CN115100109A - Method for detecting tightness state of rail elastic strip fastener - Google Patents
Method for detecting tightness state of rail elastic strip fastener Download PDFInfo
- Publication number
- CN115100109A CN115100109A CN202210545365.7A CN202210545365A CN115100109A CN 115100109 A CN115100109 A CN 115100109A CN 202210545365 A CN202210545365 A CN 202210545365A CN 115100109 A CN115100109 A CN 115100109A
- Authority
- CN
- China
- Prior art keywords
- elastic strip
- point
- point cloud
- fastener
- strip fastener
- 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
- 238000000034 method Methods 0.000 title claims abstract description 42
- 238000001514 detection method Methods 0.000 claims abstract description 16
- 238000000926 separation method Methods 0.000 claims abstract description 5
- 238000011156 evaluation Methods 0.000 claims abstract description 3
- 239000013598 vector Substances 0.000 claims description 18
- 230000000877 morphologic effect Effects 0.000 claims description 10
- 238000000605 extraction Methods 0.000 claims description 6
- 238000004891 communication Methods 0.000 claims description 3
- 230000003287 optical effect Effects 0.000 claims description 3
- 230000002093 peripheral effect Effects 0.000 claims description 3
- 238000010845 search algorithm Methods 0.000 claims description 3
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 claims description 2
- 230000002159 abnormal effect Effects 0.000 claims description 2
- 230000010339 dilation Effects 0.000 claims description 2
- 230000003628 erosive effect Effects 0.000 claims description 2
- 230000001678 irradiating effect Effects 0.000 claims description 2
- 230000007547 defect Effects 0.000 abstract description 4
- 238000012423 maintenance Methods 0.000 abstract description 3
- 229910000831 Steel Inorganic materials 0.000 description 10
- 239000010959 steel Substances 0.000 description 10
- 230000005489 elastic deformation Effects 0.000 description 6
- 238000007689 inspection Methods 0.000 description 5
- 238000003384 imaging method Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 238000006243 chemical reaction Methods 0.000 description 3
- 230000011218 segmentation Effects 0.000 description 3
- 230000000007 visual effect Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 210000000709 aorta Anatomy 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010224 classification analysis Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000001066 destructive effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
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
- 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
- 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
- G06V10/762—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using clustering, e.g. of similar faces in social networks
-
- 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
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Databases & Information Systems (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Multimedia (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a method for detecting the tightness state of a rail elastic strip fastener, and relates to the technical field of machine vision detection. It includes: acquiring point cloud data of the elastic strip fastener; converting point cloud data of the elastic strip fastener into a binary image; extracting a two-dimensional framework of the elastic strip fastener from the binary image; and selecting a Z value by taking a normal line of each point in the two-dimensional framework of the elastic strip fastener to determine the three-dimensional framework of the elastic strip fastener; finding out a plurality of characteristic points from the three-dimensional skeleton data, and calculating the seam separation height of the elastic strip fastener according to the lowest point of the central concave part of the elastic strip fastener, namely an evaluation index of the tightness state of the elastic strip fastener. Through detecting different railway spring strip fasteners, the gap height of each fastener is calculated by using a non-contact measuring method based on three-dimensional point cloud, so that the tightness state of the fasteners can be automatically and quickly detected, and the accuracy of fastener defect detection and the rail maintenance efficiency are improved.
Description
Technical Field
The invention relates to the technical field of machine vision detection, in particular to a method for detecting the tightness state of a rail elastic strip fastener.
Background
The railway is a national economy aorta, a key infrastructure and a major civil engineering, and is the backbone of a comprehensive transportation system. The railway infrastructure comprises rails, fasteners, sleepers, fishplates and the like, wherein the fasteners are used for connecting the steel rails and the sleepers, and the fasteners on the left side and the right side of the rails fix the rails on the sleepers to prevent the rails from shifting. With the development of railways towards high speed, high density and heavy load, the destructive power of trains to railway infrastructure is increased, such as rail cracks, wear, fastener functional failures and the like. The functional failure of the fastener is mainly represented by the loss of the fastener, the fracture of the retaining ring, the dislocation of the retaining ring, the loosening of the bolt and the like. The functional failure of the fastener can cause the deviation of the left and right steel rails and the vibration of the train to be strengthened, thereby influencing the riding comfort and even causing the derailment of the train under severe conditions. Therefore, how to quickly and accurately detect the state of the fastener in a limited skylight time is a practical problem to be solved urgently by the domestic and foreign rail maintenance departments.
The scale of the rail transit construction in China is gradually enlarged, the corresponding mileage is increased year by year, and more new lines need to be detected and more lines need to be maintained in the operation period. The rail inspection in China is carried out at night, and the provided skylight has short time, so that the inspection mode has the defects of low working efficiency, high labor intensity, multiple human factor influences, low inspection sampling rate and the like, and the requirement of the general inspection of the service state of the rail fastener in China cannot be met.
Currently, rail fastener buckling force detection techniques based on machine vision can be broadly classified into two-dimensional and three-dimensional visual imaging detection. In the former, a camera is used for obtaining a two-dimensional image of a track, and the tightness of a fastener is generally detected by using the deviation or the texture of the image; in the latter, a three-dimensional visual imaging is generally formed by using a structured light and a camera, a three-dimensional point cloud of a scene is collected, and the tightness state of a fastener is identified according to the depth information and the three-dimensional space structure of a measured object. Although many studies have been conducted and significant results have been achieved for defect detection using two-dimensional vision and three-dimensional vision imaging, applications in the field of fastener tightness detection are problematic:
(1) the two-dimensional image acquired by the two-dimensional visual imaging lacks depth information of a third dimension and is only suitable for the condition that the bullet strip is rotated or displaced to complete failure under the buckling pressure. It is difficult to detect whether the buckling force of the fastener is insufficient, resulting in missed detection and false detection.
(2) The line laser sensor is used for acquiring accurate three-dimensional point cloud data, the three-dimensional point cloud data can reflect bolt loosening information of the fastener, and the research on how to accurately calculate the seam separating height of the fastener from the three-dimensional point cloud is lacked.
Disclosure of Invention
In view of the above, it is necessary to provide a method for detecting the tightness of the rail elastic strip fastener.
The embodiment of the invention provides a method for detecting the tightness state of a rail elastic strip fastener, which comprises the following steps:
acquiring point cloud data of the rail elastic strip fastener;
dividing a spring strip fastener area in the track spring strip fastener point cloud data, and extracting the point cloud data of the spring strip fastener from the spring strip fastener area;
converting the point cloud data of the elastic strip fastener into a binary image, and extracting a two-dimensional skeleton of the elastic strip fastener from the binary image; selecting high position Z value corresponding points by taking a normal line of each point in the two-dimensional framework of the elastic strip fastener to determine the three-dimensional framework of the elastic strip fastener;
finding out a plurality of minimum characteristic points from the three-dimensional skeleton data, and calculating the seam separation height of the elastic strip fastener according to the normal direction of the lowest point of the central recess of the elastic strip fastener; and the height of the gap is an evaluation index of the tightness state of the elastic strip fastener.
Further, acquire track bullet strip fastener point cloud data, specifically include:
vertically installing a three-dimensional line scanning camera on a detection vehicle, and vertically irradiating line laser on a rail elastic strip fastener;
the three-dimensional line scanning camera emits a beam of laser to irradiate the surface of a measured object, reflected light forms light spots on the surface of the photosensitive element through the optical lens group, the positions of the light spots formed by the reflection of the surfaces with different heights are different, namely 3D contour height is formed, and X, Y and Z coordinates are constructed according to the position of the 3D contour height to form three-dimensional point cloud data.
Further, divide into bullet strip fastener region in the track bullet strip fastener point cloud data to extract the point cloud data of bullet strip fastener in the bullet strip fastener region, include:
dividing height direction data D _ z of the track elastic strip fastener point cloud data D (x, y, z), and excluding the track elastic strip fastener point cloud data D (x, y, z) corresponding to a point with a lower height D _ z to obtain point cloud data D' (x, y, z);
sorting the point cloud data D ' (x, y, z) according to the size of y vector data D ' _ y of the point cloud data D ' (x, y, z), solving a first derivative D (D ' _ y) of the D ' _ y, comparing the D (D ' _ y) with a confidence value fib, and segmenting a bullet strip fastener existence area corresponding to each D ' _ y value;
finding each point cloud interval in the point cloud data D (x, y, z) of the corresponding track elastic strip fastener according to the y value interval corresponding to each elastic strip fastener existing area, and recording the point cloud data of the ith elastic strip fastener existing area as Di (x, y, z);
splicing point cloud data in an upper point cloud file by using the point cloud data Di (x, y, z) of the existing area of the elastic strip fastener, which meets the formula (1), splicing point cloud data in a lower point cloud file by using the point cloud data Di (x, y, z) of the existing area of the elastic strip fastener, which meets the formula (2), and judging the remaining point cloud data Di (x, y, z) of the existing area of the elastic strip fastener by using the formula (3) to obtain the point cloud Di (x, y, z) of the existing area of the complete elastic strip fastener;
(MAX(Di_y)-MIN(Di_y))<y_fib；i＝1 (1)
(MAX(Di_y)-MIN(Di_y))<y_fib；i＝end (2)
Point(Di)>point_fib&&(MAX(Di_y)-MIN(Di_y))<y_fib；i∈[1,end] (3)
the method comprises the following steps of (1) obtaining point cloud data Di (x, y, z) of an ith elastic strip fastener existing region point cloud data Di (y, z); y _ fib is the width value of the default elastic strip fastener area in the y direction; point _ fib is the cloud amount of the minimum bearing thought elastic strip fastener point; MAX () is the maximum value; MIN () is minimum; point (di) is the point cloud amount of point cloud data of the ith elastic strip fastener existing region; end is the area where the last spring fastener in a point cloud file exists.
Further, the bullet strip fastener region to among the track bullet strip fastener point cloud data is cut apart to extract the point cloud data of bullet strip fastener in the bullet strip fastener region, still include:
classifying point cloud data Di (x, y, z) of the existing area of the complete elastic strip fastener by an Euclidean clustering method;
in each classified category, excluding the category of which the point cloud data volume is less than 1 per mill of the total point cloud number and the category of which the MEAN (Di _ z) is smaller; wherein MEAN (Di _ z) is an average value of z vectors in the ith elastic fastener existing region point cloud data Di (x, y, z);
sorting according to the z values of the rest Di (x, y, z), and solving a first derivative d (Di _ z);
and comparing the preset confidence value fid with the solved first-order Z value d (Di _ Z), and deleting the point cloud data by the dichotomy principle to obtain point cloud data DT (x, y, Z) of the elastic strip fastener.
Further, the classifying the point cloud data Di (x, y, z) of the complete elastic strip fastener existence region by the euclidean clustering method includes:
selecting one point of point cloud data Di (x, y, z) of the area where a single complete elastic strip fastener exists;
finding n points nearest to the point through a KD-Tree neighbor search algorithm, and clustering the points with the distance smaller than a set threshold value into a set; if the number of the elements in the set is not increased any more, the whole clustering process is ended; otherwise, another point in the set is selected for re-clustering until the number of elements in the set is not increased any more.
Further, the point cloud data of the elastic strip fastener is converted into a binary image, and the method comprises the following steps:
obtaining the length-width pixel proportion of the elastic strip fastener corresponding to the point cloud image according to the point cloud data DT (x, y, z) of the elastic strip fastener, and taking the length-width pixel proportion as a length-width pixel point of a binary image to be converted:
y point ＝ceil(x point *p) (5)
wherein, y point 、x point Pixel points set for converting the binary image; x is the number of point The Ax and Ay are respectively corresponding x and y column vectors in point cloud data of the elastic strip fastener as initial set values;
according to the fact that the divided binary images are similar to a grid structure, the point cloud images are divided into single grids, and the point cloud range corresponding to each grid, namely each pixel point of the binary images is as follows:
XI＝min(A x )：(x point +1)：max(A x ) (6)
YI＝min(A y )：(y point +1)：max(A y ) (7)
YI(j)＝＜A y ＜YI(j+1)i∈(1：x point ) (8)
XI(i)＝＜A x ＜XI(i+1)j∈(1：y point ) (9)
wherein, YI (j) is a y vector of a binary image corresponding to the pixel point j; xi (j) is the x vector of the binary image corresponding to pixel point j;
converting point cloud data DT (x, y, z) of the elastic strip fastener into a binary image G (x, y) according to the point cloud range;
performing morphological closed operation on the binary image by using a disc-shaped structural element SE, specifically as a formula (10), and obtaining a binary image G' (x, y) with continuous pixel points;
wherein,
is a morphological dilation treatment and Θ is a morphological erosion treatment.
Further, extract out the two-dimensional skeleton of bullet strip fastener from binary image, include:
extracting a target peripheral contour from the continuous binary image G' (x, y) of each pixel point;
corroding the boundary of a target image by using the contour, narrowing the object into a line, deleting redundant pixels, shrinking the object without holes into the line with minimum connectivity, and shrinking the object with holes into a communication ring between each hole and the outer boundary to obtain a two-dimensional framework DG (x, y);
extract out the two-dimensional skeleton of bullet strip fastener from binary image, specifically include:
judging the values of 8 neighboring points around the binary image G' (x, y) with continuous pixels;
in the first sub-iteration, the pixel point p is deleted if and only if the conditions G1, G2, and G3 are all satisfied;
in the second sub-iteration, the pixel point p is deleted if and only if the conditions G1, G2, and G3' are all satisfied;
in the skeleton extraction process, a first iteration is adopted, when the pixel points of the whole binary image are not transformed any more, a second iteration scheme is adopted, the process is repeated until the whole image is reduced into a line, namely, no pixel p is satisfied: (condition G1& condition G2& condition G3) or (condition G1& condition G2& condition G3'); wherein,
the condition G1 is that more than one directly connected point among the 8 neighboring points of the pixel point P and the left, right, upper and lower leading points are not all 1;
the condition G2 is that two connected points among 8 neighboring points of the pixel point P are combined and at least two or three of them are not empty after being combined;
the condition G3 indicates that the 4 top-right leading points of the pixel P must all be 0 or only the right neighboring point is 1;
the condition G3' is that 4 leading points at the lower left of the pixel point P must be all 0 or only left adjacent points are 1;
the specific formula is as follows:
(x 2 ∨x 3 ∨x 8 )∧x 1 not equal to 0 (condition-G3)
(x 6 ∨x 7 ∨x 4 )∧x 5 0 (condition-G3').
Further, making a normal line to each point in the two-dimensional framework of the elastic strip fastener, selecting a high-position Z-value corresponding point to determine the three-dimensional framework of the elastic strip fastener, and the method comprises the following steps:
making a normal line for each point in the two-dimensional framework DG (x, y) through a formula (11);
the normal line extends forwards and backwards along the x and y directions of the two-dimensional skeleton, corresponding x and y intervals in the point cloud data DT (x, y, z) of the elastic fastener with higher pixels are searched, and each interval of the point cloud data DT (x, y, z) of the corresponding elastic fastener is searched;
after the points with abnormal Z values are eliminated, selecting the points with higher Z values as Z value corresponding points of the elastic strip fastener three-dimensional framework to obtain the fastener elastic strip three-dimensional framework DG (x, y, Z);
the formula of the extraction method is as follows:
wherein z is all A set of z-vectors for all points; mean () is the median of the z vectors for all points.
Further, finding out a plurality of characteristic points from the three-dimensional skeleton data, and calculating the seam separating height of the elastic strip fastener according to the lowest normal point of the central recess of the elastic strip fastener, comprises:
dividing the elastic strip fastener three-dimensional framework DG (x, y, z) by adopting 2 x 3 grids to obtain 5 characteristic points A, B, C, D, E positioned on different grids; wherein, the point A is the lowest point of the central recess of the elastic strip; the point B and the point C are the lowest points of the elastic strips close to the track end; points E and D are the lowest points of the elastic strips far away from the track end; point F is the midpoint between points E and D;
the planar equation y ═ α x + β y + γ z + δ and a (x) by BCF point fitting 1 ，y 1 ，z 1 ) Point coordinates, according to a formula (12), calculating the distance d from the point A to the surface BCF, namely the seam separation height of the elastic strip fastener;
wherein, alpha, beta, gamma and delta are coefficients of a plane equation fitted according to BCF points.
Further, the method for detecting the tightness state of the rail elastic strip fastener provided by the embodiment of the invention further comprises the following steps:
distributing and classifying elastic strip deformation quantities d corresponding to a plurality of elastic strip fasteners in the same section of track, and calculating the standard deviation sigma of the deformation quantities d of the whole elastic strip fasteners;
and defining the elastic strip fastener which is more than 3 times of sigma as a failed fastener, and finding out the point cloud data D (x, y, z) of the corresponding elastic strip fastener to obtain the serial number and the position of the failed fastener.
Compared with the prior art, the tightness state detection method of the rail elastic strip fastener provided by the embodiment of the invention has the following beneficial effects:
the invention provides a method for detecting the buckling pressure failure defect of a fastener based on three-dimensional point cloud information, which is rapid, stable, accurate and high in applicability. Specifically, the deformation quantity of the steel rail fastener elastic strip is measured by utilizing three-dimensional point cloud information, and the distribution analysis is carried out according to the deformation quantity of the steel rail fastener elastic strip in a section of line, so that the actual engineering problem that the fastening pressure of the fastener can only be detected manually at present is solved; the invention can reduce the manual time for measuring the tightness state of the fastener, enhance the detection capability of the fastener state in the current track inspection instrument, and simultaneously provide a new detection method and means for the detection and operation maintenance of the railway; the invention adopts the fastener point cloud information, and carries out three-dimensional to two-dimensional to three-dimensional conversion on the data according to the clustering algorithm, the morphological operation, the KD-Tree and the like, thereby improving the overall operation efficiency and reducing the redundancy of the data.
Drawings
Fig. 1 is a flowchart of a method for detecting a tightness state of a rail elastic fastener according to an embodiment;
FIG. 2 is a schematic diagram of an embodiment of a Euclidean clustering process;
FIG. 3 is a flow diagram of a discrimination scheme provided in one embodiment;
FIG. 4 is a schematic flow chart illustrating conversion of a three-dimensional elastic bar point cloud into a three-dimensional elastic bar skeleton point cloud according to an embodiment;
FIG. 5 is a schematic illustration of a point of interest of a spring frame provided in one embodiment;
fig. 6 is a flowchart of the overall procedure provided in one embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more clearly understood, the present application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
Referring to fig. 1, in an embodiment, a method for detecting a tightness state of a rail elastic strip fastener includes:
s1, collecting point cloud information D (x, y, z) of the steel rail fastener.
S2, the fastener region in the point cloud data D (x, y, z) is divided, and a region point cloud DT (x, y, z) including only fastener clips is extracted separately.
S3, extracting the fastener elastic strip three-dimensional skeleton DG (x, y, z) according to the fastener elastic strip area point cloud DT (x, y, z).
And S4, finding out required characteristic points based on the extracted elastic strip framework DG (x, y, z), and calculating the elastic strip deformation d according to the normal direction of the characteristic points.
And S5, comparing the data of the elastic strip deformation d of the plurality of groups of fasteners measured in the circuit, thereby judging the fastening pressure tightness of a single fastener.
The step S1 specifically includes:
the three-dimensional line scanning camera is vertically arranged on the detection vehicle at a certain height so that line laser vertically irradiates on the steel rail fastener. Three-dimensional line scanning camera adopts laser triangle reflection principle, at first uses a branch of laser to shine the testee surface, and the reverberation forms the facula on the photosensitive element surface through optical lens group, and the facula position diverse that the surface reflection of co-altitude formed to form the 3D profile fast and gather rail fastener point cloud D (x, y, z).
In the step, point cloud information of the steel rail fastener is mainly acquired, and the high-precision point cloud information of the steel rail fastener can be acquired by using the line scanning camera, so that the accuracy and the applicability of the measuring method are improved.
As shown in fig. 2 and 3, the step S2 specifically includes the following steps:
s21, the height of the rail clip point cloud data D (x, y, z) is divided, and a point with a low height D _ z is excluded as D' (x, y, z). D ' _ y is sorted from small to large, a first derivative D (D ' _ y) of the D ' _ y is obtained, and possible fastener existence areas corresponding to the D ' _ y values are segmented according to the comparison between the D (D ' _ y) and the confidence value fib.
S22, finding each point cloud section in the corresponding steel rail fastener point cloud data D (x, y, z) according to the y value section corresponding to each fastener area obtained in S21, and recording point cloud of the fastener existing area as Di (x, y, z), wherein Di is the ith point cloud file after segmentation; processing the point cloud Di (x, y, z) of the fastener existing region in the formula (1) after being spliced with the upper point cloud file; and (3) processing the point cloud Di (x, y, z) of the fastener existing region in the formula (2) after being spliced with the lower point cloud file. And (4) judging all the rest Di after head and tail data processing by the formula (3), thereby obtaining point cloud Di (x, y, z) of the complete fastener existing region.
(MAX(Di_y)-MIN(Di_y))<y_fib；i＝1 (1)
(MAX(Di_y)-MIN(Di_y))<y_fib；i＝end (2)
Point(Di)>point_fib&&(MAX(Di_y)-MIN(Di_y))<y_fib；i∈[1,end] (3)
And Di _ y is data of a y vector in the point cloud Di (x, y, z) of the ith fastener existing region, y _ fib is a default fastener region y width value, and point _ fib is the minimum admissible fastener point cloud amount.
And S23, carrying out classification analysis on the complete fastener existing region point cloud Di (x, y, z) separated by the S22 by an Euclidean clustering method. And selecting a certain point from the point cloud Di (x, y, z) of the area where the single complete fastener exists, finding n points nearest to the point through a KD-Tree nearest neighbor search algorithm, and clustering the points with the distance smaller than a set threshold value into a set. If the number of elements in the set is not increased, the whole clustering process is ended; otherwise, another point is selected from the set to replace the original point, and the process is repeated until the number of the elements in the set is not increased any more, so that the point cloud Di (x, y, z) is clustered.
After classification, the category with less point cloud data amount (less than 1% of the total number of the point clouds) and the category with smaller MEAN (Di _ z), namely the average z vector, are excluded from all the categories.
S24, sequencing according to the remaining Z values of Di (x, y, Z), obtaining a first derivative d (Di _ Z), comparing the first derivative d (Di _ Z) with the first-order Z value d (Di _ Z) according to a preset confidence value (fid), and rapidly deleting the point cloud through the concept of a binary method to obtain the point cloud DT (x, y, Z) of the fastener elastic strip.
In the step, segmentation and processing are carried out according to original point cloud information, classification and segmentation are carried out according to the relation of the serial numbers of the corresponding points of all the points (x, y, z) in the point cloud data and the relation corresponding to other points, and finally the point cloud data only with fastener elastic strips are extracted.
As shown in fig. 4, the step S3 specifically includes the following steps:
and S31, obtaining the fastener length and width pixel proportion corresponding to the point cloud image by using the fastener elastic strip point cloud DT (x, y, z) obtained in S2, and taking the fastener length and width pixel proportion as the length and width pixel points of the binary image to be converted.
y point ＝ceil(x point *p) (5)
Wherein, y point 、x point Pixel points, x, set for converting binary images point And Ax and Ay are corresponding x and y column vectors in the bullet point cloud data as initial set values. The point cloud is zoomed according to the scheme, and only the approximate point cloud range of the region of interest is found by converting the two-dimensional image. The divided binary image is similar to a grid structure, the original point cloud image is divided into single grids, and each grid, namely the point cloud range corresponding to each pixel point of the binary image, is as follows:
XI＝min(A x )：(x point +1)：max(A x ) (6)
YI＝min(A y )：(y point +1)：max(A y ) (7)
YI(j)＝＜A y ＜YI(j+1)i∈(1：x point ) (8)
XI(i)＝＜A x ＜XI(i+1)j∈(1：y point ) (9)
the fastener spring point cloud DT (x, y, z) is converted into a binary image G (x, y) according to the range.
S32, performing a morphological closing operation on the binary image using a disc-shaped structuring element SE. The morphological closing operation is to expand first and then corrode, and the two operations use the same structural elements, and the specific formula is shown as (10). Finally, obtaining a continuous binary image G' (x, y) of each pixel point;
s33, extracting a skeleton of the binary image G' (x, y) with continuous pixels, extracting a target peripheral contour, corroding the boundary of the target image by using the contour until the boundary cannot be corroded, and narrowing the object into a line. And deleting redundant pixels, so that the object without holes is shrunk into a line with minimum connectivity, and the object with holes is shrunk into a communication ring between each hole and the outer boundary, thereby obtaining a two-dimensional framework DG (x, y).
And (3) extracting a skeleton of a binary image G' (x, y) with continuous pixel points, judging the values of 8 adjacent points around the binary image, and deleting the pixel point p in the first sub-iteration if and only if all conditions G1, G2 and G3 are met. In the second sub-iteration, pixel point p is deleted if and only if the conditions G1, G2, and G3' are all satisfied. In the skeleton extraction process, the first iteration is adopted, but after the pixel points of the whole binary image are not transformed any more, the second iteration scheme is adopted, and the process is repeated until the whole image is reduced into one line, namely the pixel p is not satisfied (condition G1& condition G2& condition G3) or (condition G1& condition G2& condition G3').
Wherein the condition G1 is that more than one directly connected point among 8 adjacent points of the pixel point P and the left, right, upper and lower leading points are not all 1; the condition G2 is that two connected points among 8 neighboring points of the pixel point P are combined and at least two or three of them are not empty after being combined; the condition G3 is that the 4 top-right leading points of the pixel point P must all be 0 or only the right neighboring point is 1; the condition G3' is that the 4 top-down points at the left of the pixel point P must all be 0 or only the left neighboring point is 1; the specific formula is as follows:
(x 2 ∨x 3 ∨x 8 )∧x 1 not equal to 0 (condition-G3)
(x 6 ∨x 7 ∨x 4 )∧x 5 0 (Condition-G3')
S34, each point in the two-dimensional framework DG (x, y) is taken as a normal line, the normal line is extended forwards and backwards along the x and y directions of the two-dimensional framework, corresponding x and y intervals in the three-dimensional point cloud space DT (x, y, Z) of higher pixels are searched, so that each interval of the three-dimensional point cloud space DT (x, y, Z) is corresponding, then after points with Z abnormity are eliminated, points with higher Z values are selected as Z value corresponding points of the elastic strip three-dimensional framework, so that the fastener elastic strip three-dimensional framework DG (x, y, Z) is obtained, and the formula for taking the higher points is as (11).
In the step, two-dimensional binary image conversion is carried out according to the extracted elastic strip point cloud data, the binary image is subjected to closed operation according to morphological analysis, and then a corresponding two-dimensional skeleton is extracted; and finally converting the two-dimensional elastic strip skeleton into a three-dimensional point cloud skeleton by matching and comparing the two-dimensional elastic strip skeleton with the original point cloud data.
As shown in fig. 5, the implementation method of the step S4 includes the following steps:
s41, dividing a three-dimensional framework DG (x, y, z) of the fastener elastic strip, and dividing the fastener elastic strip by adopting 2-3 grids, wherein the elastic strip is an approximately bilaterally symmetrical entity, so A, B, C, D, E five points are necessarily located in 5 different grids. And (3) adopting the distance from the point A to the surface BCF as a deformation index of the elastic strip so as to estimate whether the fastener is compressed or not.
Wherein:
the point A is the lowest point of the concave part in the center of the elastic strip;
the point B and the point C are the lowest points of the elastic strips close to the track end;
points E and D are the lowest points of the elastic strips far away from the track end;
point F is the midpoint between points E and D.
S42, a (x) obtained from the BCF point-fitted plane equation y ═ α x + β y + γ z + δ and S41 1 ， y 1 ，z 1 ) The point coordinates are obtained by calculating the distance from point A to the surface BCF according to equation (12), where d is the distance.
In the step, A, B, C, D, E five points are found through a characteristic extraction mode according to the extracted three-dimensional point cloud framework, a midpoint F is calculated according to D, E points, a plane BCF is fitted according to the points B, C, F, the distance d between the point A and the plane BCF is calculated, and the elastic deformation quantity d of the elastic strip is recorded and stored.
The method for implementing the above step S5 includes the following steps:
and S51, recording and storing the elastic deformation d of the elastic strip in each point cloud file, carrying out distribution classification on the elastic deformation d of the whole elastic strip, and calculating the standard deviation sigma of the elastic deformation d of the whole elastic strip.
S52, defining the deviation point as the point of failure, which is larger than 3 times sigma. And finding out the corresponding point cloud information D (x, y, z) of the steel rail fastener, thereby obtaining the number and the position of the failed fastener.
In the step, the elastic deformation d of the elastic strip is analyzed according to the calculated elastic deformation amount d, and finally, a failed fastener is found or all the fasteners are allowed to be reliable.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the protection scope of the present application. Therefore, the protection scope of the present patent should be subject to the appended claims.
Claims (10)
1. A method for detecting the tightness state of a rail elastic strip fastener is characterized by comprising the following steps:
acquiring point cloud data of the rail elastic strip fastener;
dividing an elastic strip fastener area in the track elastic strip fastener point cloud data, and extracting point cloud data of the elastic strip fastener from the elastic strip fastener area;
converting the point cloud data of the elastic strip fastener into a binary image, and extracting a two-dimensional skeleton of the elastic strip fastener from the binary image; selecting high position Z value corresponding points by taking a normal line of each point in the two-dimensional framework of the elastic strip fastener to determine the three-dimensional framework of the elastic strip fastener;
finding out a plurality of minimum characteristic points from the three-dimensional skeleton data, and calculating the seam separation height of the elastic strip fastener according to the lowest point of the central recess of the elastic strip fastener in the normal direction; and the height of the gap is an evaluation index of the tightness state of the elastic strip fastener.
2. The method for detecting the tightness of a rail elastic fastener according to claim 1, wherein the obtaining of the point cloud data of the rail elastic fastener specifically includes:
vertically installing a three-dimensional line scanning camera on a detection vehicle, and vertically irradiating line laser on a rail elastic strip fastener;
the three-dimensional line scanning camera emits a beam of laser to irradiate the surface of a measured object, reflected light forms light spots on the surface of the photosensitive element through the optical lens group, the positions of the light spots formed by the reflection of the surfaces with different heights are different, namely 3D contour height is formed, and X, Y and Z coordinates are constructed according to the position of the 3D contour height to form three-dimensional point cloud data.
3. The method for detecting the tightness state of a track elastic strip fastener according to claim 1, wherein the step of segmenting the elastic strip fastener area in the track elastic strip fastener point cloud data and extracting the point cloud data of the elastic strip fastener from the elastic strip fastener area comprises:
dividing height direction data D _ z of the track elastic strip fastener point cloud data D (x, y, z), and excluding the track elastic strip fastener point cloud data D (x, y, z) corresponding to a point with a lower height D _ z to obtain point cloud data D' (x, y, z);
sorting the point cloud data D ' (x, y, z) according to the size of y vector data D ' _ y of the point cloud data D ' (x, y, z), solving a derivative D (D ' _ y) of the D ' _ y, and comparing the D (D ' _ y) with a confidence value fib to segment a bullet strip fastener existence area corresponding to each D ' _ y value;
finding each point cloud interval in the point cloud data D (x, y, z) of the corresponding track elastic strip fastener according to the y value interval corresponding to each elastic strip fastener existing area, and recording the point cloud data of the ith elastic strip fastener existing area as Di (x, y, z);
splicing point cloud data in an upper point cloud file by using the point cloud data Di (x, y, z) of the elastic strip fastener existence area which meets the formula (1), splicing point cloud data in a lower point cloud file by using the point cloud data Di (x, y, z) of the elastic strip fastener existence area which meets the formula (2), and judging the remaining point cloud data Di (x, y, z) of the elastic strip fastener existence area by using the formula (3) to obtain complete point cloud Di (x, y, z) of the elastic strip fastener existence area;
(MAX(Di_y)-MIN(Di_y))<y_fib；i＝1 (1)
(MAX(Di_y)-MIN(Di_y))<y_fib；i＝end (2)
Point(Di)>point_fib&&(MAX(Di_y)-MIN(Di_y))<y_fib；i∈[1,end](3)
the data Di _ y is data of a y vector in point cloud data Di (x, y, z) of the ith elastic strip fastener existing region; y _ fib is the width value of the default elastic strip fastener area in the y direction; point _ fib is the cloud amount of the point of the minimum bearing considered elastic strip fastener; MAX () is the maximum value; MIN () is minimum; point (di) is the point cloud amount of point cloud data of the area where the ith elastic strip fastener exists; end is the area where the last spring strip fastener in a point cloud file exists.
4. The method of detecting a tightness state of a track spring bar fastener according to claim 3, wherein the step of segmenting a spring bar fastener area in the point cloud data of the track spring bar fastener and extracting the point cloud data of the spring bar fastener from the spring bar fastener area further comprises:
classifying point cloud data Di (x, y, z) of the existing area of the complete elastic strip fastener by an Euclidean clustering method;
in each classified category, excluding the category of which the point cloud data volume is less than 1 per mill of the total point cloud number and the category of which the MEAN (Di _ z) is smaller; the MEAN (Di _ z) is the average value of z vectors in the point cloud data Di (x, y, z) of the ith elastic strip fastener existing region;
sorting according to the z values of the rest Di (x, y, z), and solving a first derivative d (Di _ z);
and comparing the preset confidence value fid with the solved first-order Z value d (Di _ Z), and deleting the point cloud data by the dichotomy principle to obtain point cloud data DT (x, y, Z) of the elastic strip fastener.
5. The method for detecting the tightness of a rail elastic fastener according to claim 4, wherein the classification of the point cloud data Di (x, y, z) of the existence area of the complete elastic fastener by Euclidean clustering method comprises:
selecting one point of point cloud data Di (x, y, z) of the area where a single complete elastic strip fastener exists;
finding n points nearest to the point through a KD-Tree neighbor search algorithm, and clustering the points with the distance smaller than a set threshold value into a set; if the number of elements in the set is not increased any more, the whole clustering process is ended; otherwise, another point in the set is selected for re-clustering until the number of elements in the set is not increased any more.
6. The method for detecting the tightness of a rail spring fastener according to claim 1, wherein the converting point cloud data of the spring fastener into a binary image comprises:
obtaining the length-width pixel proportion of the elastic strip fastener corresponding to the point cloud image according to the point cloud data DT (x, y, z) of the elastic strip fastener, and taking the length-width pixel proportion as the length-width pixel point of the binary image to be converted:
y point ＝ceil(x point *p) (5)
wherein, y point 、x point Pixel points set for converting the binary image; x is the number of point The Ax and Ay are respectively corresponding x and y row vectors in point cloud data of the elastic strip fastener as an initial set value;
and (3) dividing the point cloud image into single grids according to the similarity of the divided binary image to a grid structure, wherein each grid, namely the point cloud range corresponding to each pixel point of the binary image is as follows:
XI＝min(A x )：(x point +1)：max(A x ) (6)
YI＝min(A y )：(y point +1)：max(A y ) (7)
YI(j)＝＜A y ＜YI(j+1) i∈(1：x point ) (8)
XI(i)＝＜A x ＜XI(i+1) j∈(1：y point ) (9)
wherein, YI (j) is a y vector of a binary image corresponding to the pixel point j; xi (j) is the x vector of the binary image corresponding to the pixel point j;
converting point cloud data DT (x, y, z) of the elastic strip fastener into a binary image G (x, y) according to the point cloud range;
performing morphological closed operation on the binary image by using a disc-shaped structural element SE, specifically as a formula (10), and obtaining a binary image G' (x, y) with continuous pixel points;
wherein,
is a morphological dilation treatment and Θ is a morphological erosion treatment.
7. The method for detecting the tightness state of the track elastic strip fastener according to claim 6, wherein the extracting the two-dimensional skeleton of the elastic strip fastener from the binarized image comprises:
extracting a target peripheral contour from the continuous binary image G' (x, y) of each pixel point;
corroding the boundary of a target image by using the contour, narrowing the object into a line, and deleting redundant pixels to shrink the object without holes into the line with minimum connectivity, and shrinking the object with holes into a communication ring between each hole and the outer boundary to obtain a two-dimensional framework DG (x, y);
extract out the two-dimensional skeleton of bullet strip fastener from binary image, specifically include:
judging the values of 8 neighboring points around the binary image G' (x, y) with continuous pixels;
in the first subiteration, if and only if all the conditions G1, G2, and G3 are satisfied, deleting the pixel point p;
in the second sub-iteration, if and only if the conditions G1, G2, and G3' are all satisfied, deleting pixel point p;
in the skeleton extraction process, a first iteration is adopted, when the pixel points of the whole binary image are not transformed any more, a second iteration scheme is adopted, the process is repeated until the whole image is reduced into a line, namely, no pixel p is satisfied: (condition G1& condition G2& condition G3) or (condition G1& condition G2& condition G3'); wherein,
the condition G1 is that more than one directly connected point among 8 adjacent points of the pixel point P and the left, right, upper and lower leading points are not all 1;
the condition G2 is that two connected points among 8 neighboring points of the pixel point P are combined and at least two or three of them are not empty after being combined;
the condition G3 is that the 4 top-right leading points of the pixel point P must all be 0 or only the right neighboring point is 1;
the condition G3' is that 4 leading points at the lower left of the pixel point P must be all 0 or only left adjacent points are 1;
the specific formula is as follows:
x1, x2, … are the values of the neighbors of p, numbered counterclockwise starting from the right neighbor; (Condition-G1)
(x 2 ∨x 3 ∨x 8 )∧x 1 Not equal to 0 (condition-G3)
(x 6 ∨x 7 ∨x 4 )∧x 5 0 (condition-G3').
8. The method for detecting the tightness of a rail spring clip according to claim 7, wherein the determining of the three-dimensional frame of the spring clip by selecting the corresponding point of the high Z value as the normal to each point of the two-dimensional frame of the spring clip comprises:
making a normal line for each point in the two-dimensional framework DG (x, y) through a formula (11);
the normal line extends forwards and backwards along the x and y directions of the two-dimensional skeleton, corresponding x and y intervals in the point cloud data DT (x, y and z) of the elastic strip fastener with higher pixels are searched, and each interval of the point cloud data DT (x, y and z) of the corresponding elastic strip fastener is searched;
after the points with abnormal Z values are eliminated, selecting the points with higher Z values as Z value corresponding points of the elastic strip fastener three-dimensional framework to obtain a fastener elastic strip three-dimensional framework DG (x, y, Z);
the formula of the extraction method is as follows:
z 1 ∈median(z all )±(max(z all )-min(z all ))*0.4 (11)
wherein z is all A set of z-vectors for all points; mean () is the median of the z-vectors for all points.
9. The method of claim 1, wherein the step of finding a plurality of feature points from the three-dimensional skeleton data and calculating the seam crossing height of the elastic strip fastener according to the lowest point of the central recess of the elastic strip fastener comprises:
dividing the elastic strip fastener three-dimensional framework DG (x, y, z) by adopting 2 x 3 grids to obtain 5 characteristic points A, B, C, D, E positioned on different grids; wherein, the point A is the lowest point of the central recess of the elastic strip; the point B and the point C are the lowest points of the elastic strips close to the track end; points E and D are the lowest points of the elastic strips far away from the track end; point F is the midpoint between points E and D;
the planar equation y ═ α x + β y + γ z + δ and a (x) by BCF point fitting 1 ，y 1 ，z 1 ) Point coordinates, according to a formula (12), calculating the distance d from the point A to the surface BCF, namely the seam separation height of the elastic strip fastener;
wherein α, β, γ, δ are coefficients of a plane equation fitted according to BCF points.
10. The method of detecting the tightness of a rail spring clip according to claim 9, further comprising:
distributing and classifying elastic strip deformation quantities d corresponding to a plurality of elastic strip fasteners in the same section of track, and calculating the standard deviation sigma of the deformation quantities d of the whole elastic strip fasteners;
and defining the elastic strip fastener with the sigma larger than 3 times as a failure fastener, and finding out the point cloud data D (x, y, z) of the corresponding elastic strip fastener to obtain the serial number and the position of the failure fastener.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210545365.7A CN115100109B (en) | 2022-05-19 | 2022-05-19 | Method for detecting tightness state of rail elastic strip fastener |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210545365.7A CN115100109B (en) | 2022-05-19 | 2022-05-19 | Method for detecting tightness state of rail elastic strip fastener |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115100109A true CN115100109A (en) | 2022-09-23 |
| CN115100109B CN115100109B (en) | 2023-04-18 |
Family
ID=83289738
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210545365.7A Active CN115100109B (en) | 2022-05-19 | 2022-05-19 | Method for detecting tightness state of rail elastic strip fastener |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115100109B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117635520A (en) * | 2023-03-20 | 2024-03-01 | 南昌工程学院 | Elastic strip fastener tightness detection method by fusing two-dimensional image and three-dimensional point cloud |
| CN117830199A (en) * | 2023-04-14 | 2024-04-05 | 南昌工程学院 | Comprehensive detection method for defects of track elastic strip fastener by fusing image and point cloud |
| CN118761994A (en) * | 2024-09-03 | 2024-10-11 | 中国铁路设计集团有限公司 | Quick detection and judgment method for service state of fastener |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109870459A (en) * | 2019-02-21 | 2019-06-11 | 武汉武大卓越科技有限责任公司 | The track plates crack detection method of non-fragment orbit |
| CN110634122A (en) * | 2018-06-05 | 2019-12-31 | 成都精工华耀科技有限公司 | Fastener elastic strip displacement detection method |
| CN111391881A (en) * | 2020-04-03 | 2020-07-10 | 中国铁道科学研究院集团有限公司基础设施检测研究所 | Method and device for detecting fastening state of steel rail fastener |
| CN113469940A (en) * | 2021-05-26 | 2021-10-01 | 成都唐源电气股份有限公司 | Fastener looseness detection method based on three-dimensional point cloud and two-dimensional image processing technology |
| CN113781537A (en) * | 2021-09-17 | 2021-12-10 | 南昌工程学院 | Track elastic strip fastener defect identification method and device and computer equipment |
| CN113808112A (en) * | 2021-09-22 | 2021-12-17 | 北京京东乾石科技有限公司 | Track fastener detection method, electronic device and computer readable medium |
-
2022
- 2022-05-19 CN CN202210545365.7A patent/CN115100109B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110634122A (en) * | 2018-06-05 | 2019-12-31 | 成都精工华耀科技有限公司 | Fastener elastic strip displacement detection method |
| CN109870459A (en) * | 2019-02-21 | 2019-06-11 | 武汉武大卓越科技有限责任公司 | The track plates crack detection method of non-fragment orbit |
| CN111391881A (en) * | 2020-04-03 | 2020-07-10 | 中国铁道科学研究院集团有限公司基础设施检测研究所 | Method and device for detecting fastening state of steel rail fastener |
| CN113469940A (en) * | 2021-05-26 | 2021-10-01 | 成都唐源电气股份有限公司 | Fastener looseness detection method based on three-dimensional point cloud and two-dimensional image processing technology |
| CN113781537A (en) * | 2021-09-17 | 2021-12-10 | 南昌工程学院 | Track elastic strip fastener defect identification method and device and computer equipment |
| CN113808112A (en) * | 2021-09-22 | 2021-12-17 | 北京京东乾石科技有限公司 | Track fastener detection method, electronic device and computer readable medium |
Non-Patent Citations (1)
| Title |
|---|
| 寇东华: "基于多传感器集成的无砟轨道轨道板裂缝与离缝自动检测装置", 《中国铁路》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117635520A (en) * | 2023-03-20 | 2024-03-01 | 南昌工程学院 | Elastic strip fastener tightness detection method by fusing two-dimensional image and three-dimensional point cloud |
| CN117830199A (en) * | 2023-04-14 | 2024-04-05 | 南昌工程学院 | Comprehensive detection method for defects of track elastic strip fastener by fusing image and point cloud |
| CN117830199B (en) * | 2023-04-14 | 2024-10-18 | 南昌工程学院 | Comprehensive detection method for defects of track elastic strip fastener by fusing image and point cloud |
| CN118761994A (en) * | 2024-09-03 | 2024-10-11 | 中国铁路设计集团有限公司 | Quick detection and judgment method for service state of fastener |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115100109B (en) | 2023-04-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115100109B (en) | Method for detecting tightness state of rail elastic strip fastener | |
| Liu et al. | A review of applications of visual inspection technology based on image processing in the railway industry | |
| Yan et al. | Automated extraction of structural elements in steel girder bridges from laser point clouds | |
| KR102565689B1 (en) | Method of Mobile Safety Diagnosis and monitoring for facilities based on deep learning and Apparatus thereof | |
| CN110866430B (en) | License plate recognition method and device | |
| Guldur et al. | Condition assessment of bridges using terrestrial laser scanners | |
| Salari et al. | Pavement distress detection and severity analysis | |
| CN107798293A (en) | A kind of crack on road detection means | |
| CN114581764B (en) | Underground structure crack disease discriminating method based on deep learning algorithm | |
| CN112651988B (en) | Finger-shaped image segmentation, finger-shaped plate staggering and abnormal fastener detection method based on double-pointer positioning | |
| CN108508023B (en) | Defect detection system for contact end jacking bolt in railway contact network | |
| CN112508911A (en) | Rail joint touch net suspension support component crack detection system based on inspection robot and detection method thereof | |
| CN117036326A (en) | Defect detection method based on multi-mode fusion | |
| KR102311558B1 (en) | System and method for detecting structure damage using artificial intelligence, and a recording medium having computer readable program for executing the method | |
| CN115841477A (en) | Multi-mode visual fusion ballastless track structure hidden defect detection method and system | |
| CN112508913A (en) | Cable section edge detection method based on image detection | |
| CN108931206A (en) | Method for distinguishing is known for rail profile outlier detection and effective profile | |
| CN115239666A (en) | Method for positioning and segmenting defect target area image of rolling contact fatigue defect | |
| Wang et al. | Automated shape-based pavement crack detection approach | |
| Ibragimov et al. | Automated Pavement Condition Index Assessment with Deep Learning and Image Analysis: An End-to-End Approach | |
| CN110672632A (en) | Tunnel disease identification method | |
| Guinard et al. | Fast Weakly Supervised Detection of Railway-Related Infrastructures in Lidar Acquisitions | |
| Wang et al. | 3D tensor-based point cloud and image fusion for robust detection and measurement of rail surface defects | |
| Yao et al. | Automated detection and identification of area-based distress in concrete pavements | |
| Shao et al. | Automated Two-Stage Approach for Damage Detection of Surface Defects in Historical Buildings |
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 |