CN115601366B - Vehicle bottom bolt looseness detection algorithm - Google Patents

Vehicle bottom bolt looseness detection algorithm Download PDF

Info

Publication number
CN115601366B
CN115601366B CN202211609718.1A CN202211609718A CN115601366B CN 115601366 B CN115601366 B CN 115601366B CN 202211609718 A CN202211609718 A CN 202211609718A CN 115601366 B CN115601366 B CN 115601366B
Authority
CN
China
Prior art keywords
bolt
image
area
point cloud
difference
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.)
Active
Application number
CN202211609718.1A
Other languages
Chinese (zh)
Other versions
CN115601366A (en
Inventor
邹爱刚
程坦
刘涛
马伟
胡冰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhongkehaituo Wuxi Technology Co ltd
Original Assignee
Zhongkehaituo Wuxi Technology Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Zhongkehaituo Wuxi Technology Co ltd filed Critical Zhongkehaituo Wuxi Technology Co ltd
Priority to CN202211609718.1A priority Critical patent/CN115601366B/en
Publication of CN115601366A publication Critical patent/CN115601366A/en
Application granted granted Critical
Publication of CN115601366B publication Critical patent/CN115601366B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G06T7/001—Industrial image inspection using an image reference approach
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
    • G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
    • G06T7/337—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving reference images or patches
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
    • G06T7/35—Determination of transform parameters for the alignment of images, i.e. image registration using statistical methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/60—Analysis of geometric attributes
    • G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
    • 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
    • G06T2207/20—Special algorithmic details
    • G06T2207/20076—Probabilistic image processing
    • 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
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30108—Industrial image inspection
    • G06T2207/30136—Metal
    • 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
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30108—Industrial image inspection
    • G06T2207/30164—Workpiece; Machine component
    • 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
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Physics & Mathematics (AREA)
  • Geometry (AREA)
  • Quality & Reliability (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Probability & Statistics with Applications (AREA)
  • Image Analysis (AREA)
  • Length Measuring Devices By Optical Means (AREA)

Abstract

The invention discloses a vehicle bottom bolt looseness detection algorithm, which belongs to the technical field of high-speed rail vehicle bottom fault detection methods and comprises the following steps: calculating the area of a difference image based on an image registration technology and a target region bolt looseness identification technology; step two: extracting a bolt target area, performing biplane fitting on the bolt target area by combining a point cloud growth division and numerical simulation technology of the area, and calculating the distance between two planes; step three: and (4) integrating the distance between the two planes through the area of the difference image, and evaluating the looseness degree of the bolt. According to the invention, the difference image area between the reference image and the image to be detected can be calculated by an image registration technology and a bolt target area identification technology, the distance between the top surface and the base surface of the bolt can be calculated by combining the point cloud growth segmentation of the bolt target area and a numerical simulation technology, the bolt loosening degree is calculated by combining the bolt target area and the point cloud growth segmentation and the numerical simulation technology, and the loosening degree of the bolt at the bottom of the high-speed rail car is more accurately judged.

Description

Vehicle bottom bolt looseness detection algorithm
Technical Field
The invention relates to a bolt looseness detection algorithm, in particular to a vehicle bottom bolt looseness detection algorithm, and belongs to the technical field of high-speed rail vehicle bottom fault detection methods.
Background
Bolts at the bottom of a high-speed rail are vital to normal operation of the high-speed rail, the normal fastening degree of each bolt at the bottom of the high-speed rail can be ensured to ensure the safety, comfort and correct point of each high-speed rail vehicle group, and two image registration methods are mentioned in the text of the technical research on the bolt looseness detection of the power grid iron tower based on machine vision, namely an image registration method based on an area and an image registration method based on characteristics: the image registration based on the region is also called as gray level-based registration, the similarity between the image to be measured and the reference image is directly measured according to the gray level value size and distribution of the whole image, and then an extreme point with higher similarity is searched, so that the transformation model parameters between the reference image and the floating image are determined; the image registration method based on the characteristics is used for matching each part of the image by extracting the significant characteristics of the image, has strong anti-interference performance, and has the advantages that the global information of the image is not concerned any more, and the calculated amount is much smaller than that of a region-based method.
Disclosure of Invention
The invention mainly aims to solve the problem that the existing power grid iron tower bolt looseness detection method cannot be applied to high-speed rail car bottom bolt looseness detection due to different working conditions of bolts, and provides a car bottom bolt looseness detection algorithm.
The purpose of the invention can be achieved by adopting the following technical scheme:
a vehicle bottom bolt looseness detection algorithm comprises the following steps:
the method comprises the following steps: calculating a difference image area based on an image registration technology and a 3D camera point cloud technology;
step two: extracting a bolt target area, performing biplane fitting on the bolt target area by combining an area point cloud growth division and a numerical simulation technology, and calculating the distance between two planes;
step three: the distance between the two planes is fused through the area of the difference image, and the looseness degree of the bolt is evaluated;
the first step comprises the following steps:
s1, selecting a reference image and an image to be detected: selecting an image without bolt looseness from historical images of points to be detected shot by a camera at the bottom of a high-speed rail as a reference image, and selecting a to-be-detected image;
s2, image registration: taking the central point of the camera as the origin of a space coordinate system, searching the space coordinate position relation, the attitude relation and the matrix relation between two images
Figure 168564DEST_PATH_IMAGE001
And &>
Figure 201243DEST_PATH_IMAGE002
Respectively marked as the positions of the centers of the reference image and the image to be measured in a spatial coordinate system, wherein &>
Figure 57203DEST_PATH_IMAGE003
Solving for G 1 G is 1 Is recorded as a position relation transformation solution between the reference image and the image to be measured, and the matrix->
Figure 394381DEST_PATH_IMAGE004
And &>
Figure 753818DEST_PATH_IMAGE005
Respectively recorded as the orientation of the reference image and the image to be measured in space, wherein->
Figure 19715DEST_PATH_IMAGE006
Solving for G 2 G is 2 Recording as the space orientation transformation relation between the reference image and the image to be measured, using G 1 And G 2 Converting the image to be detected to make the converted image to be detected consistent with the reference image in space, respectively taking eight alignment points from the two images, wherein the points corresponding to the same position in space in the eight points are in one-to-one correspondence;
s3, extracting a bolt area: extracting a bolt area through a target detection means;
s4, correcting image registration errors: correcting the bolt matching by using a matching method based on the bolt center point, recalculating a transformation matrix according to a matching result, and correcting an image registration error;
s5, calculating a difference image: calculating a difference image of a corresponding bolt area in the reference image and the registered image to be detected, and judging whether a target bolt in the image to be detected is loosened or not according to the area of the difference image;
the second step comprises the following steps:
roi region extraction: extracting a bolt area by a YOLO target detection technology;
B2. point cloud target area extraction: mapping the target detection ROI extracted by the B1 to a point cloud region;
B3. point cloud segmentation: extracting point clouds of the top surface and the base surface of the bolt by using a random sampling consistency method;
B4. two plane distances are calculated: fitting the top surface of the bolt with a plane of a base surface, and calculating the distance between the two planes;
in the third step, the looseness of the bolt is positively correlated with the difference length of the difference image and the distance between the two surfaces, the difference length of the difference image is set as a quotient of the area of the difference image and the peripheral boundary length of the difference image, and the difference length formula of the difference image is as follows:
Figure 97392DEST_PATH_IMAGE007
(ii) a In the formula: l is a radical of an alcohol c Difference length, S, noted as difference image c Recording the area of the difference image, and recording the D as the perimeter of the peripheral contour line of the difference image; the bolt looseness degree evaluation formula is as follows: />
Figure 739726DEST_PATH_IMAGE008
(ii) a In the formula: />
Figure 219249DEST_PATH_IMAGE009
Is recorded as the bolt loosening degree and is used for judging whether the bolt is loosened or not>
Figure 656047DEST_PATH_IMAGE010
Is recorded as the distance between the top surface and the base surface of the bolt.
As a further scheme of the invention, when the S2 image is aligned, the central points of the images are aligned, then the straight line of the boundary is fitted, the rotation angle of the images is calculated while the boundaries of the bolts are regulated, for the images with uneven exposure, the depth information of the reference image and the depth information of the target to be detected are extracted, and the missing bolt boundary is extracted from the depth information.
As a further aspect of the present invention, said B2 and said B3 comprise: preprocessing the bolt point cloud, removing all-zero invalid point cloud, removing point cloud outliers by adopting a Gaussian filtering method, and downsampling point cloud elements.
As a further scheme of the invention, when the point clouds of the top surface and the base surface of the bolt are extracted in B3, the cluster with the largest clustering is selected, clusters with normal vectors consistent with the top surface are utilized on the base surface, the normal vectors are calculated by using local neighbors of the centers of different point cloud clusters, the normal vectors of the base surface are estimated by using the point cloud normal vectors before the region is grown and separated, and the point cloud clusters with the normal vectors identical with the top surface are found.
As a further scheme of the present invention, when B4 calculates the distance between the two planes, the central point of the base plane is taken, and the distance from the central point of the base plane to the top plane is calculated.
The invention has the beneficial technical effects that: according to the vehicle bottom bolt looseness detection algorithm, the difference image area between the reference image and the image to be detected can be calculated through an image registration technology and a bolt target area identification technology, so that whether the high-speed rail bolt is loosened or not can be preliminarily judged through the difference image area between the reference image and the image to be detected, the distance between the top surface of the bolt and the base surface can be calculated through the combination of point cloud growth segmentation of a bolt target area and a numerical simulation technology, whether the high-speed rail vehicle bottom bolt is loosened or not can be further judged through the distance between the top surface of the bolt and the base surface, the bolt looseness degree can be calculated through the combination of the distance between the difference image area and the top surface of the bolt, and the looseness degree of the high-speed rail vehicle bottom bolt can be more accurately judged.
Drawings
Fig. 1 is a flow chart of a vehicle bottom bolt looseness detection algorithm according to the invention.
Detailed Description
In order to make the technical solutions of the present invention more clear and definite for those skilled in the art, the present invention is further described in detail below with reference to the examples and the accompanying drawings, but the embodiments of the present invention are not limited thereto.
As shown in fig. 1, the algorithm for detecting loosening of underbody bolts provided by this embodiment includes the following steps:
the method comprises the following steps: calculating a difference image area based on an image registration technology and a 3D camera point cloud technology;
step two: extracting a bolt target area, performing biplane fitting on the bolt target area by combining an area point cloud growth division and a numerical simulation technology, and calculating the distance between two planes;
step three: the distance between the two planes is fused through the area of the difference image, and the loosening degree of the bolt is evaluated;
the first step comprises the following steps:
s1, selecting a reference image and an image to be detected: selecting an image without bolt looseness from historical images of points to be detected shot by a camera at the bottom of a high-speed rail as a reference image, and selecting a to-be-detected image;
s2, image registration: using the camera center point of the camera as the origin of the space coordinate system, finding the space coordinate position relation and posture relation between the two images, and obtaining the matrix
Figure 17758DEST_PATH_IMAGE001
And &>
Figure 198203DEST_PATH_IMAGE002
Respectively recorded as the position of the center of the reference image and the image to be measured in the spatial coordinate system, wherein ^ is greater than or equal to ^>
Figure 30768DEST_PATH_IMAGE003
Solving for G 1 G is 1 Is recorded as a position relation transformation solution between the reference image and the image to be measured, and the matrix->
Figure 638467DEST_PATH_IMAGE004
And &>
Figure 690736DEST_PATH_IMAGE005
Respectively recorded as the orientation of the reference image and the image to be measured in space, wherein->
Figure 143714DEST_PATH_IMAGE006
Solving for G 2 G is 2 Recording as a spatial orientation transformation relation between the reference image and the image to be measured, using G 1 And G 2 Converting the image to be detected to make the converted image to be detected consistent with the reference image in space, respectively taking eight alignment points from the two images, wherein the points corresponding to the same position in space correspond one to one;
s3, extracting a bolt area: extracting a bolt area through a target detection means;
s4, correcting image registration errors: correcting the bolt matching by using a matching method based on the bolt center point, recalculating a transformation matrix according to a matching result, and correcting an image registration error;
s5, calculating a difference image: calculating a difference image of a corresponding bolt area in the reference image and the registered image to be detected, and judging whether a target bolt in the image to be detected is loosened or not according to the area of the difference image;
the second step comprises the following steps:
roi region extraction: extracting a bolt area by using a YOLO target detection technology;
B2. extracting a point cloud target area: mapping the target detection ROI extracted by the B1 to a point cloud region;
B3. point cloud segmentation: extracting point clouds of the top surface and the base surface of the bolt by using a random sampling consistency method;
B4. two plane distances are calculated: fitting the top surface of the bolt with a plane of a base surface, and calculating the distance between the two planes;
in the third step, the loosening degree of the bolt is positively correlated with the difference length of the difference image and the distance between the two surfaces, the difference length of the difference image is set as the quotient of the area of the difference image and the peripheral boundary length of the difference image, and the difference length formula of the difference image is as follows:
Figure 597830DEST_PATH_IMAGE007
(ii) a In the formula: l is a radical of an alcohol c Difference length, S, noted as difference image c Recording the area of the difference image, and recording the D as the perimeter of the peripheral contour line of the difference image; the bolt looseness evaluation formula is as follows:
Figure 376430DEST_PATH_IMAGE008
(ii) a In the formula: />
Figure 915995DEST_PATH_IMAGE009
Is recorded as the bolt loosening degree and is used for judging whether the bolt is loosened or not>
Figure 438244DEST_PATH_IMAGE010
Is recorded as the distance between the top surface and the base surface of the bolt.
The image registration technology and the bolt target area identification technology are used for calculating the difference image area between the reference image and the image to be detected, so that whether the high-speed rail bolt is loosened or not is preliminarily judged by using the difference image area between the reference image and the image to be detected, the distance between the top surface of the bolt and the base surface can be calculated by combining point cloud growth segmentation of the bolt target area and the numerical simulation technology, whether the high-speed rail vehicle bottom bolt is loosened or not is further judged by combining the distance between the top surface of the bolt and the base surface, the bolt loosening degree is calculated by combining the distance between the difference image area and the top surface of the bolt, and the loosening degree of the high-speed rail vehicle bottom bolt is more accurately judged.
And S2, when the images are aligned, aligning the central points of the images, fitting a straight line of the boundary, regulating the boundaries of the bolts, calculating the rotation angle of the images, extracting the depth information of the reference image and the depth information of the target to be detected for the image with uneven exposure, and extracting the missing bolt boundary from the depth information.
The missing bolt boundary is extracted through the depth information, the algorithm pre-estimation can be provided when the bolt image can regularly extract the bolt boundary, meanwhile, the calculation of the rotation degree between the image to be detected and the reference image can be realized while the bolt boundary is regulated, the image registration efficiency and speed are improved, and the quick investment of maintenance work is facilitated.
Said B2 and said B3 comprise: preprocessing the bolt point cloud, removing all-zero invalid point cloud, removing point cloud outliers by adopting a Gaussian filtering method, and downsampling point cloud elements.
Through the pretreatment of the bolt point cloud, all-zero invalid point clouds can be removed, the invalid point cloud is prevented from being put into the image extraction process, the efficiency and the accuracy of extraction of a bolt target area are improved, point cloud outliers in an image can be removed through the arrangement of a Gaussian filter method, point cloud noise points during image point cloud extraction are further reduced, adverse effects caused by extraction of the point cloud target area by the noise points are avoided, the top surface mentioned in B3 selects the cluster with the maximum clustering, clusters with the normal vectors consistent with the top surface are utilized on a base surface, the normal vectors of the base surface are estimated by using the point cloud normal vectors before the area growth and separation, and the point cloud clusters with the normal vectors identical with the top surface are found out.
The point cloud of the top surface of the bolt can be found out by selecting the maximum cluster of the top surface cluster, so that the outline of the top surface of the bolt can be extracted conveniently and quickly.
And B4, when the distance between the two planes is calculated, taking the central point of the base plane, and calculating the distance from the central point of the base plane to the plane of the top surface.
By selecting the center point of the plane of the base surface, the alignment and distance calculation with the center of the top surface of the bolt can be completed at the center of the base surface, and further the distance between the two planes can be conveniently calculated.
In summary, in this embodiment, according to the algorithm for detecting loosening of underbody bolts in this embodiment, the missing bolt boundaries are extracted through depth information, so that the algorithmic pre-estimation can be provided when the bolt boundaries can be extracted regularly from the bolt images, and meanwhile, the calculation of the rotation degree between the image to be detected and the reference image can be realized while the bolt boundaries are regulated, so that the efficiency and speed of image registration are improved, and the quick investment of maintenance work is facilitated. Through the preliminary treatment of bolt point cloud, can get rid of all zero invalid point clouds, avoid invalid point cloud to drop into the image extraction process, help improving the efficiency and the degree of accuracy that power target area drawed, through the setting of gauss filtering method, can get rid of the point cloud outlier in the image, and then reduce the point cloud noise point when the image point cloud drawed, avoid the noise point to draw the harmful effects that brings to the point cloud target area, through choosing the selection of the biggest cluster of top surface cluster, can find out the point cloud of bolt top surface, and then be convenient for draw bolt top surface profile fast, through utilizing the cluster that normal vector is unanimous with the top surface on the base surface, with the local neighbour of different point cloud cluster centers to ask the normal vector of top surface to draw the base surface, and then perfect the definite of base surface when drawing the normal vector of top surface. By selecting the center point of the plane of the base surface, the alignment and distance calculation with the center of the top surface of the bolt can be completed at the center of the base surface, and further the distance between the two planes can be conveniently calculated.
The above description is only for the purpose of illustrating the present invention and is not intended to limit the scope of the present invention, and any person skilled in the art can substitute or change the technical solution of the present invention and its conception within the scope of the present invention.

Claims (5)

1. A method for detecting looseness of bolts at bottoms of vehicles is characterized by comprising the following steps:
the method comprises the following steps: calculating a difference image area based on an image registration technology and a 3D camera point cloud technology;
step two: extracting a bolt target area, performing biplane fitting on the bolt target area by combining an area point cloud growth division and a numerical simulation technology, and calculating the distance between two planes;
step three: the distance between the two planes is fused through the area of the difference image, and the loosening degree of the bolt is evaluated; the first step comprises the following steps:
s1, selecting a reference image and an image to be detected: selecting an image without bolt looseness from historical images of points to be detected shot by a camera at the bottom of a high-speed rail as a reference image, and selecting a to-be-detected image;
s2, image registration: using the camera center point of the camera as the origin of the space coordinate system, finding the space coordinate position relation and posture relation between the two images, and obtaining the matrix w j ＝[x j ，y j ，z j ]And w i ＝[x i ，y i ，z i ]Respectively, as the positions of the centers of the reference image and the image to be measured in the space coordinate system, wherein w j ＝G 1 w i Solving for G 1 G is 1 Is recorded as a position relation transformation solution, matrix between the reference image and the image to be measured
Figure FDA0004065815540000011
And &>
Figure FDA0004065815540000012
Respectively as the orientation of the reference image and the image to be measured in space, where R j ＝G 2 R i Solving for G 2 G is to be 2 Recording as a spatial orientation transformation relation between the reference image and the image to be measured, using G 1 And G 2 Converting the image to be detected to make the converted image to be detected consistent with the reference image in space, respectively taking eight alignment points from the two images, wherein the points corresponding to the same position in space in the eight points are in one-to-one correspondence;
s3, extracting a bolt area: extracting a bolt area through a target detection means;
s4, correcting image registration errors: correcting and using a matching method based on the central point of the bolt to match the bolt, recalculating a transformation matrix according to a matching result, and correcting an image registration error;
s5, calculating a difference image: calculating a difference image of a corresponding bolt area in the reference image and the registered image to be detected, and judging whether a target bolt in the image to be detected is loosened or not according to the area of the difference image;
the second step comprises the following steps:
roi region extraction: extracting a bolt area by a YOLO target detection technology;
B2. point cloud target area extraction: mapping the target detection ROI extracted by the B1 to a point cloud area;
B3. point cloud segmentation: extracting point clouds of the top surface and the base surface of the bolt by using a random sampling consistency method;
B4. two plane distances are calculated: fitting the top surface of the bolt with a plane of a base surface, and calculating the distance between the two planes;
in the third step, the loosening degree of the bolt is positively correlated with the difference length of the difference image and the distance between the two surfaces, the difference length of the difference image is set as the quotient of the area of the difference image and the peripheral boundary length of the difference image, and the difference length formula of the difference image is as follows:
Figure FDA0004065815540000021
in the formula: l is a radical of an alcohol C Difference length, S, noted as difference image C Recording the area of the difference image, and recording the D as the perimeter of the peripheral contour line of the difference image;
the bolt looseness evaluation formula is as follows:
Figure FDA0004065815540000022
in the formula: r loose And recording the bolt loosening degree, and recording the distance between the top surface and the base surface of the bolt by deltaH.
2. The method for detecting loosening of bolts at bottom of car according to claim 1, characterized in that when the S2 image is aligned, the center points of the image are aligned, then the boundary straight line is fitted, the rotation angle of the image is calculated while the bolt boundary is normalized, for the image with uneven exposure, the depth information of the reference image and the depth information of the target to be detected are extracted, and the missing bolt boundary is extracted from the depth information.
3. The method for detecting loosening of bolts at bottoms of vehicles as claimed in claim 1, wherein when extracting the B2 midpoint cloud target area, preprocessing the bolt point cloud, removing all-zero invalid point cloud, removing point cloud outliers by adopting a Gaussian filtering method, and downsampling point cloud elements.
4. The vehicle bottom bolt looseness detection method of claim 1, wherein B3. Point cloud segmentation: and when the point clouds of the top surface and the base surface of the bolt are extracted in the step B3, selecting the cluster with the largest clustering, solving normal vectors by using clusters with normal vectors consistent with the top surface on the base surface, using local neighbors of different point cloud cluster centers to solve the normal vectors, estimating the normal vectors of the base surface by using the point cloud normal vectors before region growing and separating, and finding out the point cloud clusters with the normal vectors identical with the top surface.
5. The vehicle bottom bolt looseness detection method as claimed in claim 1, wherein when the distance between the two planes is calculated by B4, the distance between the plane center of the base plane and the plane of the top surface is calculated by taking the plane center of the base plane.
CN202211609718.1A 2022-12-15 2022-12-15 Vehicle bottom bolt looseness detection algorithm Active CN115601366B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211609718.1A CN115601366B (en) 2022-12-15 2022-12-15 Vehicle bottom bolt looseness detection algorithm

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211609718.1A CN115601366B (en) 2022-12-15 2022-12-15 Vehicle bottom bolt looseness detection algorithm

Publications (2)

Publication Number Publication Date
CN115601366A CN115601366A (en) 2023-01-13
CN115601366B true CN115601366B (en) 2023-04-14

Family

ID=84854149

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211609718.1A Active CN115601366B (en) 2022-12-15 2022-12-15 Vehicle bottom bolt looseness detection algorithm

Country Status (1)

Country Link
CN (1) CN115601366B (en)

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102520401B (en) * 2011-12-21 2013-05-08 南京大学 Building area extraction method based on LiDAR data
CN104835175B (en) * 2015-05-26 2019-11-05 西南科技大学 Object detection method in a kind of nuclear environment of view-based access control model attention mechanism
CN105512665B (en) * 2015-12-11 2018-10-30 中国测绘科学研究院 A kind of airborne laser radar point cloud data edge extracting method

Also Published As

Publication number Publication date
CN115601366A (en) 2023-01-13

Similar Documents

Publication Publication Date Title
CN109903327B (en) Target size measurement method of sparse point cloud
CN112405123B (en) Shoe sole roughing track planning method and device based on clustering algorithm
CN111311679B (en) Free floating target pose estimation method based on depth camera
CN108597009B (en) Method for detecting three-dimensional target based on direction angle information
CN107862735B (en) RGBD three-dimensional scene reconstruction method based on structural information
CN103727930A (en) Edge-matching-based relative pose calibration method of laser range finder and camera
CN111179321B (en) Point cloud registration method based on template matching
CN113884002B (en) Pantograph slide plate upper surface detection system and method based on two-dimensional and three-dimensional information fusion
CN114888692B (en) Polishing and grinding mechanical arm control system and method
CN113516695B (en) Point cloud registration strategy in laser profiler flatness measurement
CN111415376A (en) Automobile glass sub-pixel contour extraction method and automobile glass detection method
CN111311618A (en) Circular arc workpiece matching and positioning method based on high-precision geometric primitive extraction
CN112184725B (en) Method for extracting center of structured light bar of asphalt pavement image
CN107144234A (en) A kind of city rail vehicle wheel tread contour fitting method
CN116358449A (en) Aircraft rivet concave-convex amount measuring method based on binocular surface structured light
CN115601366B (en) Vehicle bottom bolt looseness detection algorithm
CN116958714B (en) Automatic identification method for wafer back damage defect
CN105260559A (en) Paper pulp fiber morphology parameter calculation method based on contour area and contour refinement
CN116883446B (en) Real-time monitoring system for grinding degree of vehicle-mounted camera lens
CN113723314A (en) Sugarcane stem node identification method based on YOLOv3 algorithm
CN112414318B (en) Steel rail abrasion rapid measurement algorithm based on structured light
CN112016354B (en) Method for detecting loading state of grain tank of grain transporting vehicle based on visual identification
CN113103226A (en) Visual guide robot system for ceramic biscuit processing and manufacturing
CN116958264A (en) Bolt hole positioning and pose estimation method based on three-dimensional vision
CN113639643B (en) Crop seedling stage height detection method based on RGB-D depth camera

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