curvature, including; judging whether the normal vector of the point n is smaller than a first preset threshold value or not; if the normal vector of the point n is smaller than the first preset threshold, judging whether the curvature of the point n is smaller than a second preset threshold; if the normal vector of the point n is not smaller than the first preset threshold value, performing neighbor range search on the point K to select other random points n, updating the point n through the other random points, and returning: and judging whether the point n is the label-free point.
Further, if the normal vector of the point n is smaller than a first preset threshold, determining whether the curvature of the point n is smaller than a second preset threshold, including; if the normal vector of the point n is smaller than a first preset threshold value, judging whether the curvature of the point n is smaller than a second preset threshold value; when the normal vector of the point n is smaller than a first preset threshold value, marking the point n, and giving an index value to the point n according to the size of the normal vector of the point n; judging whether the curvature of the point n is smaller than a second preset threshold value or not; if the curvature of the point n is smaller than the second preset threshold value, adding the point n into seed point cloud data, and dividing the points n with the same index value into the same type of seed point cloud data; sorting various kinds of seed point cloud data according to the size of an index value, and taking other kinds of seed point cloud data except the seed point cloud data with the maximum index value as defect point cloud data; and generating the defect point cloud set according to the defect point cloud data.
Further, the determining whether the curvature of the point n is smaller than a second preset threshold further includes:
and if the curvature of the point n is not smaller than a second preset threshold value, searching the neighboring range of the point K to select other random points, and updating the point n through the other random points until each point in the down-sampling point cloud set is completed.
Further, the acquiring of the initial three-dimensional point cloud data of the surface to be detected based on the preset extraction method includes: and acquiring the three-dimensional point cloud data of the surface to be detected as the initial three-dimensional point cloud data by an external camera, a line laser and a laser triangulation method, wherein the preset extraction method is the laser triangulation method.
In addition, in order to achieve the above object, the present invention further provides a method and an apparatus for detecting cloud defects of a welding spot of a top cover, wherein the apparatus for detecting cloud defects of a welding spot of a top cover comprises: the human face image conversion module is used for converting the target human face image into a cartoon human face image with a target style according to the human face conversion model; the body image conversion module is used for determining a body conversion model according to the target style and converting the target body image into a cartoon body image of the target style through the body conversion model; and the cartoon image generation module is used for splicing the cartoon face image and the cartoon body image according to the first image size of the cartoon face image and the second image size of the cartoon body image to generate a target character cartoon image.
In addition, in order to achieve the above object, the present invention further provides a method and an apparatus for detecting cloud defects of a top cover welding spot, where the apparatus for detecting cloud defects of a top cover welding spot includes a processor, a memory, and a top cover welding spot cloud defect detection program stored in the memory and executable by the processor, and when the top cover welding spot cloud defect detection program is executed by the processor, the method for detecting cloud defects of a top cover welding spot is implemented.
In addition, to achieve the above object, the present invention further provides a computer readable storage medium, in which a top cover welding point cloud defect detecting program is stored, wherein when the top cover welding point cloud defect detecting program is executed by a processor, the steps of the top cover welding point cloud defect detecting method are implemented.
The invention provides a method for detecting welding spot cloud defects of a top cover, which is characterized by acquiring initial three-dimensional point cloud data of a surface to be detected based on a preset extraction method; optimizing the initial data based on a preset optimization method to generate down-sampling point cloud data; processing the down-sampled point cloud data to obtain a point cloud normal vector and a point cloud curvature, and extracting a defect point cloud based on a region growing algorithm, the point cloud normal vector and the point cloud curvature; and reporting the defect position corresponding to the defect point cloud. Therefore, the method avoids the problems of insufficient precision, visual fatigue and high labor cost of manual visual inspection, the point cloud of the top cover welding to-be-detected area is obtained by using a machine visual method, defect pits and bumps are identified in the point cloud, the defect identification of the scanning area can be completed faster and better than the manual visual inspection, the detection efficiency and the precision of the defect detection of the to-be-detected area of the top cover welding are improved, and the technical problems of low efficiency and high labor cost of manual top cover welding defect detection at present are solved.
Drawings
Fig. 1 is a schematic diagram of a hardware structure of a top cover welding spot cloud defect detecting device according to an embodiment of the present invention;
FIG. 2 is a schematic flow chart illustrating a method for detecting defects in a welding point cloud of a top cover according to a first embodiment of the present invention;
FIG. 3 is a schematic flow chart of a method for detecting defects in a welding point cloud of a top cover according to a third embodiment of the present invention;
FIG. 4 is a schematic flowchart illustrating a method for detecting defects in a welding point cloud of a top cover according to a fourth embodiment of the present invention;
fig. 5 is a functional block diagram of a defect detection apparatus for a welding point cloud of a top cover according to a first embodiment of the present invention.
The implementation, functional features and advantages of the present invention will be further described with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The method for detecting the cloud defects of the welding spots of the top cover is mainly applied to equipment for detecting the cloud defects of the welding spots of the top cover, and the equipment for detecting the cloud defects of the welding spots of the top cover can be equipment with display and processing functions, such as a PC (personal computer), a portable computer, a mobile terminal and the like.
Referring to fig. 1, fig. 1 is a schematic diagram of a hardware structure of a top cover welding spot cloud defect detecting apparatus according to an embodiment of the present invention. In the embodiment of the present invention, the top cover solder joint cloud defect detecting apparatus may include a processor 1001 (e.g., a CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used for realizing connection communication among the components; the user interface 1003 may include a Display screen (Display), an input unit such as a Keyboard (Keyboard); the network interface 1004 may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface); the memory 1005 may be a high-speed RAM memory, or may be a non-volatile memory (e.g., a magnetic disk memory), and optionally, the memory 1005 may be a storage device independent of the processor 1001.
Those skilled in the art will appreciate that the hardware configuration shown in FIG. 1 does not constitute a limitation of the apparatus for detecting defects in a weld-top cloud, and may include more or fewer components than those shown, or some components in combination, or a different arrangement of components.
With continued reference to fig. 1, the memory 1005 of fig. 1, which is one type of computer-readable storage medium, may include an operating system, a network communication module, and a top lid solder cloud defect detection program.
In fig. 1, the network communication module is mainly used for connecting to a server and performing data communication with the server; the processor 1001 may call a top cover solder joint cloud defect detection program stored in the memory 1005, and execute the top cover solder joint cloud defect detection method provided by the embodiment of the present invention.
The embodiment of the invention provides a method for detecting cloud defects of welding spots of a top cover.
Referring to fig. 2, fig. 2 is a schematic flow chart of a method for detecting a defect of a welding point cloud of a top cover according to a first embodiment of the present invention.
In this embodiment, the method for detecting the defect of the welding point cloud of the top cover comprises the following steps:
s10, acquiring initial three-dimensional point cloud data of a to-be-detected surface based on a preset extraction method;
s20, optimizing the initial three-dimensional point cloud data based on a preset optimization method to generate down-sampled point cloud data;
step S30, processing each down-sampled point cloud data to obtain a point cloud normal vector and a point cloud curvature corresponding to each down-sampled point cloud data;
s40, extracting defect point cloud data to form a defect point cloud set based on a region growing algorithm, the point cloud normal vector and the point cloud curvature;
and S50, reporting the defect coordinates corresponding to the defect point cloud set.
In this embodiment, a three-dimensional point cloud of the welding surface of the top cap is obtained by using a machine vision method in the first step, a laser triangulation method can obtain more accurate point cloud data under a severe condition, a camera and a line laser are externally arranged, the line laser is applied to the surface of an object to be measured and moves a certain distance at a constant speed in a fixed direction, the camera acquires an image where the laser line is located, laser line information is extracted, and the point cloud of the surface of the part is obtained by using a triangulation method. And then, optimizing the initial three-dimensional point cloud data, wherein the processing amount of the point cloud data obtained by over-scanning is usually too large, and the detection speed is slow.
It can be understood that, through handling each down-sampling point cloud data afterwards, acquire the point cloud normal vector and the point cloud camber that each down-sampling point cloud data corresponds, based on predetermined defect design, handle and filter each down-sampling point cloud data's point through normal vector and point cloud camber as 2 judgement conditions, select the defect point at last, assemble into defect point cloud data.
The embodiment provides a method for detecting cloud defects of a welding spot of a top cover, which is based on a preset extraction method to obtain initial three-dimensional point cloud data of a surface to be detected; optimizing the initial data based on a preset optimization method to generate down-sampling point cloud data; processing the down-sampled point cloud data to obtain a point cloud normal vector and a point cloud curvature, and extracting a defect point cloud based on a region growing algorithm, the point cloud normal vector and the point cloud curvature; and reporting the defect position corresponding to the defect point cloud. Therefore, the method avoids the problems of insufficient precision, visual fatigue and high labor cost of manual visual inspection, the point cloud of the top cover welding to-be-detected area is obtained by using a machine visual method, defect pits and bumps are identified in the point cloud, the defect identification of the scanning area can be completed faster and better than the manual visual inspection, the detection efficiency and the precision of the defect detection of the to-be-detected area of the top cover welding are improved, and the technical problems of low efficiency and high labor cost of manual top cover welding defect detection at present are solved.
Based on the embodiment shown in fig. 2, in a second embodiment of the method for detecting a cloud defect of a welding spot on a top cover, the step S10 further includes:
and acquiring the three-dimensional point cloud data of the surface to be detected as the initial three-dimensional point cloud data by an external camera, a line laser and a laser triangulation method, wherein the preset extraction method is the laser triangulation method.
In this embodiment, a machine vision method is used to obtain three-dimensional point cloud of the top cover welding surface, a laser triangulation method can obtain more accurate point cloud data under the severe light environment conditions inside factories and workshops, line laser is applied to the surface of an object to be measured and moves a certain distance at a constant speed in a fixed direction through an external camera and line laser scanning, the camera collects an image of the laser line, laser line information is extracted, and three-dimensional point cloud data of the surface of the object part to be measured is obtained by a triangulation method and is used as initial three-dimensional point cloud data.
Referring to fig. 3, fig. 3 is a schematic flow chart of a method for detecting defects of a welding point cloud of a top cover according to a third embodiment of the present invention.
Based on the foregoing embodiment shown in fig. 2, in this embodiment, the step S20 specifically includes:
s21, optimizing the initial three-dimensional point cloud data based on a voxel filtering algorithm to obtain point cloud gravity center data;
and S22, optimizing the point cloud gravity center data based on a statistical filtering algorithm to obtain the downsampled point cloud data.
In this embodiment, the original point cloud data of the initial three-dimensional point cloud data is large, in order to increase the detection speed, the point cloud is down-sampled by using voxel grid filtering, a certain spatial scale is set, the point cloud is divided into spatial grids of a given scale, and the gravity center of the point cloud in each grid is used as the down-sampled point cloud. Let the dimension in the x-direction in space be x a The dimension in the y direction being y a The dimension in the z direction is z a Then, the x, y and z directions respectively use the three dimensions as step lengths to divide the space into a form of a tetragonal mesh, and assuming a certain tetragonal in the space, for the vertex of the tetrahedron, the point facing the ZOY plane and having a smaller value of x at the lower left corner is set as (x) i ，y j ，z k ) The point with the larger value of x at the upper right corner is (x) l ，y m ，z n ) If the point p = (x, y, z) exists in the original point cloud data, wherein x is i ≤x≤x l ，y l ≤y≤y j ，z k ≤z≤z n Then, the gravity center of the tetragonal mesh is the point cloud after down-sampling, and the gravity center is:
further, in this embodiment, some noise points may exist in the original point cloud of the initial three-dimensional point cloud data, outliers are filtered by using a statistical filtering method, a target point is given, distances from k neighborhood points to the point are calculated, distances of all the points should form a gaussian distribution, and points with mean values and variances outside a given threshold are removed.
Referring to fig. 4, fig. 4 is a schematic flow chart of a method for detecting defects in a welding point cloud of a top cover according to a fourth embodiment of the present invention.
Based on the foregoing embodiment shown in fig. 2, in this embodiment, the step S40 specifically includes:
step S41, forming a plurality of downsampled point cloud data into a downsampled point cloud set, wherein each point in the downsampled point cloud set is a labeled point or a non-labeled point, the labeled point has a corresponding index value and is not a null value, and the index value of the non-labeled point is a null value;
step S42, selecting any point in the down-sampling point cloud set as a point K, carrying out neighbor range search on the point K to select a random point n, and judging whether the point n is the label-free point;
step S43, if the point n is the label-free point, performing normal vector judgment and curvature judgment on the point n based on the point cloud normal vector and the point cloud curvature;
and S44, determining whether the point n is defect point cloud data or not based on the normal vector judgment result and the curvature judgment result of the point n until all defect point cloud data in the downsampling point cloud set are extracted to be used as the defect point cloud set.
In this embodiment, a normal vector of the point cloud is calculated, when the surface of the point cloud is smooth, information of the normal vector of a certain point in the point cloud is equivalent to calculating a plane fitting problem between the point and surrounding points, and the normal vector of the plane is the normal vector of the point, for the certain point in the space, a k-nearest point that is nearest to the certain point is first searched, and then a plane equation that fits the k-nearest point is calculated by using a least square method, where the plane can be represented as:
to find the normal vector, the eigenvalues and eigenvectors of matrix a are first solved.
In the above formula, λ i (j) ，
Respectively representing the eigenvalue and the eigenvector of A, and respectively setting the three eigenvalues as lambda 0 、λ 1 、λ 2 Corresponding feature vector is v 0 、v 1 、v 2 Let us assume λ 0 <λ 1 <λ 2 Then the minimum eigenvalue λ 0 Corresponding feature vector v 0 I.e. the normal vector found.
In this embodiment, curvature reflects the degree of irregularity of the surface of the point cloud, and the curvature estimation calculation of the point cloud is performed based on the estimation of the normal vector by using the above-mentioned PCA method for estimating the normal vector, because the matrix a is positive and semi-constant, all feature values are real feature values, and its feature vectors form a set of orthogonal coordinate systems, which represent the change of the point along the direction of the corresponding feature value, and quantitatively describe the change of k neighboring point sets of the point cloud along the normal direction, and the distance of the points from the tangent plane is estimated as follows:
sigma represents the surface variation of the point cloud P at k nearby points, if sigma =0, all points are on one plane, the maximum value of sigma is 1/3, which represents that the point set is completely isotropic distribution, the surface variation is not the intrinsic feature of one point sampling surface, but depends on the size of the neighborhood, and lambda 1 And λ 2 Representing the distribution of the point cloud on the tangent plane and thus can be used to estimate the local anisotropy, the curvature H of the point cloud in the point cloud set can be approximated as σ, i.e., H ≈ c.
In a specific embodiment, regarding how to extract point cloud defect data, the method of region growing is utilized to extract salient point and concave point defects from the point cloud. The specific idea of the region growing method is as follows: and establishing an index for each point in the preprocessed point cloud set for marking which set the point cloud belongs to, if the index has a value, then the point is marked, otherwise, the point is marked. And then, if the point cloud data has the non-labeled points, establishing an empty seed point set, and adding the point with the minimum curvature in the non-labeled points into the seed point set. And then, performing K neighbor search on the seed point, traversing all points in the K neighbor range, if the point has no label, calculating an included angle between a normal vector and the seed point, if the included angle is smaller than a set threshold value, marking the point as i (i =0,1,2, \8230;) and if the curvature of the point is smaller than a threshold value, adding the point into a seed point set, and performing the K neighbor search cycle on the newly added seed point until all the points in the seed point set are cycled and then finished. And after the round circulation, marking the same label i on part of points in the point cloud set, wherein the points with the same label are the same point cloud cluster, and repeating the step 2 until all the points are labeled. And finally, deleting the point cloud set with the maximum label by using the same point cloud set with the same label, wherein the rest point cloud set is the extracted defect point cloud.
Based on the foregoing embodiment shown in fig. 4, in this embodiment, the step S43 specifically includes:
step S431, judging whether the normal vector of the point n is smaller than a first preset threshold value;
step S432, if the normal vector of the point n is smaller than the first preset threshold, determining whether the curvature of the point n is smaller than the second preset threshold;
step S433, if the normal vector of the point n is not smaller than the first preset threshold, performing neighbor range search on the point K to select the remaining random points n, updating the point n through the remaining random points, and returning: and judging whether the point n is the label-free point.
In a specific embodiment, the first preset threshold is a point cloud normal vector correlation value, and the second preset threshold is a point cloud curvature correlation value. And calculating a normal vector of the point cloud, wherein under the condition that the surface of the point cloud is smooth, the normal vector information of a certain point in the point cloud is equivalent to calculating the plane fitting problem between the point and the surrounding points, and the normal vector of the plane is the normal vector of the point and is used for the certain point in the space. The curvature reflects the concave-convex degree of the point cloud surface, and the curvature estimation calculation of the point cloud is carried out on the basis of the normal vector estimation by utilizing the PCA method for estimating the normal vector, because the matrix A is positive and semi-definite, all characteristic values are real characteristic values, and the characteristic vectors form a group of orthogonal coordinate systems which represent the change of the point along the direction of the corresponding characteristic value. And when the selected point n meets the first preset threshold, judging a second preset threshold.
Based on the foregoing embodiment, in this embodiment, the step S432 specifically includes:
when the normal vector of the point n is smaller than a first preset threshold value, marking the point n, and giving an index value to the point n according to the size of the normal vector of the point n;
judging whether the curvature of the point n is smaller than a second preset threshold value or not;
if the curvature of the point n is smaller than the second preset threshold value, adding the point n into seed point cloud data, and dividing the points n with the same index value into the same type of seed point cloud data;
sorting various kinds of seed point cloud data according to the size of an index value, and taking other kinds of seed point cloud data except the seed point cloud data with the maximum index value as defect point cloud data;
and generating the defect point cloud set according to the defect point cloud data.
In a specific embodiment, a point K is subjected to neighbor search, all points in a K neighbor range are traversed, if a random point n near the point K has no label, an included angle between a normal vector and a seed point of the point K is calculated, if the included angle is smaller than a set threshold value, the point is labeled and is marked as i (i =0,1,2, \\8230;), if the curvature of the point is smaller than a threshold value, the point is added into a seed point set, and a K neighbor search cycle is performed on the newly added seed point until all points in the seed point set are cycled and then ended.
Based on the foregoing embodiment, in this embodiment, the determining whether the curvature of the point n is smaller than a second preset threshold further includes:
and if the curvature of the point n is not smaller than a second preset threshold value, searching the neighboring range of the point K to select other random points, and updating the point n through the other random points until each point in the down-sampling point cloud set is completed.
In a specific embodiment, regarding how to extract point cloud defect data, the method of region growing is utilized to extract salient point and concave point defects from the point cloud. The specific idea of the region growing method is as follows: and establishing an index for each point in the preprocessed point cloud set for marking which set the point cloud belongs to, if the index has a value, the point is marked, otherwise, the point is marked. And then, if the point cloud data has the non-labeled points, establishing an empty seed point set, and adding the point with the minimum curvature in the non-labeled points into the seed point set. And then, performing K neighbor search on the seed point, traversing all points in the K neighbor range, if the point has no label, calculating an included angle between a normal vector and the seed point, if the included angle is smaller than a set threshold value, marking the point as i (i =0,1,2, \8230;) and if the curvature of the point is smaller than a threshold value, adding the point into a seed point set, and performing the K neighbor search cycle on the newly added seed point until all the points in the seed point set are cycled and then finished. And after the round circulation, marking the same label i on part of points in the point cloud set, wherein the points with the same label are the same point cloud cluster, and repeating the step 2 until all the points are labeled. And finally, the points with the same labels are the same point cloud set, the point cloud set with the largest labels is deleted, and the remaining point cloud set is the extracted defect point cloud.
In addition, the embodiment of the invention also provides a method and a device for detecting the cloud defects of the welding spots of the top cover.
Referring to fig. 5, fig. 5 is a functional module schematic diagram of a first embodiment of a top cover welding point cloud defect detecting device of the present invention.
In this embodiment, the top cap welding point cloud defect detecting device includes:
the initial data acquisition module 10 is used for acquiring initial three-dimensional point cloud data of the surface to be detected based on a preset extraction method;
the data optimization module 20 is used for optimizing the initial three-dimensional point cloud data based on a preset optimization method to generate downsampled point cloud data;
the data processing module 30 is used for processing each down-sampled point cloud data to obtain a point cloud normal vector and a point cloud curvature corresponding to each down-sampled point cloud data;
the defect data extraction module 40 is used for extracting defect point cloud data to form a defect point cloud set based on a region growing algorithm, the point cloud normal vector and the point cloud curvature;
and a defect coordinate reporting module 50 for reporting the defect coordinates corresponding to the defect point cloud set.
Further, the initial data obtaining module 10 specifically includes:
and the initial data acquisition unit acquires the three-dimensional point cloud data of the surface to be detected as the initial three-dimensional point cloud data by an external camera, a line laser and a laser triangulation extraction method, wherein the preset extraction method is the laser triangulation extraction method.
Further, the data optimization module 20 specifically includes:
the point cloud gravity center extraction module is used for optimizing the initial three-dimensional point cloud data based on a voxel filtering algorithm to obtain point cloud gravity center data;
and the downsampling data acquisition module is used for optimizing the point cloud gravity center data based on a statistical filtering algorithm to acquire the downsampling point cloud data.
Further, the defect data extraction module 40 includes:
the tag identification unit is used for forming a plurality of downsampled point cloud data into a downsampled point cloud set, wherein each point in the downsampled point cloud set is a tagged point or a non-tagged point, the tagged point has a corresponding index value and is not a null value, and the index value of the non-tagged point is a null value;
a point n selection unit, which selects any point in the down-sampling point cloud set as a point K, performs neighbor range search on the point K to select a random point n, and judges whether the point n is the label-free point;
a point n discrimination unit configured to perform normal vector discrimination and curvature discrimination on the point n based on the point cloud normal vector and the point cloud curvature if the point n is the unlabeled point;
and the defect point cloud summarizing unit is used for determining whether the point n is defect point cloud data or not based on the normal vector judgment result and the curvature judgment result of the point n until all defect point cloud data in the downsampling point cloud set are extracted and used as the defect point cloud set.
Further, the defect data extraction module 40 includes:
a first threshold judgment unit, which judges whether the normal vector of the point n is smaller than a first preset threshold;
a second threshold judgment unit, configured to, if the normal vector of the point n is smaller than the first preset threshold, judge whether the curvature of the point n is smaller than a second preset threshold;
and a circular selection unit, if the normal vector of the point n is not smaller than the first preset threshold, performing neighbor range search on the point K to select other random points n, updating the point n through the other random points, and returning: and judging whether the point n is the label-free point.
Further, the defect data extraction module 40 includes:
an index value giving unit, configured to, when the normal vector of the point n is smaller than a first preset threshold, mark the point n, and give an index value to the point n according to the magnitude of the normal vector of the point n;
a progress second threshold unit, which is used for judging whether the curvature of the point n is smaller than a second preset threshold;
adding a seed point cloud unit, if the curvature of the point n is smaller than a second preset threshold value, adding the point n into seed point cloud data, and dividing the points n with the same index value into the same type of seed point cloud data;
the sorting unit sorts the various kinds of seed point cloud data according to the size of the index value, and takes other kinds of seed point cloud data except the seed point cloud data with the maximum index value as defect point cloud data;
and the defect point cloud generating unit is used for generating the defect point cloud set according to the defect point cloud data.
Further, the defect data extraction module 40 includes:
and a curvature judgment negation unit, wherein if the curvature of the point n is not less than a second preset threshold value, the point K is subjected to neighbor range search to select other random points, and the point n is updated through the other random points until each point in the downsampling point cloud set is completed.
Each module in the device for detecting defects of the welding point cloud of the top cover corresponds to each step in the embodiment of the method for detecting defects of the welding point cloud of the top cover, and the functions and the implementation process of the device are not repeated one by one here.
In addition, the embodiment of the invention also provides a computer readable storage medium.
The computer readable storage medium of the invention stores a top cover welding spot cloud defect detection program, wherein when the top cover welding spot cloud defect detection program is executed by a processor, the steps of the top cover welding spot cloud defect detection method are realized.
The method implemented when the top cover welding point cloud defect detection program is executed may refer to each embodiment of the top cover welding point cloud defect detection method of the present invention, and details are not described here.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or system. Without further limitation, an element defined by the phrases "comprising a," "8230," "8230," or "comprising" does not exclude the presence of other like elements in a process, method, article, or system comprising the element.
The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.
The application is operational with numerous general purpose or special purpose computing system environments or configurations. For example: personal computers, server computers, hand-held or portable devices, tablet-type devices, multiprocessor systems, microprocessor-based systems, set top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments that include any of the above systems or devices, and the like. The application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium (e.g., ROM/RAM, magnetic disk, optical disk) as described above and includes instructions for enabling a terminal device (e.g., a mobile phone, a computer, a server, an air conditioner, or a network device) to execute the method according to the embodiments of the present invention.
The above description is only a preferred embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes, which are made by using the contents of the present specification and the accompanying drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.
Claims (10)
1. A method for detecting cloud defects of welding spots of a top cover is characterized by comprising the following steps:
acquiring initial three-dimensional point cloud data of a to-be-detected surface based on a preset extraction method;
optimizing the initial three-dimensional point cloud data based on a preset optimization method to generate down-sampled point cloud data;
processing each down-sampled point cloud data to obtain a point cloud normal vector and a point cloud curvature corresponding to each down-sampled point cloud data;
extracting defect point cloud data to form a defect point cloud set based on a region growing algorithm, the point cloud normal vector and the point cloud curvature;
and reporting the defect coordinates corresponding to the defect point cloud set.
2. The method for detecting the cloud defects of the welding spots of the top cover according to claim 1, wherein the optimizing the initial data based on a preset optimization method to generate down-sampled point cloud data comprises:
optimizing the initial three-dimensional point cloud data based on a voxel filtering algorithm to obtain point cloud gravity center data;
and optimizing the point cloud gravity center data based on a statistical filtering algorithm to obtain the downsampled point cloud data.
3. The method for detecting the cloud defects of the welding spots of the top cover as claimed in claim 1, wherein the extracting of the data of the defect point cloud to form the collection of the defect point cloud based on the region growing algorithm, the normal vector of the point cloud and the curvature of the point cloud comprises:
forming a plurality of downsampled point cloud data into a downsampled point cloud set, wherein each point in the downsampled point cloud set is a labeled point or an unlabeled point, the labeled point has a corresponding index value and is not a null value, and the index value of the unlabeled point is a null value;
selecting any point in the down-sampling point cloud set as a point K, searching the neighbor range of the point K to select a random point n, and judging whether the point n is the label-free point;
if the point n is the label-free point, performing normal vector discrimination and curvature discrimination on the point n based on the point cloud normal vector and the point cloud curvature;
and determining whether the point n is defect point cloud data or not based on the normal vector discrimination result and the curvature discrimination result of the point n until all defect point cloud data in the downsampling point cloud set are extracted and used as the defect point cloud set.
4. The method for detecting the cloud defects of the welding spots of the top cover as claimed in claim 3, wherein if the point n is the label-free point, normal vector discrimination and curvature discrimination are performed on the point n based on the point cloud normal vector and the point cloud curvature, including;
judging whether the normal vector of the point n is smaller than a first preset threshold value or not;
if the normal vector of the point n is smaller than the first preset threshold, judging whether the curvature of the point n is smaller than a second preset threshold;
if the normal vector of the point n is not smaller than the first preset threshold value, performing neighbor range search on the point K to select other random points n, updating the point n through the other random points, and returning: and judging whether the point n is the label-free point.
5. The method for detecting the cloud defect of the welding spot on the top cover of claim 4, wherein if the normal vector of the point n is smaller than a first preset threshold, judging whether the curvature of the point n is smaller than a second preset threshold, including;
when the normal vector of the point n is smaller than a first preset threshold value, labeling the point n, and giving an index value to the point n according to the size of the normal vector of the point n;
judging whether the curvature of the point n is smaller than a second preset threshold value or not;
if the curvature of the point n is smaller than the second preset threshold value, adding the point n into seed point cloud data, and dividing the points n with the same index value into the same type of seed point cloud data;
sorting various kinds of seed point cloud data according to the size of an index value, and taking other kinds of seed point cloud data except the seed point cloud data with the maximum index value as defect point cloud data;
and generating the defect point cloud set according to the defect point cloud data.
6. The method for detecting the cloud defect of the welding spot on the top cover of claim 5, wherein the step of judging whether the curvature of the point n is smaller than a second preset threshold value further comprises the following steps:
and if the curvature of the point n is not smaller than a second preset threshold value, searching the neighbor range of the point K to select other random points, and updating the point n through the other random points until each point in the down-sampling point cloud set is completed.
7. The method for detecting the cloud defect of the welding spot on the top cover according to any one of claims 1 to 6, wherein the step of acquiring the initial three-dimensional point cloud data of the surface to be detected based on a preset extraction method comprises the following steps:
and acquiring three-dimensional point cloud data of the surface to be detected as the initial three-dimensional point cloud data by an external camera, a line laser and a laser triangulation method, wherein the preset extraction method is the laser triangulation method.
8. A method and a device for detecting cloud defects of welding spots of a top cover are characterized by comprising the following steps:
the initial data acquisition module is used for acquiring initial three-dimensional point cloud data of the surface to be detected based on a preset extraction method;
the data optimization module is used for optimizing the initial three-dimensional point cloud data based on a preset optimization method to generate down-sampling point cloud data;
the data processing module is used for processing each down-sampled point cloud data to obtain a point cloud normal vector and a point cloud curvature corresponding to each down-sampled point cloud data;
the defect data extraction module is used for extracting defect point cloud data to form a defect point cloud set based on a region growing algorithm, the point cloud normal vector and the point cloud curvature;
and the defect coordinate reporting module is used for reporting the defect coordinates corresponding to the defect point cloud set.
9. A top cover welding point cloud defect detection method device, wherein the top cover welding point cloud defect detection device comprises a processor, a memory, and a top cover welding point cloud defect detection program stored on the memory and executable by the processor, wherein when the top cover welding point cloud defect detection program is executed by the processor, the steps of the top cover welding point cloud defect detection method according to any one of claims 1 to 7 are implemented.
10. A computer-readable storage medium, wherein the computer-readable storage medium has a top cover solder joint cloud defect detection program stored thereon, and wherein the top cover solder joint cloud defect detection program, when executed by a processor, implements the steps of the top cover solder joint cloud defect detection method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210703097.7A CN115239625B (en) | 2022-06-21 | 2022-06-21 | Top cover welding spot cloud defect detection method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210703097.7A CN115239625B (en) | 2022-06-21 | 2022-06-21 | Top cover welding spot cloud defect detection method, device, equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115239625A true CN115239625A (en) | 2022-10-25 |
| CN115239625B CN115239625B (en) | 2023-05-09 |
Family
ID=83668637
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210703097.7A Active CN115239625B (en) | 2022-06-21 | 2022-06-21 | Top cover welding spot cloud defect detection method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115239625B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116309576A (en) * | 2023-05-19 | 2023-06-23 | 厦门微亚智能科技有限公司 | Lithium battery weld defect detection method, system and storage medium |
| CN116452583A (en) * | 2023-06-14 | 2023-07-18 | 南京信息工程大学 | Point cloud defect detection method, device and system and storage medium |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120256916A1 (en) * | 2009-12-11 | 2012-10-11 | Kazuo Kitamura | Point cloud data processing device, point cloud data processing method, and point cloud data processing program |
| CN105354846A (en) * | 2015-11-05 | 2016-02-24 | 沈阳东软医疗系统有限公司 | Method and apparatus for segmenting three-dimensional medical image |
| CN105631936A (en) * | 2014-10-31 | 2016-06-01 | 富泰华工业（深圳）有限公司 | Point cloud repair method and system |
| CN109118574A (en) * | 2018-07-04 | 2019-01-01 | 北京航空航天大学 | A kind of fast reverse modeling method extracted based on three-dimensional feature |
| CN110033447A (en) * | 2019-04-12 | 2019-07-19 | 东北大学 | A kind of high-speed rail heavy rail detection method of surface flaw based on cloud method |
| CN110517193A (en) * | 2019-06-28 | 2019-11-29 | 西安理工大学 | A kind of bottom mounted sonar Processing Method of Point-clouds |
| CN112098419A (en) * | 2020-09-11 | 2020-12-18 | 江苏理工学院 | System and method for detecting surface defects of automobile outer covering part |
| CN112489025A (en) * | 2020-12-07 | 2021-03-12 | 南京钢铁股份有限公司 | Method for identifying pit defects on surface of continuous casting billet |
| CN113469968A (en) * | 2021-06-29 | 2021-10-01 | 湖北中烟工业有限责任公司 | Cigarette packaging box seam height difference detection method and device and electronic equipment |
| CN114255326A (en) * | 2020-09-10 | 2022-03-29 | 广东博智林机器人有限公司 | Point cloud data processing method and device, electronic equipment and storage medium |
| CN114543666A (en) * | 2022-01-20 | 2022-05-27 | 大连理工大学 | Stockpile surface prediction method based on mine field environment perception |
-
2022
- 2022-06-21 CN CN202210703097.7A patent/CN115239625B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120256916A1 (en) * | 2009-12-11 | 2012-10-11 | Kazuo Kitamura | Point cloud data processing device, point cloud data processing method, and point cloud data processing program |
| CN105631936A (en) * | 2014-10-31 | 2016-06-01 | 富泰华工业（深圳）有限公司 | Point cloud repair method and system |
| CN105354846A (en) * | 2015-11-05 | 2016-02-24 | 沈阳东软医疗系统有限公司 | Method and apparatus for segmenting three-dimensional medical image |
| CN109118574A (en) * | 2018-07-04 | 2019-01-01 | 北京航空航天大学 | A kind of fast reverse modeling method extracted based on three-dimensional feature |
| CN110033447A (en) * | 2019-04-12 | 2019-07-19 | 东北大学 | A kind of high-speed rail heavy rail detection method of surface flaw based on cloud method |
| CN110517193A (en) * | 2019-06-28 | 2019-11-29 | 西安理工大学 | A kind of bottom mounted sonar Processing Method of Point-clouds |
| CN114255326A (en) * | 2020-09-10 | 2022-03-29 | 广东博智林机器人有限公司 | Point cloud data processing method and device, electronic equipment and storage medium |
| CN112098419A (en) * | 2020-09-11 | 2020-12-18 | 江苏理工学院 | System and method for detecting surface defects of automobile outer covering part |
| CN112489025A (en) * | 2020-12-07 | 2021-03-12 | 南京钢铁股份有限公司 | Method for identifying pit defects on surface of continuous casting billet |
| CN113469968A (en) * | 2021-06-29 | 2021-10-01 | 湖北中烟工业有限责任公司 | Cigarette packaging box seam height difference detection method and device and electronic equipment |
| CN114543666A (en) * | 2022-01-20 | 2022-05-27 | 大连理工大学 | Stockpile surface prediction method based on mine field environment perception |
Non-Patent Citations (1)
| Title |
|---|
| 魏永超: "基于曲面分割的三维点云物体识别" * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116309576A (en) * | 2023-05-19 | 2023-06-23 | 厦门微亚智能科技有限公司 | Lithium battery weld defect detection method, system and storage medium |
| CN116309576B (en) * | 2023-05-19 | 2023-09-08 | 厦门微亚智能科技股份有限公司 | Lithium battery weld defect detection method, system and storage medium |
| CN116452583A (en) * | 2023-06-14 | 2023-07-18 | 南京信息工程大学 | Point cloud defect detection method, device and system and storage medium |
| CN116452583B (en) * | 2023-06-14 | 2023-09-12 | 南京信息工程大学 | Point cloud defect detection method, device and system and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115239625B (en) | 2023-05-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115239625B (en) | Top cover welding spot cloud defect detection method, device, equipment and storage medium | |
| CN105243374B (en) | Three-dimensional face identification method, system and the data processing equipment using it | |
| US9064151B2 (en) | Device and method for detecting plantation rows | |
| CN109410192B (en) | Fabric defect detection method and device based on multi-texture grading fusion | |
| US20110229020A1 (en) | Learning method and apparatus for pattern recognition | |
| JP2016031764A (en) | Image analysis method | |
| CN105354866A (en) | Polygon contour similarity detection method | |
| CN111144325A (en) | Fault identification and positioning method, device and equipment for power equipment of transformer substation | |
| CN106780503A (en) | Remote sensing images optimum segmentation yardstick based on posterior probability information entropy determines method | |
| CN114881955B (en) | Annular point cloud defect extraction method, device and equipment storage medium based on slice | |
| Zelener et al. | Cnn-based object segmentation in urban lidar with missing points | |
| JP2019109839A (en) | Model generation device, generation method, and program | |
| CN112683169A (en) | Object size measuring method, device, equipment and storage medium | |
| CN113393442B (en) | Train part abnormality detection method, system, electronic equipment and storage medium | |
| CN113934877A (en) | Wheel hub model retrieval method, storage medium and equipment | |
| CN109389309A (en) | A kind of wind-driven generator performance estimating method, system, equipment and computer media | |
| CN116363319B (en) | Modeling method, modeling device, equipment and medium for building roof | |
| CN114494765B (en) | Method and device for identifying true and false smoke discrimination points, electronic equipment and storage medium | |
| Zhang et al. | A YOLOv3-Based Industrial Instrument Classification and Reading Recognition Method | |
| CN115546116A (en) | Method and system for extracting and calculating spacing of discontinuous surface of fully-covered rock mass | |
| CN113657179A (en) | Image recognition and modeling method and device, electronic equipment and storage medium | |
| CN103870846B (en) | A kind of image representing method and its application in images match, identification | |
| CN114037993A (en) | Substation pointer instrument reading method and device, storage medium and electronic equipment | |
| CN112598059A (en) | Worker dressing detection method and device, storage medium and electronic equipment | |
| CN113569600A (en) | Method and device for identifying weight of object, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |