CN111207304B - Railway tunnel leaky cable vision inspection device and product positioning detection method - Google Patents
Railway tunnel leaky cable vision inspection device and product positioning detection method Download PDFInfo
- Publication number
- CN111207304B CN111207304B CN201811397287.0A CN201811397287A CN111207304B CN 111207304 B CN111207304 B CN 111207304B CN 201811397287 A CN201811397287 A CN 201811397287A CN 111207304 B CN111207304 B CN 111207304B
- Authority
- CN
- China
- Prior art keywords
- fixture
- positioning
- module
- image
- processing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F17—STORING OR DISTRIBUTING GASES OR LIQUIDS
- F17D—PIPE-LINE SYSTEMS; PIPE-LINES
- F17D5/00—Protection or supervision of installations
- F17D5/02—Preventing, monitoring, or locating loss
Abstract
The invention provides a visual inspection device for a leaky cable of a railway tunnel, which is integrally arranged on a rail engineering vehicle and comprises a camera system, a processing system and a display system, wherein the camera system is used for acquiring imaging data of the leaky cable and a fixture thereof and sending the imaging data to the processing system; and the processing system is used for receiving the imaging data sent by the camera system, analyzing and processing the image data by a fixture defect intelligent identification processing method and a leaky cable skeleton tracking processing method based on a machine learning algorithm, and identifying and positioning the leaky cable and the fixture with the defects. Meanwhile, an identification and positioning method based on the device is also provided. The invention relates to a machine learning-based fixture processing and identifying process, which adopts proper processing methods under the modes of small samples and large samples, and meets the requirements of different use scenes. The leakage cable framework tracking and processing function is achieved, and deformation of the leakage cable can be quickly discriminated. The fixture positioning, extracting and matching based on the texture are realized, and the identification accuracy is higher.
Description
Technical Field
The invention relates to the technical field of railway communication, in particular to a visual inspection device for a railway tunnel leaky cable and a product positioning detection method.
Background
Railway tunnel leaky cable causes serious influence to the normal operation of vehicle, and the method of patrolling and examining among the prior art is as follows:
1. hebei Huaheng communication technology, Inc., tunnel cable fault patrol system and corresponding tunnel cable fault patrol method, application number 201410451867.9. The system structure of the system comprises a patrol car, a control unit (PC), an information acquisition unit (namely a visual imaging system), an illumination unit, a power supply unit and a positioning unit. The inspection vehicle comprises a vehicle head and a vehicle tail, the control unit and the power supply unit are fixedly arranged in the vehicle head, a fixed support is fixedly arranged on the vehicle tail, the information acquisition unit and the illumination unit are fixedly arranged on the fixed support, and the information acquisition unit is connected with the control unit in the vehicle head through a line. The information acquisition unit comprises a linear array camera, a laser range finder and an illumination system. The linear array camera carries out focusing and aperture control through a gear meshing mechanism according to a ranging signal of the laser range finder. And (4) carrying out mileage calculation by adopting a speed measuring sensor of the train, and positioning the position where the defect occurs. For the control output, the visual observation of workers is mentioned, the automatic identification can also be carried out, and for the automatic identification function, the comparison between the cable path outline and the model is simply described.
2. China railway science research institute, a system and a method for detecting the appearance of communication leaky cables in vehicle-mounted tunnels, and application number 201610365363.4. The system mainly comprises a high-speed identification camera, a laser range finder, an image acquisition device and a controller; the high-speed identification camera, the laser range finder and the image acquisition device are arranged at the tops of two sides of the train; the controller is connected with the high-speed identification camera, the laser range finder and the image acquisition device, after the communication leaky cable is identified by the high-speed identification camera at the top of one side of the train, the laser range finder at the side is started to measure the distance between the laser range finder and the communication leaky cable in first preset time, the focal length of the image acquisition device at the side is adjusted according to the distance measured by the laser range finder at the side, and the image acquisition device at the side is started to acquire image data of the communication leaky cable in second preset time. The focal length of the image acquisition device can be adjusted in real time according to the distance information measured by the laser range finder, and the accuracy of image acquisition is improved.
3. Beijing university of chemical industry, railway leaky cable fault vehicle-mounted detection system based on visual image, academic paper, thesis number 1001020150064. Introduction of contents: the thesis firstly adopts a JPEG image compression algorithm to compress the acquired image on the basis of researching the image compression algorithm, and selects an FPGA with large capacity and high processing speed to realize image data compression, thereby facilitating image transmission. On the basis of analyzing and researching leaky cable image fault characteristics, positioning of a cable is realized by adopting a Snake characteristic detection method, characteristics of a fixture are extracted by adopting a Haar method, and positioning of a leaky cable fixture is realized by combining with an SVM; and extracting texture features of the fixture based on an LBP method and carrying out fault identification by combining an SVM method. Aiming at the problems to be solved in a leaky cable fault vehicle-mounted detection system based on a visual image, on the basis of researching key technologies such as image acquisition speed, effective view field, illumination and automatic acquisition, a hardware overall structure and each unit module are designed according to the specific requirements of image acquisition and fault processing identification and the characteristics of a railway tunnel; designing and developing a software system and software modules for image acquisition, compression storage, image fault identification and the like. And finally, carrying the vehicle-mounted leaky cable fault detection system on a train running at a high speed for experimental verification.
The inventor finds that the following defects exist in the prior art in the process of research:
1. application No. 201410451867.9: (1) the acquisition system is used for vertically installing the linear array camera, the laser range finder, the light source and the like, and is difficult to adjust; (2) the practical environmental factors of the train on site are not considered in the aspects of appearance protection and electric communication of the system; (3) the linear array acquisition system describes the focusing and aperture adjusting functions of a micro motor, is a feed-forward type preset parameter adjusting focal length system depending on a range finder, can cause errors due to mechanism abrasion, motor loss and the like, and does not adopt a limit sensor to prevent over-travel and overload of the mechanism; (4) the speed measurement sensor is adopted for mileage calculation, no trigger mechanism is adopted to synchronize image acquisition, and deformation phenomena such as image stretching and compression can occur when the vehicle speed changes; (5) for the automatic identification aspect, no implementation scheme and method is proposed.
2. Application No. 201610365363.4: (1) the content comparison system mainly lies in the aspect of image acquisition modes; (2) the image acquisition device does not explicitly indicate whether a line scanning mode is used; (3) the acquisition mode that the high-speed identification camera is adopted to identify the leaky cable and then the laser range finder is opened has no practical significance, and the laser range finder can open an acquisition program according to the distance change to acquire images.
3. Paper number 1001020150064: (1) in the aspect of system design, in order to meet the requirement of carrying out experiments on a motor car site, a simple one-time mounting mechanism is adopted, and the requirements of engineering detection processing are difficult to meet due to different conditions for acquiring images each time; (2) the functions of automatic focusing, attitude adjustment, synchronous triggering and the like are not provided, and the image deformation cannot be controlled; (3) in the aspect of an image processing algorithm, a basic positioning, training and identifying method is described, only feasibility is verified, and 19 196 experimental pictures have misjudgment and are far away from practical distance.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a visual inspection device for a railway tunnel leaky cable and a product positioning detection method, and the technical problems to be solved are as follows: (1) the image information acquisition and processing system hardware platform meets the requirements of the actual engineering application site; (2) the method has feasible defect detection image processing.
In order to achieve the purpose, the invention is concretely realized by the following technical scheme:
a railway tunnel leaky cable vision inspection device is integrally arranged on a rail engineering vehicle and comprises a camera system arranged outside a carriage and a processing system arranged inside the carriage;
the camera system is used for acquiring image data of the leaky cable and the fixture thereof and sending the image data to the processing system;
the processing system is used for receiving the image data sent by the camera system, analyzing and processing the image data by a fixture defect intelligent identification processing method and a leaky cable skeleton tracking processing method based on a machine learning algorithm, and identifying and positioning the defective leaky cable and the fixture thereof.
Furthermore, the camera system comprises an imaging acquisition end, wherein the imaging acquisition end consists of an image acquisition module, an optical lens real-time focusing module, an imaging system pose adjusting module, an illumination light source control module and a PC host which are arranged in a closed integral box body structure;
the illumination light source control module is used for realizing reflected illumination with high optical power density on an imaging area, integrally moves along with the pose change of the imaging system pose adjusting module and is driven by a high-precision direct current driving power supply in a program control mode;
the image acquisition module is used for receiving the reflected light of the illumination light source control module and then synchronously acquiring the motion image of the target to be detected by using a position trigger mode based on the linear array camera system;
the optical lens real-time focusing module preferably adopts a zooming optical lens and is used for carrying out triple real-time closed-loop adjustment on focal length, micro-focusing and aperture according to the actual imaging distance.
The imaging system pose adjusting module is used for realizing the pitching and rotation program-controlled electric adjustment of the optical path of the imaging acquisition end by adopting an electric angular displacement table and an electric rotating table mechanism;
and the PC host is used for controlling the image acquisition module to acquire image data and carrying out data transmission communication with the PC of the processing system through the gigabit network interface.
Furthermore, a shockproof structure is arranged on a base of the imaging acquisition end.
Further, the synchronous acquisition method of the image acquisition module comprises the following steps: the rotary encoder synchronously installed with wheels of the rail engineering vehicle is subjected to pulse acquisition and subdivision through a speed acquisition circuit, and is transmitted to an image acquisition card through an RS-422 communication protocol to trigger each line of image acquisition of the linear array camera.
Further, the processing system comprises:
the leaky cable processing system is used for receiving the image data sent by the camera system, identifying and positioning leaky cable images, and identifying and positioning defective leaky cables based on a leaky cable skeleton tracking processing method;
and the fixture processing system is used for receiving the image data sent by the camera system, identifying and positioning the fixture image, analyzing and processing the image data by using the intelligent fixture defect identification processing method based on the machine learning algorithm, positioning, extracting and matching the fixture based on the texture, and identifying and positioning the defective fixture.
Further, the leaky cable processing system comprises:
the rough positioning module is used for combining kilometer sign information acquired by the train and pulse counting of a speed measuring sensor, firstly performing rough positioning on image data in a web acquisition mode, and performing frame-by-frame segmentation according to synchronously acquired image data;
the segmentation module is used for carrying out image segmentation processing on each frame of image data after frame-by-frame segmentation;
the extraction module is used for performing morphological processing on the segmented image, extracting a leaking cable framework and a central line by an image gradient direction Histogram (HOG) method, calculating the curvature of the axis of the leaking cable, and judging whether a clamp is missing or the leaking cable is excessively bent;
and the positioning module is used for tracking and analyzing the width of the communication area in the vertical direction of the leaky cable by using the leaky cable framework extracted by the extraction module and positioning the fixture at the position with the width suddenly changed.
Further, the fixture processing system comprises a fixture identification and positioning module and a fixture analysis and detection module;
the fixture identification and positioning module is used for preliminarily positioning the position of the fixture according to the image data prior information; extracting local texture features of the image by using a Local Binary Pattern (LBP) method, and positioning the clamp area based on shape matching, identifying and positioning the clamp;
and the fixture analysis and detection module is used for automatically detecting the morphological defects of the whole fixture by establishing a fixture library and a machine learning-based method after the fixture identification and positioning module identifies and positions the fixture, identifying the small sample fixture by adopting an SVM method, identifying the large sample fixture by adopting a deep learning method and positioning the position of the fixture with a problem.
Further, the method for establishing the card library and based on the machine learning comprises the following steps:
firstly, manually appointing key position characteristics of a positioned fixture image, marking whether the fixture image is qualified or not, and obtaining a standard for judging whether the fixture image is qualified or not through machine learning training by the system through providing a large number of patterns of qualified and unqualified fixtures for the system.
The invention also provides a product positioning detection method, which comprises a method for carrying out product positioning detection by using the device.
According to the visual inspection device for the leaky cable of the railway tunnel and the product positioning detection method, the fixture processing and identifying process based on machine learning is adopted, and appropriate processing methods are adopted under the small sample mode and the large sample mode, so that the requirements of different use scenes are met. The leakage cable framework tracking and processing function is achieved, and deformation of the leakage cable can be quickly discriminated. The fixture positioning, extracting and matching based on the texture are realized, and the identification accuracy is higher.
Drawings
The invention is explained in more detail below with reference to the figures and examples.
FIG. 1 is a schematic structural diagram of a visual inspection device for a leaky cable of a railway tunnel according to an embodiment of the invention;
fig. 2 is a second schematic structural diagram of the visual inspection device for the leaky cable of the railway tunnel according to the embodiment of the invention.
Detailed Description
As shown in fig. 1, an embodiment of the present invention provides a visual inspection device 100 for a leaky cable in a railway tunnel, which is integrally installed on a rail engineering vehicle, and comprises a camera system installed outside a carriage and a processing system disposed inside the carriage;
the camera system 11 is used for acquiring image data of the leaky cable and the fixture thereof and sending the image data to the processing system;
the processing system 12 is configured to receive image data sent by the camera system, analyze and process the image data by a fixture defect intelligent identification processing method and a leaky cable skeleton tracking processing method based on a machine learning algorithm, and identify and locate a defective leaky cable and a fixture thereof.
The processing system is arranged in the engineering truck carriage and mainly used as a computing processing system for analyzing and processing the acquired image data.
Furthermore, the camera system comprises an imaging acquisition end, wherein the imaging acquisition end consists of an image acquisition module, an optical lens real-time focusing module, an imaging system pose adjusting module, an illumination light source control module and a PC host which are arranged in a closed integral box body structure;
the illumination light source control module is used for realizing reflected illumination with high optical power density on an imaging area, integrally moves along with the pose change of the imaging system pose adjusting module and is driven by a high-precision direct current driving power supply in a program control mode;
the lighting light source is mainly a high-brightness LED spotlight or a laser line light source, realizes the reflection lighting of high optical power density on an imaging area, and is driven by a high-precision direct current driving power supply in a program control mode.
And the image acquisition module is used for receiving the reflected light of the illumination light source control module and then synchronously acquiring the motion image of the target to be detected by using a position trigger mode based on the linear array camera system.
As the leaky cables and the clamps in the tunnel are in a continuous distribution state, a linear array camera system is needed to acquire the motion images of the target to be detected. In order to ensure that the acquired images are not stretched or compressed due to the change of the running speed, a position trigger mode is used for synchronous acquisition, namely, a speed acquisition circuit is used for acquiring and subdividing pulses of a rotary encoder synchronously installed with train wheels, the pulses are transmitted to a Cameralink image acquisition card through an RS-422 communication protocol, and each line of image acquisition of the linear array camera is triggered.
The optical lens real-time focusing module preferably adopts a zooming optical lens and is used for carrying out triple real-time closed-loop adjustment on focal length, micro-focusing and aperture according to the actual imaging distance.
In order to meet the environmental adaptability of practical use, a zoom optical lens is adopted. The system carries out triple real-time closed-loop adjustment of focal length, micro-focusing and aperture according to the actual imaging distance, the distance value is measured by a laser range finder in real time, the motion control part rapidly adjusts the stroke of a corresponding mechanism according to the distance value, and the adjustment parameters are stored by a lookup table obtained by the system through experiments in an off-line state; the motion control card controls the focusing motor through the CAN-open bus to drive the mechanism to move so as to realize precise adjustment.
The imaging system pose adjusting module is used for realizing the pitching and rotation program-controlled electric adjustment of the optical path of the imaging acquisition end by adopting an electric angular displacement table and an electric rotating table mechanism;
in the actual test process, due to factors such as cable installation height and fixture type change, the pose of the imaging system needs to be adjusted. The system adopts a holder mechanism with two degrees of freedom to realize pitching and rotating adjustment of a light path, and simultaneously adopts an area array CCD imaging system (with automatic focusing and aperture adjustment) horizontally coplanar with an optical axis of a linear array system to continuously shoot a leaky cable detection area, and mainly realizes the following steps: real-time images are collected in the system installation or operation process to assist manual observation; in the operation process, timing inspection is carried out by adopting an image processing method, and a closed loop is controlled by combining an electric pitching mechanism, so that the leaky cable is ensured to be positioned in the center of a view field.
And the PC host is used for controlling the image acquisition module to acquire image data and carrying out data transmission communication with the PC of the processing system through the gigabit network interface.
After the PC of the camera system is subjected to initialization, pose adjustment, image acquisition, image compression storage and other steps, data are transmitted to a processing terminal PC, image preprocessing, defect detection, fault position confirmation and other operations are carried out, and finally, repairing operation of the corresponding position is carried out manually.
Furthermore, a shockproof structure is arranged on a base of the imaging acquisition end. To solve the problem of site vibration.
Further, the synchronous acquisition method of the image acquisition module comprises the following steps: the rotary encoder synchronously installed with wheels of the rail engineering vehicle is subjected to pulse acquisition and subdivision through a speed acquisition circuit, and is transmitted to an image acquisition card through an RS-422 communication protocol to trigger each line of image acquisition of the linear array camera.
Further, the processing system comprises:
the leaky cable processing system is used for receiving the image data sent by the camera system, identifying and positioning leaky cable images, and identifying and positioning defective leaky cables based on a leaky cable skeleton tracking processing method;
and the fixture processing system is used for receiving the image data sent by the camera system, identifying and positioning the fixture image, analyzing and processing the image data by using the intelligent fixture defect identification processing method based on the machine learning algorithm, positioning, extracting and matching the fixture based on the texture, and identifying and positioning the defective fixture.
Further, the leaky cable processing system comprises:
the rough positioning module is used for combining kilometer sign information acquired by the train and pulse counting of a speed measuring sensor, firstly performing rough positioning on image data in a web acquisition mode, and performing frame-by-frame segmentation according to synchronously acquired image data;
the segmentation module is used for carrying out image segmentation processing on each frame of image data after frame-by-frame segmentation;
the extraction module is used for performing morphological processing on the segmented image, extracting a leaking cable framework and a central line by an image gradient direction Histogram (HOG) method, calculating the curvature of the axis of the leaking cable, and judging whether a clamp is missing or the leaking cable is excessively bent;
and the positioning module is used for tracking and analyzing the width of the communication area in the vertical direction of the leaky cable by using the leaky cable framework extracted by the extraction module and positioning the fixture at the position with the width suddenly changed.
Further, the fixture processing system comprises a fixture identification and positioning module and a fixture analysis and detection module;
the fixture identification and positioning module is used for preliminarily positioning the position of the fixture according to the image data prior information; extracting local texture features of the image by using a Local Binary Pattern (LBP) method, and positioning the clamp area based on shape matching, identifying and positioning the clamp;
and the fixture analysis and detection module is used for automatically detecting the morphological defects of the whole fixture by establishing a fixture library and a machine learning-based method after the fixture identification and positioning module identifies and positions the fixture, identifying the small sample fixture by adopting an SVM method, identifying the large sample fixture by adopting a deep learning method and positioning the position of the fixture with a problem.
Further, the method for establishing the card library and based on the machine learning comprises the following steps:
firstly, manually appointing key position characteristics of a positioned fixture image, marking whether the fixture image is qualified or not, and obtaining a standard for judging whether the fixture image is qualified or not through machine learning training by the system through providing a large number of patterns of qualified and unqualified fixtures for the system.
The invention also provides a product positioning detection method, which comprises a method for carrying out product positioning detection by using the device.
The method is applied to the embodiment, after an engineering truck enters a tunnel, a light source arranged on the side face of a truck body is directed at a leaky cable for irradiation, a camera receives reflected light to realize acquisition and shooting, real-time conditions of the leaky cable and a fixture of the leaky cable in the whole tunnel are scanned and stored, after image acquisition is completed, a core software system is used for performing deep processing on image data, and identification, positioning and analysis processing methods suitable for corresponding characteristics are respectively adopted for the leaky cable and the fixture. In the aspect of fixture processing, after identifying and positioning the fixture, whether the shape of the fixture is normal or not and whether the part is complete or not need to be analyzed. The system uses a method of establishing a card library and based on machine learning to implement this process. Before the system is used, images of qualified and unqualified fixtures need to be collected firstly, the key positions of the fixtures need to be designated manually, a large number of training samples are provided, and the system obtains a standard for judging the quality through autonomous learning. For the feature recognition of the small sample fixture, a Machine learning algorithm based on SVM (Support Vector Machine) is used; in order to meet the requirements of large-sample, multi-type and multi-characteristic fixture identification, the machine Learning algorithm based on Deep Learning (Deep Learning) can be adopted to automatically detect the form defects of the whole leaky cable and the fixture, the position with problems is positioned, and the maintenance personnel can conveniently and timely replace and maintain the leaky cable fixture.
The image processing method of the leaky cable part is as follows:
(1) combining kilometer sign information acquired by a train and pulse counting of a speed measuring sensor, firstly performing coarse positioning on image data in a web acquisition mode, and performing frame-by-frame segmentation according to synchronously acquired image data;
(2) firstly, carrying out image segmentation processing on each frame of image data;
(3) performing morphological processing and HOG (Histogram of image Gradient direction) method on the segmented image, extracting a leaky cable skeleton and a central line, calculating the curvature of the axis of the leaky cable, and judging whether a clamp is missing or the leaky cable is excessively bent;
(4) and tracking and analyzing the width of the communication area in the vertical direction of the leaky cable by using the leaky cable framework extracted in the previous step, and positioning a fixture at the position with the width suddenly changed.
The fixture part identification and positioning method comprises the following steps:
(1) according to the prior information (image frame positioning and leakage cable width sudden change positions), the possible positions of the fixture are preliminarily positioned;
(2) local texture features of the image are extracted by an LBP (Local Binary Pattern) method and are used for fixture region positioning based on shape matching.
The fixture analysis and detection method comprises the following steps:
after the fixture is identified and positioned, whether the shape of the fixture is normal or not and whether the part is complete or not need to be analyzed. The system uses a method of establishing a card library and based on machine learning to implement this process. Before the system is used, key position features are manually appointed on the positioned fixture images, whether the fixture images are qualified or not is marked, and the system obtains a standard for judging the quality of the fixture images through machine learning training by providing a large number of patterns of qualified and unqualified fixtures for the system.
When the Machine learning processing is performed, a small-sample and nonlinear fixture feature recognition can be realized by using a Machine learning algorithm based on an SVM (Support Vector Machine). In order to meet the requirements of fixture identification of large samples, multiple types and multiple characteristics, the algorithm based on deep learning is adopted for characteristic identification, and the method has the advantages that: the depth of the model structure is emphasized, and a plurality of layers of hidden layer nodes are provided; the importance of feature learning is clearly highlighted, and the feature representation of the sample in the original space is transformed into a new feature space through layer-by-layer feature transformation, so that the classification or the prediction is easier.
According to the visual inspection device for the leaky cable of the railway tunnel and the product positioning detection method, the fixture processing and identifying process based on machine learning is adopted, the SVM method is adopted for small samples, the deep learning method is adopted for large samples, and appropriate processing methods are adopted under the small samples and the large samples, so that the requirements of different use scenes are met. Coarse positioning extraction of a leaky cable area, namely extracting a framework and a central line by using a method combining image segmentation, gradient direction histogram processing and morphology, and the leaky cable framework tracking processing function is realized, so that leaky cable deformation can be quickly discriminated. The texture features of the fixture rough positioning area are extracted by a method based on a local binary pattern, fixture positioning based on shape matching is carried out, the fixture rough positioning area has a fixture positioning, extracting and matching function based on the texture, and the identification accuracy is higher.
Finally, it should be noted that: although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that changes may be made in the embodiments and/or equivalents thereof without departing from the spirit and scope of the invention as defined in the following claims. Any modification, equivalent replacement, or improvement made within the spirit and principle of the invention should be included in the protection scope of the invention.
Claims (7)
1. A railway tunnel leaky cable vision inspection device is integrally mounted on a rail engineering truck and is characterized by comprising a camera system mounted outside a carriage and a processing system arranged inside the carriage;
the camera system is used for acquiring image data of the leaky cable and the fixture thereof and sending the image data to the processing system;
the processing system is used for receiving image data sent by the camera system, analyzing and processing the image data by a fixture defect intelligent identification processing method and a leaky cable skeleton tracking processing method based on a machine learning algorithm respectively, and identifying and positioning a defective leaky cable and a fixture thereof;
the processing system comprises:
the leaky cable processing system is used for receiving the image data sent by the camera system, identifying and positioning leaky cable images, and identifying and positioning defective leaky cables based on a leaky cable skeleton tracking processing method;
the fixture processing system is used for receiving the image data sent by the camera system, identifying and positioning fixture images, analyzing and processing the image data by using a fixture defect intelligent identification processing method based on a machine learning algorithm, positioning, extracting and matching fixtures based on textures, and identifying and positioning defective fixtures;
wherein, leaky cable processing system includes:
the rough positioning module is used for combining kilometer sign information acquired by the train and pulse counting of a speed measuring sensor, firstly performing rough positioning on image data in a web acquisition mode, and performing frame-by-frame segmentation according to synchronously acquired image data;
the segmentation module is used for carrying out image segmentation processing on each frame of image data after frame-by-frame segmentation;
the extraction module is used for performing morphological processing on the segmented image, extracting a leaking cable framework and a central line by an image gradient direction Histogram (HOG) method, calculating the curvature of the axis of the leaking cable, and judging whether a clamp is missing or the leaking cable is excessively bent;
and the positioning module is used for tracking and analyzing the width of the communication area in the vertical direction of the leaky cable by using the leaky cable framework extracted by the extraction module and positioning the fixture at the position with the width suddenly changed.
2. The device of claim 1, wherein the camera system comprises an imaging acquisition end, and the imaging acquisition end consists of an image acquisition module, an optical lens real-time focusing module, an imaging system pose adjusting module, an illumination light source control module and a PC host which are arranged in a closed integral box body structure;
the illumination light source control module is used for realizing reflected illumination with high optical power density on an imaging area, integrally moves along with the pose change of the imaging system pose adjusting module and is driven by a high-precision direct current driving power supply in a program control mode;
the image acquisition module is used for receiving the reflected light of the illumination light source control module and then synchronously acquiring the motion image of the target to be detected by using a position trigger mode based on the linear array camera system;
the optical lens real-time focusing module is used for carrying out triple real-time closed-loop adjustment on focal length, micro focusing and aperture according to the actual imaging distance;
the imaging system pose adjusting module is used for realizing the pitching and rotation program-controlled electric adjustment of the optical path of the imaging acquisition end by adopting an electric angular displacement table and an electric rotating table mechanism;
and the PC host is used for controlling the image acquisition module to acquire image data and carrying out data transmission communication with the PC of the processing system through the gigabit network interface.
3. The apparatus of claim 2, wherein the base of the imaging acquisition end is mounted with an anti-vibration structure.
4. The apparatus of claim 2, wherein the synchronous acquisition method of the image acquisition module comprises: the rotary encoder synchronously installed with wheels of the rail engineering vehicle is subjected to pulse acquisition and subdivision through a speed acquisition circuit, and is transmitted to an image acquisition card through an RS-422 communication protocol to trigger each line of image acquisition of the linear array camera.
5. The apparatus of claim 1, wherein the fixture processing system comprises a fixture identification and positioning module and a fixture analysis and detection module;
the fixture identification and positioning module is used for preliminarily positioning the position of the fixture according to the image data prior information; extracting local texture features of the image by using a Local Binary Pattern (LBP) method, and positioning the clamp area based on shape matching, identifying and positioning the clamp;
and the fixture analysis and detection module is used for automatically detecting the morphological defects of the whole fixture by establishing a fixture library and a machine learning-based method after the fixture identification and positioning module identifies and positions the fixture, identifying the small sample fixture by adopting an SVM method, identifying the large sample fixture by adopting a deep learning method and positioning the position of the fixture with a problem.
6. The apparatus of claim 5, wherein the method of establishing a card library and machine learning based comprises:
firstly, manually appointing key position characteristics of a positioned fixture image, marking whether the fixture image is qualified or not, and obtaining a standard for judging whether the fixture image is qualified or not through machine learning training by the system through providing a large number of patterns of qualified and unqualified fixtures for the system.
7. A method for product location detection, comprising performing product location detection by the apparatus of any of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811397287.0A CN111207304B (en) | 2018-11-22 | 2018-11-22 | Railway tunnel leaky cable vision inspection device and product positioning detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811397287.0A CN111207304B (en) | 2018-11-22 | 2018-11-22 | Railway tunnel leaky cable vision inspection device and product positioning detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111207304A CN111207304A (en) | 2020-05-29 |
| CN111207304B true CN111207304B (en) | 2021-09-14 |
Family
ID=70785333
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811397287.0A Active CN111207304B (en) | 2018-11-22 | 2018-11-22 | Railway tunnel leaky cable vision inspection device and product positioning detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111207304B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112683911B (en) * | 2020-11-17 | 2023-02-07 | 国网山东省电力公司济南供电公司 | Cable tunnel intelligence unmanned aerial vehicle inspection check out test set with high stability |
| CN112984388B (en) * | 2021-03-08 | 2022-12-02 | 上海海事大学 | BIM technology-based pipeline maintenance real-time monitoring system and method |
| CN113504416A (en) * | 2021-06-11 | 2021-10-15 | 中国人民解放军海军工程大学 | Leakage cable testing system and method |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101566586A (en) * | 2008-04-22 | 2009-10-28 | 宁波诺可电子科技发展有限公司 | External safety early-warning and positioning system of optical cable and optical-electrical compound cable |
| CN102866158A (en) * | 2012-07-24 | 2013-01-09 | 华北电力大学 | Detection method of power transmission and distribution cables for tunnel routing inspection |
| CN103376093A (en) * | 2012-04-28 | 2013-10-30 | 宝山钢铁股份有限公司 | Method and device for imaging non-uniform velocity moving object by TDI (Time Delayed Integration) CCD (Charge Coupled Device) |
| CN103669182A (en) * | 2013-11-15 | 2014-03-26 | 上海嘉珏实业有限公司 | Pavement crack recognition device based on camera and line laser and recognition method of pavement crack recognition device |
| CN203630050U (en) * | 2013-11-26 | 2014-06-04 | 成都术有科技有限公司 | Tunnel cable fixture detecting device |
| CN104167027A (en) * | 2014-09-05 | 2014-11-26 | 河北华恒信通信技术有限公司 | Tunnel cable fault inspection system and method |
| CN105321173A (en) * | 2015-09-23 | 2016-02-10 | 电子科技大学 | Machine vision based automatic defect detection method for train tunnel cable clamp |
| CN105866131A (en) * | 2016-05-27 | 2016-08-17 | 中国铁道科学研究院 | Vehicle-mounted appearance detection system and method for communication leaky coaxial cable in tunnel |
| CN107014352A (en) * | 2017-05-02 | 2017-08-04 | 中铁第四勘察设计院集团有限公司 | A kind of railcar and detection method for railway tunnel tunneling boring comprehensive detection |
| CN206772843U (en) * | 2017-05-17 | 2017-12-19 | 无锡精质视觉科技有限公司 | The classified detection system with reference to the defects of CCD line scan cameras and deep learning |
-
2018
- 2018-11-22 CN CN201811397287.0A patent/CN111207304B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101566586A (en) * | 2008-04-22 | 2009-10-28 | 宁波诺可电子科技发展有限公司 | External safety early-warning and positioning system of optical cable and optical-electrical compound cable |
| CN103376093A (en) * | 2012-04-28 | 2013-10-30 | 宝山钢铁股份有限公司 | Method and device for imaging non-uniform velocity moving object by TDI (Time Delayed Integration) CCD (Charge Coupled Device) |
| CN102866158A (en) * | 2012-07-24 | 2013-01-09 | 华北电力大学 | Detection method of power transmission and distribution cables for tunnel routing inspection |
| CN103669182A (en) * | 2013-11-15 | 2014-03-26 | 上海嘉珏实业有限公司 | Pavement crack recognition device based on camera and line laser and recognition method of pavement crack recognition device |
| CN203630050U (en) * | 2013-11-26 | 2014-06-04 | 成都术有科技有限公司 | Tunnel cable fixture detecting device |
| CN104167027A (en) * | 2014-09-05 | 2014-11-26 | 河北华恒信通信技术有限公司 | Tunnel cable fault inspection system and method |
| CN105321173A (en) * | 2015-09-23 | 2016-02-10 | 电子科技大学 | Machine vision based automatic defect detection method for train tunnel cable clamp |
| CN105866131A (en) * | 2016-05-27 | 2016-08-17 | 中国铁道科学研究院 | Vehicle-mounted appearance detection system and method for communication leaky coaxial cable in tunnel |
| CN107014352A (en) * | 2017-05-02 | 2017-08-04 | 中铁第四勘察设计院集团有限公司 | A kind of railcar and detection method for railway tunnel tunneling boring comprehensive detection |
| CN206772843U (en) * | 2017-05-17 | 2017-12-19 | 无锡精质视觉科技有限公司 | The classified detection system with reference to the defects of CCD line scan cameras and deep learning |
Non-Patent Citations (1)
| Title |
|---|
| 柔性印刷电路板银线缺陷在线视觉检测方法研究;魏智鹏;《中国优秀硕士学位论文全文数据库（电子期刊）》;20170815;正文第40页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111207304A (en) | 2020-05-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111207304B (en) | Railway tunnel leaky cable vision inspection device and product positioning detection method | |
| CN109271872B (en) | Device and method for judging on-off state and diagnosing fault of high-voltage isolating switch | |
| CN111016932B (en) | Track inspection vehicle and detection method of energy-saving air rail system | |
| CN202793389U (en) | On-line routing inspection device for high-speed rail overhead line system | |
| CN105675610A (en) | Online detection system for object surface texture characteristics and working principle | |
| CN109030498A (en) | The vision inspection apparatus and method of a kind of blade knife mark and surface defect | |
| CN105651782A (en) | Engine cylinder cover defect automatic detection equipment based on machine vision | |
| CN109794948A (en) | Distribution network live line work robot and recognition positioning method | |
| CN109300161A (en) | A kind of localization method and device based on binocular vision | |
| CN103646544B (en) | Based on the vehicle behavioural analysis recognition methods of The Cloud Terrace and camera apparatus | |
| CN113723325B (en) | Frock defect detecting system of prefabricated component of assembled | |
| CN111208146B (en) | Tunnel cable detection system and detection method | |
| CN113418933B (en) | Flying shooting visual imaging detection system and method for detecting large-size object | |
| CN103481910A (en) | Train part image collecting system and train part anomaly detection system | |
| CN109406526A (en) | A kind of resin lens smog defect detecting device | |
| CN113935971A (en) | Method and device for detecting surface defects of composite material | |
| CN210572007U (en) | LED support defect image capturing device and detection equipment | |
| CN204440490U (en) | A kind of Model For The Bush-axle Type Parts surface imperfection on-line measuring device based on compressed sensing | |
| CN111724445A (en) | Method and system for identifying large-view small-size identification code | |
| CN108489830A (en) | Crack propagation measuring system | |
| CN113189005A (en) | Portable surface defect integrated detection device and surface defect automatic detection method | |
| CN114529811A (en) | Rapid and automatic identification and positioning method for foreign matters in subway tunnel | |
| CN113723841B (en) | On-line detection method for tool missing in assembled prefabricated part | |
| CN114392940B (en) | Pin detection method and device for special-shaped component | |
| CN206523438U (en) | Cylinder cap defect detecting system |
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 | ||
| CP01 | Change in the name or title of a patent holder |
Address after: 102600 Beijing Daxing District Zhongguancun Science and Technology Park Daxing Biomedical Industry Base, No. 1 Building, 21 Tianrong Street Patentee after: Beijing Century Oriental Zhihui Technology Co.,Ltd. Address before: 102600 Beijing Daxing District Zhongguancun Science and Technology Park Daxing Biomedical Industry Base, No. 1 Building, 21 Tianrong Street Patentee before: BEIJING CENTURY DONGFANG COMMUNICATION EQUIPMENT Co.,Ltd. |
|
| CP01 | Change in the name or title of a patent holder |