CN115272287A - Fault detection method, medium and system for rail wagon buffer and slave plate - Google Patents
Fault detection method, medium and system for rail wagon buffer and slave plate Download PDFInfo
- Publication number
- CN115272287A CN115272287A CN202210999696.8A CN202210999696A CN115272287A CN 115272287 A CN115272287 A CN 115272287A CN 202210999696 A CN202210999696 A CN 202210999696A CN 115272287 A CN115272287 A CN 115272287A
- Authority
- CN
- China
- Prior art keywords
- buffer
- slave plate
- slave
- straight line
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/02—Knowledge representation; Symbolic representation
- G06N5/022—Knowledge engineering; Knowledge acquisition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/04—Inference or reasoning models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/168—Segmentation; Edge detection involving transform domain methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T9/00—Image coding
-
- 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
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/20084—Artificial neural networks [ANN]
-
- 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
- G06T2207/30248—Vehicle exterior or interior
-
- 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)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Quality & Reliability (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
A method, a medium and a system for detecting faults of a wagon buffer and a slave plate relate to the technical field of image processing, and aim at the problem that the accuracy rate of artificially identifying the wagon buffer and the slave plate faults is low in the prior art, the method and the system introduce learning-RCNN training by using fast-RCNN. The position relation between the slave plates and the buffer, the position relation between the slave plates and the damaged slave plates and the position relation between the damaged buffer and the damaged buffer are used as characteristics to be introduced into the network, and therefore the fault detection accuracy rate is improved. According to the method, a canny operator is modified, the wheel point position is output, the gradient direction of the contour point is output at the same time, a Hough transform straight line detection method is improved, and normal direction parameters on the contour point are introduced. The accuracy and the precision rate of detecting the straight line are improved.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a railway wagon buffer and slave plate fault detection method based on deep learning.
Background
The truck buffer mainly comprises a front slave plate, a rear slave plate and the buffer, and the buffer is easy to break, displace and other faults due to collision and other reasons in the running process of a truck. And (4) detecting faults of the truck buffer and the slave plate, and generally carrying out fault maintenance in a manual troubleshooting mode at the present stage. The detection operation is greatly influenced by factors such as the quality of service, the responsibility, the labor intensity and the like of an operator, the conditions of missed detection or simplified operation and the like are easy to occur, and further the problem of low accuracy is caused.
Disclosure of Invention
The purpose of the invention is: aiming at the problem that the accuracy rate of artificially identifying faults of a wagon buffer and a slave plate is low in the prior art, the fault detection method for the wagon buffer and the slave plate is provided.
The technical scheme adopted by the invention for solving the technical problems is as follows:
a rail wagon buffer and slave plate fault detection method comprises the following steps:
the method comprises the following steps: acquiring a buffer of the rail wagon and an image of a slave plate;
step two: marking the buffer position, the slave plate position, the buffer damage position and the slave plate damage position in the railway wagon buffer and slave plate image, and constructing a training set by using the marked image, the railway wagon buffer and the slave plate image;
step three: introducing Reasoning-RCNN into a FasterRcnn model to obtain an improved FasterRcnn model, and training the improved FasterRcnn model by utilizing a training set to obtain a well-trained improved FasterRcnn model;
the modified FasterRcnn model specifically performs the following steps:
firstly, generating the characteristics of each label category by using a basic detection network of a FasterRcnn model, and simultaneously generating a global semantic pool;
then, introducing the position relation between the slave plates and the buffer, the position relation between the slave plates and the damaged slave plates, and the position relation between the damaged buffer and the damaged buffer as prior knowledge into the map, defining a class-to-class undirected graph, coding the prior knowledge of each edge in the undirected graph, coding all the edges in the undirected graph to obtain a knowledge map, then spreading the knowledge map in a global semantic pool, and generating reinforced features through a global reasoning module;
finally, connecting the enhanced features with features generated by a basic detection network to obtain new features;
step four: inputting an image to be recognized into a trained improved FasterRcnn model to obtain a buffer position, a slave plate position, a buffer damage position and a slave plate damage position in the image;
step five: judging whether a fault exists according to the result obtained in the fourth step;
the concrete steps of the fifth step are as follows: determining a failure based on the buffer position, the slave plate position, the buffer damage position, and the slave plate damage position in the obtained image, if at least one of them is true;
the relative position of the buffer position and the slave plate position in the image is less than a threshold;
the coincidence degree of the buffer position and the buffer damage position in the image exceeds a threshold value;
the coincidence of the slave plate position and the slave plate breakage position in the image exceeds a threshold value.
Further, the method comprises the following steps:
if no fault is detected in the fifth step, executing a sixth step;
step six: cutting the slave plate graph according to the slave plate position obtained in the step four;
step seven: acquiring contour points of the slave plate graph, and acquiring the positions and gradient directions of the contour points;
step eight: and judging whether the slave plate fails according to the position and gradient direction of the contour point and by combining a Hough transform detection straight line method.
Further, the step eight includes the specific steps of:
step eight one: introducing the position and the gradient direction of the contour points in a Hough transform detection straight line method, and obtaining r and theta on a straight line where each contour point is located according to the position and the gradient direction of the contour points, wherein r is the distance from an original point to the nearest point on the straight line, and theta is an included angle between an x axis and the straight line connecting the original point and the nearest point;
step eight two: constructing a two-dimensional matrix according to r and theta on the straight line where each contour point is located;
step eight and three: and calculating a secondary moment by taking the point of the maximum value in the two-dimensional matrix as a center and the number of contour points as radii, wherein the maximum point of the secondary moment is r and theta of the slave plate boundary straight line, finally obtaining the straight line slope according to theta, and judging that the slave plate has a fault when the straight line slope is greater than a threshold value.
Further, the slope of the line is expressed as:
k＝-cosθ/sinθ。
further, r is represented as:
r＝x*cosθ+y*sinθ。
further, the seventh step is preceded by a step of denoising the board graph.
Further, the denoising is performed by gaussian filtering.
Further, the contour points of the plate graph are obtained through a Canny operator.
The beneficial effects of the invention are:
1. the application uses fast-RCNN to introduce learning-RCNN training. The position relation between the slave plates and the buffer, the position relation between the slave plates and the damaged slave plates and the position relation between the damaged buffer and the damaged buffer are used as characteristics to be introduced into the network, and therefore the fault detection accuracy rate is improved.
2. And modifying a canny operator, outputting the wheel point position and the gradient direction of the contour point, improving a Hough transform detection straight line method, and introducing normal direction parameters on the contour point. The accuracy and the precision rate of detecting the straight line are improved.
Drawings
FIG. 1 is a schematic diagram of a network architecture of the present application;
fig. 2 is a schematic diagram of a buffer and a slave board.
Detailed Description
It should be noted that, in the case of conflict, the various embodiments disclosed in the present application may be combined with each other.
The first embodiment is as follows: specifically describing the present embodiment with reference to fig. 1, a method for detecting a trouble in a buffer and slave plate of a railway freight car according to the present embodiment includes the steps of:
the method comprises the following steps: acquiring a buffer of the rail wagon and an image of a slave plate;
step two: marking the buffer position, the slave plate position, the buffer damage position and the slave plate damage position in the railway wagon buffer and slave plate image, and constructing a training set by using the marked image, the railway wagon buffer and the slave plate image;
step three: introducing Reasoning-RCNN into a FasterRcnn model to obtain an improved FasterRcnn model, and training the improved FasterRcnn model by utilizing a training set to obtain a well-trained improved FasterRcnn model;
the modified FasterRcnn model specifically performs the following steps:
firstly, generating the characteristics of each label category by using a basic detection network of a FasterRcnn model, and simultaneously generating a global semantic pool;
then, introducing the position relation between the slave plates and the buffer, the position relation between the slave plates and the damaged slave plates, and the position relation between the damaged buffer and the damaged buffer as prior knowledge into the map, defining a class-to-class undirected graph, coding the prior knowledge of each edge in the undirected graph, coding all the edges in the undirected graph to obtain a knowledge map, then spreading the knowledge map in a global semantic pool, and generating reinforced features through a global reasoning module;
finally, connecting the enhanced features with features generated by a basic detection network to obtain new features;
step four: inputting an image to be recognized into a trained improved FasterRcnn model to obtain a buffer position, a slave plate position, a buffer damage position and a slave plate damage position in the image;
step five: and judging whether a fault exists according to the result obtained in the fourth step.
The second embodiment is as follows: this embodiment mode is a further description of the first embodiment mode, and is different from the first embodiment mode in that the method further includes the steps of:
if no fault is detected in the fifth step, executing a sixth step;
step six: cutting out a slave plate graph according to the slave plate position obtained in the fourth step;
step seven: acquiring contour points of the slave plate graph, and acquiring the positions and gradient directions of the contour points;
step eight: and judging whether the slave plate fails according to the position and gradient direction of the contour point and by combining a Hough transform detection straight line method.
The third concrete implementation mode: this embodiment mode is a further description of the second embodiment mode, and is different from the second embodiment mode in that the specific step of step eight is:
step eight one: introducing the position and the gradient direction of the contour points in a Hough transform detection straight line method, and obtaining r and theta on a straight line where each contour point is located according to the position and the gradient direction of the contour points, wherein r is the distance from an original point to the nearest point on the straight line, and theta is an included angle between an x axis and the straight line connecting the original point and the nearest point;
and eighthly: constructing a two-dimensional matrix according to r and theta on a straight line where each contour point is located;
step eight and three: and calculating a secondary moment by taking the point of the maximum value in the two-dimensional matrix as the center and the number of contour points as the radius, wherein the maximum point of the secondary moment is the r and theta of the slave plate boundary straight line, finally obtaining the slope of the straight line according to the theta, and judging that the slave plate fails when the slope of the straight line is greater than a threshold value.
The fourth concrete implementation mode is as follows: this embodiment mode is a further description of a third embodiment mode, and is different from the third embodiment mode in that the slope of the straight line is expressed as:
k＝-cosθ/sinθ。
the fifth concrete implementation mode: this embodiment mode is a further description of a fourth embodiment mode, and is different from the fourth embodiment mode in that r is represented by:
r＝x*cosθ+y*sinθ。
the sixth specific implementation mode: the present embodiment is a further description of a fifth embodiment, and the difference between the present embodiment and the fifth embodiment is that a step of removing noise from a board graph is further included before the seventh step.
The seventh embodiment: the present embodiment is a further description of a sixth specific embodiment, and the difference between the present embodiment and the sixth specific embodiment is that denoising is performed by gaussian filtering.
The specific implementation mode is eight: the present embodiment is further described with respect to a seventh embodiment, and the difference between the present embodiment and the seventh embodiment is that contour points from a board graph are acquired by Canny operator.
The specific implementation method nine: a medium having a program embodied therein for executing the method of claims 1 through 8.
The specific implementation mode is ten: a system having a fault detection module embodied to perform the method of claims 1 to 8.
1. Improved fast-RCNN network model
The training network adopts a fast-RCNN model, and introduces learning-RCNN to improve the accuracy of the algorithm. To increase the speed of operation of the algorithm, the VGG16 is used as a backbone feature extraction network. The overall training model is shown in fig. 1. The learning-RCNN traditional fast-RCNN is enhanced by using a class-level knowledge graph to encode existing semantic knowledge. And introducing the position relation between the slave plates and the buffer, the position relation between the slave plates and the broken slave plates, and the position relation between the broken buffer and the broken buffer into a knowledge spectrogram as prior knowledge, and carrying out evolution and propagation in the knowledge spectrogram. The enhancement of classification characteristics is realized in the detection. As shown in fig. 2: the buffers and slave plates are located on the upper and lower sides of the pallet in the middle of the image, in the image at similar horizontal positions, and the upper and lower parts are in the same position in the vertical direction. Breakage of a component only occurs in the area where the corresponding component is located. The method establishes an adaptive global reasoning module by introducing external knowledge. Firstly, generating a global semantic pool by collecting the weight of a classification layer of the previous layer, namely integrating the high-level semantic representation of each category; then, defining a class-to-class undirected graph in training and testing, wherein each edge encodes knowledge between two nodes, and the knowledge graph is propagated in a global semantic pool to enhance regional characteristics; and finally, connecting the enhanced features with the original features to obtain new features for enhancing the detection effect.
2. And (5) training the model.
Intercepting a middle area image shot by a bottom camera, simultaneously acquiring subgraphs including damage of the slave plate and damage of the buffer, and performing a large amount of simulation aiming at possible damage faults, and marking a coupler buffer, the slave plate and a corresponding damage position. Training was performed using a modified faster-rcnn model. The input is a buffer and slave board graph, the buffer and slave board position in the output image, the buffer damage and the slave board damage failure.
3. And (6) judging a fault.
And intercepting a coupler buffer position subgraph in the identification process, inputting the subgraph into a trained false-rcnn model, acquiring the positions of the buffer and the slave plate in the image, and acquiring the possible damage fault positions. By calculating whether the relative position between the components in the image is less than a threshold value, the failure is determined to be valid when the relative position of the slave board and the buffer is less than the threshold value. And when the coincidence degree of the damaged fault area, the buffer and the slave plate area exceeds a threshold value, judging that the damaged fault identification is effective, and outputting a damaged position.
4. After confirming in the above step that no fault is detected, the slave board graph is intercepted.
The plate may be inclined and deformed after being broken due to the impact during the running of the train. The common contour algorithm with small deformation is difficult to calculate the inclination angle. And after confirming that no fault of the slave board damage class exists, acquiring the slave board position, and then intercepting the corresponding sub-graph.
5. Calculated from the plate inclination angle.
In the process of obtaining the image contour, the original Canny operator only outputs the position of the contour point and does not output the direction of the contour point, the Canny operator source code is modified, and the sobel operator is used for obtaining the gradient direction of the contour point at the same time of obtaining the position of the contour point.
Firstly, removing noise in the subgraph by using a Gaussian filtering mode, and then acquiring the position and the direction of a contour point in an image by using a modified canny operator.
The straight line in the original Hough transform image is r = x cos theta + y sin theta, the gradient direction of the point (x, y) in the way is not considered in the formula, the normal direction of the straight line is f = sin theta/cos theta, and when the point is on the straight line, the included angle between the gradient direction and the normal direction of the straight line is smaller than a threshold value. According to the principle, the positions and directions of the contour points obtained by using a canny operator can be used for reversely calculating the intervals of r and theta of the straight line where each contour point is located. After the calculation results of all the points are accumulated, a set of r and theta can be obtained. r and theta form a binary matrix, and the maximum value of each area of the matrix is calculated. And taking the point of the maximum value as the center, taking the point number threshold value on the straight line as the radius, and calculating the second moment, wherein the maximum value point of the second moment is the r and the theta of the straight line from the plate boundary. And obtaining the slope k = -cos theta/sin theta of the straight line, and judging that the slave plate has a fault when the inclination angle of the straight line is greater than a threshold value.
It should be noted that the detailed description is only for explaining and explaining the technical solution of the present invention, and the scope of protection of the claims is not limited thereby. It is intended that all such modifications and variations be included within the scope of the invention as defined in the following claims and the description.
Claims (10)
1. A rail wagon buffer and slave plate fault detection method is characterized by comprising the following steps:
the method comprises the following steps: acquiring a buffer of the rail wagon and an image of a slave plate;
step two: marking the buffer position, the slave plate position, the buffer damage position and the slave plate damage position in the railway wagon buffer and slave plate image, and constructing a training set by using the marked image, the railway wagon buffer and the slave plate image;
step three: introducing the responding-RCNN into a FasterRcnn model to obtain an improved FasterRcnn model, and training the improved FasterRcnn model by utilizing a training set to obtain a well-trained improved FasterRcnn model;
the modified FasterRcnn model specifically performs the following steps:
firstly, generating the characteristics of each label category by using a basic detection network of a FasterRcnn model, and simultaneously generating a global semantic pool;
then, introducing the position relation between the slave plates and the buffer, the position relation between the slave plates and the damaged slave plates, and the position relation between the damaged buffer and the damaged buffer as prior knowledge into the map, defining a class-to-class undirected graph, coding the prior knowledge of each edge in the undirected graph, coding all the edges in the undirected graph to obtain a knowledge map, then spreading the knowledge map in a global semantic pool, and generating reinforced features through a global reasoning module;
finally, connecting the enhanced features with features generated by a basic detection network to obtain new features;
step four: inputting an image to be recognized into a trained improved FasterRcnn model to obtain a buffer position, a slave plate position, a buffer damage position and a slave plate damage position in the image;
step five: judging whether a fault exists according to the result obtained in the fourth step;
the concrete steps of the fifth step are as follows: determining a failure based on the buffer position, the slave plate position, the buffer damage position, and the slave plate damage position in the obtained image, if at least one of them is true;
the relative position of the buffer position and the slave plate position in the image is less than a threshold;
the coincidence degree of the buffer position and the buffer damage position in the image exceeds a threshold value;
the coincidence of the slave plate position with the slave plate breakage position in the image exceeds a threshold value.
2. A method of detecting a railcar buffer and slave plate failure, according to claim 1, wherein said method further comprises the steps of:
if no fault is detected in the fifth step, executing a sixth step;
step six: cutting the slave plate graph according to the slave plate position obtained in the step four;
step seven: acquiring contour points of the slave plate graph, and acquiring the positions and gradient directions of the contour points;
step eight: and judging whether the slave plate fails according to the position and gradient direction of the contour point and by combining a Hough transform detection straight line method.
3. The method for detecting faults of buffers and slave plates of railway freight cars as claimed in claim 2, wherein said step eight comprises the following steps:
step eight one: introducing the position and the gradient direction of the contour points in a Hough transform detection straight line method, and obtaining r and theta on a straight line where each contour point is located according to the position and the gradient direction of the contour points, wherein r is the distance from an original point to the nearest point on the straight line, and theta is an included angle between an x axis and the straight line connecting the original point and the nearest point;
step eight two: constructing a two-dimensional matrix according to r and theta on the straight line where each contour point is located;
step eight and three: and calculating a secondary moment by taking the point of the maximum value in the two-dimensional matrix as a center and the number of contour points as radii, wherein the maximum point of the secondary moment is r and theta of the slave plate boundary straight line, finally obtaining the straight line slope according to theta, and judging that the slave plate has a fault when the straight line slope is greater than a threshold value.
4. A rail wagon buffer and slave plate failure detection method as claimed in claim 3, wherein the slope of the straight line is represented as:
k＝-cosθ/sinθ。
5. a rail wagon buffer and slave plate failure detection method as claimed in claim 4, wherein r is represented as:
r＝x*cosθ+y*sinθ。
6. the method as claimed in claim 5, further comprising a step of removing noise from the slave board image before the seventh step.
7. The rail wagon buffer and slave plate failure detection method of claim 6, wherein the denoising is performed by Gaussian filtering.
8. The rail wagon buffer and slave plate fault detection method as claimed in claim 7, wherein the obtaining is performed from contour points of the slave plate graph by a Canny operator.
9. A medium, characterized in that a program is provided in the medium for executing the method of claim 1 to claim 8.
10. A system, characterized in that a fault detection module is provided in the system, said fault detection module particularly performing the method according to claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210999696.8A CN115272287B (en) | 2022-08-19 | 2022-08-19 | Fault detection method, medium and system for rail wagon buffer and slave plate |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210999696.8A CN115272287B (en) | 2022-08-19 | 2022-08-19 | Fault detection method, medium and system for rail wagon buffer and slave plate |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115272287A true CN115272287A (en) | 2022-11-01 |
| CN115272287B CN115272287B (en) | 2023-04-07 |
Family
ID=83751887
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210999696.8A Active CN115272287B (en) | 2022-08-19 | 2022-08-19 | Fault detection method, medium and system for rail wagon buffer and slave plate |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115272287B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115901299A (en) * | 2023-02-15 | 2023-04-04 | 慧铁科技有限公司 | Method for analyzing and processing faults of buffer component of train coupler hook |
Citations (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1562686A (en) * | 2004-04-09 | 2005-01-12 | 北京瑞威世纪软件科技发展有限公司 | Dynamic fault detection device for bed timber type vehicle |
| CN103308056A (en) * | 2013-05-23 | 2013-09-18 | 中国科学院自动化研究所 | Road marking detection method |
| CN103523043A (en) * | 2013-10-25 | 2014-01-22 | 齐齐哈尔轨道交通装备有限责任公司大连研发中心 | Coupler buffer and shell thereof used for compression overload protection |
| CN105974904A (en) * | 2016-04-27 | 2016-09-28 | 中铁第四勘察设计院集团有限公司 | Subway vehicle safety protection system and method thereof |
| CN206719224U (en) * | 2017-04-24 | 2017-12-08 | 宝鸡比优特材料有限公司 | A kind of rail vehicles automatic car coupler buffer maintenance craft equipment |
| CN109359731A (en) * | 2018-09-27 | 2019-02-19 | 中科物栖（北京）科技有限责任公司 | A kind of Processing with Neural Network method and device based on chip design defect |
| CN109900459A (en) * | 2017-12-08 | 2019-06-18 | 嘉兴博感科技有限公司 | A kind of state monitoring method and system of rail traffic hook buffer |
| CN110211101A (en) * | 2019-05-22 | 2019-09-06 | 武汉理工大学 | A kind of rail surface defect rapid detection system and method |
| CN111080614A (en) * | 2019-12-12 | 2020-04-28 | 哈尔滨市科佳通用机电股份有限公司 | Method for identifying damage to rim and tread of railway wagon wheel |
| CN111091543A (en) * | 2019-12-12 | 2020-05-01 | 哈尔滨市科佳通用机电股份有限公司 | Railway wagon swing bolster spring loss fault target detection method |
| CN111091558A (en) * | 2019-12-12 | 2020-05-01 | 哈尔滨市科佳通用机电股份有限公司 | Railway wagon swing bolster spring jumping fault image identification method |
| CN111272405A (en) * | 2020-03-26 | 2020-06-12 | 广西电网有限责任公司电力科学研究院 | High-voltage circuit breaker mechanical fault diagnosis method and system |
| CN112132820A (en) * | 2020-09-29 | 2020-12-25 | 江苏新绿能科技有限公司 | Catenary positioning pipe stay cable fault detection method based on YOLOv2 algorithm |
| CN112989354A (en) * | 2021-01-27 | 2021-06-18 | 中标软件有限公司 | Attack detection method based on neural network and focus loss |
| CN113298787A (en) * | 2021-05-25 | 2021-08-24 | 江门职业技术学院 | Plate edge sealing defect detection method, controller and computer readable storage medium |
| CN113335951A (en) * | 2021-07-15 | 2021-09-03 | 赵忠卫 | Intelligent carton loading robot system and method |
| CN113724277A (en) * | 2021-08-27 | 2021-11-30 | 池州学院 | Power line detection method based on Radon transformation |
-
2022
- 2022-08-19 CN CN202210999696.8A patent/CN115272287B/en active Active
Patent Citations (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1562686A (en) * | 2004-04-09 | 2005-01-12 | 北京瑞威世纪软件科技发展有限公司 | Dynamic fault detection device for bed timber type vehicle |
| CN103308056A (en) * | 2013-05-23 | 2013-09-18 | 中国科学院自动化研究所 | Road marking detection method |
| CN103523043A (en) * | 2013-10-25 | 2014-01-22 | 齐齐哈尔轨道交通装备有限责任公司大连研发中心 | Coupler buffer and shell thereof used for compression overload protection |
| CN105974904A (en) * | 2016-04-27 | 2016-09-28 | 中铁第四勘察设计院集团有限公司 | Subway vehicle safety protection system and method thereof |
| CN206719224U (en) * | 2017-04-24 | 2017-12-08 | 宝鸡比优特材料有限公司 | A kind of rail vehicles automatic car coupler buffer maintenance craft equipment |
| CN109900459A (en) * | 2017-12-08 | 2019-06-18 | 嘉兴博感科技有限公司 | A kind of state monitoring method and system of rail traffic hook buffer |
| CN109359731A (en) * | 2018-09-27 | 2019-02-19 | 中科物栖（北京）科技有限责任公司 | A kind of Processing with Neural Network method and device based on chip design defect |
| CN110211101A (en) * | 2019-05-22 | 2019-09-06 | 武汉理工大学 | A kind of rail surface defect rapid detection system and method |
| CN111080614A (en) * | 2019-12-12 | 2020-04-28 | 哈尔滨市科佳通用机电股份有限公司 | Method for identifying damage to rim and tread of railway wagon wheel |
| CN111091543A (en) * | 2019-12-12 | 2020-05-01 | 哈尔滨市科佳通用机电股份有限公司 | Railway wagon swing bolster spring loss fault target detection method |
| CN111091558A (en) * | 2019-12-12 | 2020-05-01 | 哈尔滨市科佳通用机电股份有限公司 | Railway wagon swing bolster spring jumping fault image identification method |
| CN111272405A (en) * | 2020-03-26 | 2020-06-12 | 广西电网有限责任公司电力科学研究院 | High-voltage circuit breaker mechanical fault diagnosis method and system |
| CN112132820A (en) * | 2020-09-29 | 2020-12-25 | 江苏新绿能科技有限公司 | Catenary positioning pipe stay cable fault detection method based on YOLOv2 algorithm |
| CN112989354A (en) * | 2021-01-27 | 2021-06-18 | 中标软件有限公司 | Attack detection method based on neural network and focus loss |
| CN113298787A (en) * | 2021-05-25 | 2021-08-24 | 江门职业技术学院 | Plate edge sealing defect detection method, controller and computer readable storage medium |
| CN113335951A (en) * | 2021-07-15 | 2021-09-03 | 赵忠卫 | Intelligent carton loading robot system and method |
| CN113724277A (en) * | 2021-08-27 | 2021-11-30 | 池州学院 | Power line detection method based on Radon transformation |
Non-Patent Citations (5)
| Title |
|---|
| HOOGENDOORN S P等: "Lessons Learned from Field Operational Test of Integrated Network Management in Amsterdam", 《TRANSPORTATION RESEARCH RECORD》 * |
| XU H等: "Reasoning-rcnn: Unifying adaptive global reasoning into large-scale object detection", 《PROCEEDINGS OF THE IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION》 * |
| ZHANG Q等: "Vehicle-damage-detection segmentation algorithm based on improved mask RCNN", 《IEEE ACCESS》 * |
| 朱云飞等: "一种改进Faster R-CNN的车辆识别算法", 《软件导刊》 * |
| 黄东林等: "关于提高TFDS动态检查质量的探讨", 《铁道车辆》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115901299A (en) * | 2023-02-15 | 2023-04-04 | 慧铁科技有限公司 | Method for analyzing and processing faults of buffer component of train coupler hook |
| CN115901299B (en) * | 2023-02-15 | 2023-06-06 | 慧铁科技有限公司 | Method for analyzing and processing faults of train coupler buffering component |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115272287B (en) | 2023-04-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115272334B (en) | Method for detecting tiny defects on surface of steel rail under complex background | |
| CN111080598B (en) | Bolt and nut missing detection method for coupler yoke key safety crane | |
| CN111079819B (en) | Method for judging state of coupler knuckle pin of railway wagon based on image recognition and deep learning | |
| CN115272287B (en) | Fault detection method, medium and system for rail wagon buffer and slave plate | |
| CN111091558B (en) | Railway wagon swing bolster spring jumping fault image identification method | |
| RU2010131032A (en) | METHOD AND SYSTEM FOR IDENTIFICATION OF TRAIN INFORMATION | |
| CN104599249A (en) | Cable rope bearing bridge deck vehicle load distribution real-time detection method | |
| CN111091548B (en) | Railway wagon adapter dislocation fault image identification method and system based on deep learning | |
| CN105160668A (en) | Image segmentation method and system, and cell image segmentation method and system | |
| CN103020626A (en) | Method and system for detecting elliptical target in image | |
| CN113516629A (en) | Intelligent detection system for TFDS passing operation | |
| CN114399744A (en) | Vehicle type recognition method and device, electronic equipment and storage medium | |
| CN113221839B (en) | Automatic truck image identification method and system | |
| CN111079630B (en) | Fault identification method for railway wagon brake beam with incorrect installation position | |
| CN115527170A (en) | Method and system for identifying closing fault of door stopper handle of automatic freight car derailing brake device | |
| CN113553916A (en) | Orbit dangerous area obstacle detection method based on convolutional neural network | |
| CN105469099A (en) | Sparse-representation-classification-based pavement crack detection and identification method | |
| CN108226180A (en) | A kind of crack detection method for CRTS II plate-type tracks | |
| CN117011290A (en) | Defect diagnosis method, system and computer equipment | |
| CN112651343B (en) | Railway wagon brake beam breaking fault identification method based on image processing | |
| CN114743166A (en) | Method for detecting brake of railway wagon | |
| CN107985349B (en) | Method and device for realizing single hardware and multiple software and computer storage medium | |
| CN116452858B (en) | Rail wagon connecting pull rod round pin breaking fault identification method and system | |
| CN112365481B (en) | Method for detecting bolt loss of cross beam assembly based on image processing | |
| CN112949644B (en) | Fault image identification method for lower pull rod of truck |
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 |