CN116993814A - Contact net suspension part displacement detection method based on monocular camera - Google Patents
Contact net suspension part displacement detection method based on monocular camera Download PDFInfo
- Publication number
- CN116993814A CN116993814A CN202310789664.XA CN202310789664A CN116993814A CN 116993814 A CN116993814 A CN 116993814A CN 202310789664 A CN202310789664 A CN 202310789664A CN 116993814 A CN116993814 A CN 116993814A
- Authority
- CN
- China
- Prior art keywords
- contact net
- camera
- suspension
- brightness
- displacement
- 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.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 104
- 239000000725 suspension Substances 0.000 title claims abstract description 101
- 238000006073 displacement reaction Methods 0.000 title claims abstract description 43
- 238000000034 method Methods 0.000 claims abstract description 26
- 238000000605 extraction Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000012545 processing Methods 0.000 abstract description 4
- 238000012937 correction Methods 0.000 abstract description 3
- 238000004458 analytical method Methods 0.000 abstract description 2
- 238000009434 installation Methods 0.000 description 4
- 238000005259 measurement Methods 0.000 description 4
- 238000012423 maintenance Methods 0.000 description 3
- 230000014509 gene expression Effects 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000011897 real-time detection Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- 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
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/64—Three-dimensional objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Multimedia (AREA)
- Software Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Databases & Information Systems (AREA)
- Medical Informatics (AREA)
- Train Traffic Observation, Control, And Security (AREA)
Abstract
The invention discloses a contact net suspension part displacement detection method based on a monocular camera, which is characterized in that a camera is controlled by a master control module to shoot and return pictures, an exposure adjustment module, a contact net suspension part integral detection module and a contact net suspension part position detection module are controlled to run, the contact net suspension part displacement detection method is used for detecting displacement of a track traffic contact net suspension part by a picture selection model, a key point detection model, a key point correction algorithm and a dp dynamic programming algorithm, higher detection precision is realized under the condition that depth information is limited and scale information is lacking, the detected relative position information of the part is recorded in a database, and relative pictures and alarm data are stored for subsequent analysis and processing of image data returned by camera equipment under the condition that the displacement value of the part exceeds a threshold value, and the contact net suspension part is identified in the running process of a train, the positions of all connection parts in the suspension part are calculated and compared, and whether the suspension part is displaced is judged.
Description
Technical Field
The invention belongs to the technical field of rail transit vision measurement, and particularly relates to a contact net suspension part displacement detection method based on a monocular camera.
Background
The displacement detection of overhead line system suspension components is an important problem in the railway field. The conventional displacement detection method mainly depends on physical sensors (such as an accelerometer and a displacement sensor), but has the problems of complex installation, high maintenance cost and interference to railway operation.
The conventional non-contact overhead line system suspension component detection scheme comprises a monocular camera-based algorithm and a binocular camera-based algorithm, and compared with the binocular camera, the monocular camera has the following advantages: 1. the cost is lower: the use of a monocular camera for displacement detection is less costly than a binocular camera. Binocular cameras require two cameras and an additional processing unit for image processing and matching, whereas monocular cameras require only one camera. 2. The installation is more convenient: the monocular camera only needs to be provided with one camera, and the installation is simpler compared with the binocular camera. This may reduce installation time and labor costs. 3. The requirements for the scene are lower: binocular camera-based algorithms typically require a certain baseline distance between the cameras in order to perform parallax calculations and depth estimation. The algorithm based on the monocular camera does not need the limitation, so that the method is more suitable for detecting the displacement of the overhead line system suspension components with different scales and distances. 4. Wider application range: the monocular camera based algorithm can be applied to various railway environments and scenes including straight track, curved track, bridge, tunnel, etc. It has higher flexibility and adaptability. 5. The algorithm is simpler: the displacement detection algorithm based on a monocular camera is generally simpler than the algorithm of a binocular camera. This is because the binocular camera requires complex computation processes such as parallax computation and depth estimation, whereas the monocular camera can realize displacement detection by simple computation such as feature extraction and feature matching. However, the disadvantages of the monocular camera are: 1. depth information is limited: the monocular camera can only acquire two-dimensional image information, and cannot directly acquire depth information of an object. In contrast, binocular cameras can obtain three-dimensional depth information of an object through parallax computation and stereoscopic vision methods, thereby providing more accurate displacement measurements. 2. Lack of scale information: the monocular camera cannot directly obtain the scale information of the object. In displacement detection, scale information is critical to accurately measuring displacement. The binocular camera can acquire the scale information of the object through the triangulation principle due to the baseline distance, so that a more accurate displacement measurement result is provided. 3. High dependence on texture and features: monocular cameras are more dependent on texture and feature points in the image in displacement detection. If the surface texture of the overhead line system suspension component is less or the characteristic points are not obvious, the displacement detection of the monocular camera may be affected, and the accuracy may be lower.
Disclosure of Invention
The invention aims to overcome the defects of the prior art, and provides a contact net suspension part displacement detection method based on a monocular camera, which is used for detecting the displacement of a rail transit contact net suspension part, realizes higher detection precision under the condition that depth information is limited and scale information is lacking, and records the detected relative position information of the part in a database. And recording alarm information aiming at the condition that the displacement value of the component exceeds a threshold value, and storing related pictures and alarm data for subsequent analysis. And analyzing and processing image data transmitted back by the camera equipment, identifying the suspension parts of the overhead contact system in the running process of the train, calculating the positions of all the connection parts in the suspension parts, comparing, and judging whether the suspension parts are displaced or not.
The invention provides a contact net suspension part displacement detection method based on a monocular camera, which comprises the following steps of:
s1, a total control module stores configuration information, wherein the configuration information comprises camera configuration information and inertial navigation configuration information, and the total control module judges the running state of a train according to the inertial navigation configuration information and controls the operation of an exposure adjustment module, an integral detection module of a contact net suspension part and a position detection module of the contact net suspension part, and controls a camera to shoot and return pictures;
s2, the exposure adjustment module automatically adjusts the exposure of the camera, and the brightness average value of the pixels of the brightness detection area in the returned picture of the camera is analyzed in real time and judged by setting the brightness detection area, so that the exposure of the camera is adjusted to enable the brightness average value of the pixels of the returned picture of the camera to be within a preset threshold range;
s3, automatically grabbing and recording an optimal image of the suspension component by the integral detection module of the suspension component of the overhead line system;
s4, positioning the positions of all parts of the suspension component through a key point detection model by using a contact net suspension component position detection module;
s5, comparing the detected positioning information of the positions of all parts of the suspension component, judging whether displacement occurs according to the comparison result, and if so, recording information and giving an alarm.
Further, the step S1 specifically further includes: the general control module adopts json format file to store configuration information, and judges the running state of the train according to inertial navigation configuration information, when the train is detected to start, the general control module initializes and self-checks, after self-checking, the camera is controlled according to frequency to shoot and return pictures, when the train is detected to be temporarily idle, the operation is continuously maintained, and when the train is detected to be idle for a long time, the general control module is automatically closed.
Further, the step S2 specifically further includes: calculating the pixel brightness average value of the brightness detection area, if the pixel brightness average value exceeds the pixel brightness threshold range, calling an exposure adjustment interface of the camera to dynamically adjust the exposure value of the camera by the exposure adjustment module, reserving the time of 3 frames of images as exposure adjustment time by the exposure adjustment module, and then detecting and judging the pixel brightness average value of the brightness detection area again until the pixel brightness average value of the brightness detection area is within the pixel brightness threshold range.
Further, the specific method for calculating the pixel brightness average value of the brightness detection area comprises the following steps:
wherein x, y represents the coordinates of the pixels of the brightness detection area, g (x, y) is the gray value of the coordinates of the pixels of the brightness detection area, th is the gray threshold of the brightness of the pixels, result is the calculation result, and the result is compared with the threshold range of the brightness of the pixels to judge whether the exposure needs to be adjusted.
Further, the step S3 specifically further includes: and the whole detection module of the overhead line system hanging component adopts a YOLO series target detection model to detect each frame of image of the camera return image in real time, a frame counter is arranged and enters a picture selection mode after the hanging component is detected and identified, whether the image is the best image of the current hanging component is comprehensively judged by detecting the size information of the identified hanging component within the counting range of the frame counter, and the best camera return image and related parameters of the current hanging component are recorded after the counting of the frame counter reaches the threshold value of the counting range.
Further, the step S4 specifically further includes: the contact net suspension part position detection module adopts a YOLO-POSE key point detection model to run by a method of additionally opening threads, positions the positions of all parts of the suspension part of the contact net, and detects in real time.
Further, the step S4 specifically further includes: the contact net suspension part position detection module corrects the offset pixel position by a local binarization and skeleton extraction method, so that the detection precision is improved.
Further, the specific method for comparing the detected positioning information of the positions of the respective parts of the suspension member in S5 is as follows: and (3) calculating the maximum matching scheme in the two-pass line data by using a dp dynamic programming algorithm and taking the type sequence of the hanging part as a matching basis, realizing the accurate matching of the two-pass line data, judging that the hanging part is displaced when the difference value of the coordinates of the pixels of the return picture of the same hanging part in the two-pass line data is larger than a preset displacement threshold value, and recording the current hanging part image, hanging part number information and alarm information.
Further, the specific comparison of the positioning information of the detected positions of the respective parts of the suspension component in S5 further includes: the sensed data of the suspension elements using the current single pass line is compared with the sensed data of the manually created historical suspension elements.
The invention has the beneficial effects that: the scheme of the monocular camera is adopted in the scheme of the patent, so that the high-precision displacement detection of the overhead line system suspension component is realized, the calculation force requirement on equipment is greatly reduced compared with the scheme of the binocular camera, and the equipment purchasing and maintenance cost is reduced. According to the technical scheme, the inertial navigation configuration information of the main control module is utilized to automatically detect the start and stop state of the train, so that the train starts to collect the contact net data without manual intervention, and meanwhile, the high-efficiency exposure adjustment module, the contact net hanging part integral detection module and the contact net hanging part position detection module are combined, so that all-weather high-precision displacement measurement of the hanging parts is realized, and data support is provided for development of subsequent operation and maintenance work.
Flexibly adjustable configuration parameters: the general control module adopts json as a format of configuration parameters, can flexibly adjust related configuration information according to different camera types, lens types, inertial navigation, computing equipment and operating environments, and provides accurate detection data on the premise of being compatible with various equipment configurations.
The whole detection module of the overhead line system hanging parts adopts a YOLO series target detection model, achieves 95% of the detection rate of the hanging parts, and captures the best picture of each hanging part on the basis of ensuring that the hanging parts are not repeatedly detected and not missed by a picture selection algorithm.
The contact net suspension part position detection module detects pixel coordinates of each part of the suspension part by adopting a YOLO-POSE key point detection model, corrects offset pixel positions by a key point correction algorithm, and improves detection precision.
And calculating the maximum matching scheme in the two-pass data by using the type sequence of the suspension component in the data as a matching basis through a dp dynamic programming algorithm, so that the accurate matching of the two-pass line data is realized.
Drawings
For a further understanding of the nature and technical aspects of the present invention, reference should be made to the following detailed description of the invention and to the accompanying drawings, which are provided for purposes of reference only and are not intended to limit the invention.
Fig. 1 is a flow chart of a method for detecting displacement of overhead line system suspension components based on a monocular camera;
fig. 2 is a schematic diagram of a detection result of the catenary suspension component position detection module.
Detailed Description
Referring to fig. 1-2, an embodiment of the present invention provides a method for detecting displacement of a suspension member of a contact net based on a monocular camera, as shown in fig. 1, including the following steps:
s1, a total control module stores configuration information, wherein the configuration information comprises camera configuration information and inertial navigation configuration information, and the total control module judges the running state of a train according to the inertial navigation configuration information and controls the operation of an exposure adjustment module, an integral detection module of a contact net suspension part and a position detection module of the contact net suspension part, and controls a camera to shoot and return pictures;
s2, the exposure adjustment module automatically adjusts the exposure of the camera, and the brightness average value of the pixels of the brightness detection area in the returned picture of the camera is analyzed in real time and judged by setting the brightness detection area, so that the exposure of the camera is adjusted to enable the brightness average value of the pixels of the returned picture of the camera to be within a preset threshold range;
s3, automatically grabbing and recording an optimal image of the suspension component by the integral detection module of the suspension component of the overhead line system;
s4, positioning the positions of all parts of the suspension component through a key point detection model by using a contact net suspension component position detection module;
s5, comparing the detected positioning information of the positions of all parts of the suspension component, judging whether displacement occurs according to the comparison result, and if so, recording information and giving an alarm.
The general control module adopts json format files to store configuration information, judges the running state of the train according to inertial navigation configuration information, when the starting of the train is detected, the general control module automatically starts all hardware equipment, reserves starting time of 30 to 60 seconds for the hardware equipment, loads json configuration parameters, initializes other modules, checks whether the equipment operates normally, controls a camera to shoot and return pictures according to frequency after self-checking, and when the temporary non-speed of the train is detected, the equipment is not closed immediately, but operation is kept continuously, and algorithm detection is prevented from being stopped because of a stop station of the train; if no speed signal is detected for a long time, the master control module is automatically turned off, all the called devices are turned off, and the power supply of the devices is cut off.
Because the flexible contact net is hung and erected outdoors, in order to ensure that the camera can shoot images with clear imaging under various brightness conditions, the method realizes the automatic exposure adjustment function of the camera, an exposure adjustment module presets a brightness detection area, analyzes the brightness value of the area in the returned image in real time to judge whether the current ambient brightness is in a reasonable interval, calculates the pixel brightness average value of the brightness detection area, and if the current ambient brightness exceeds the pixel brightness threshold range, invokes an exposure adjustment interface of the camera to dynamically adjust the exposure value of the camera, and because the exposure adjustment of the camera has certain delay, the exposure adjustment module reserves the time of 3 frames of images as exposure adjustment time, and then detects and judges the pixel brightness average value of the brightness detection area again until the pixel brightness average value of the brightness detection area is in the pixel brightness threshold range. Therefore, the detection equipment can realize all-weather contact net detection, and meanwhile, the exposure can be quickly adjusted at the junction of the tunnel section and the outdoor section, so that the imaging quality of the camera is improved. The specific method for calculating the pixel brightness average value of the brightness detection area is as follows:
wherein x, y represents the coordinates of the pixels of the brightness detection area, g (x, y) is the gray value of the coordinates of the pixels of the brightness detection area, th is the gray threshold of the brightness of the pixels, result is the calculation result, and the result is compared with the threshold range of the brightness of the pixels to judge whether the exposure needs to be adjusted.
A suspension element is arranged at intervals outdoors. In order to make the proportion of each photographed suspension component to the image as large as possible (which is helpful for subsequent recognition) on the basis of no repeated detection and no missing detection of the suspension components, the integral detection module of the suspension components of the overhead line system realizes the function of automatically grabbing the image of the optimal suspension component. The whole detection module of the overhead line system suspension component adopts a YOLO series target detection model to detect each frame of image of the camera return picture in real time, and can identify different types of suspension components such as a forward positioning suspension, a reverse positioning suspension and the like; and after detecting and identifying the hanging component, setting a frame counter and entering a picture selection mode, comprehensively judging whether the picture is the best picture of the current hanging component by detecting the size information of the identified hanging component in the counting range of the frame counter, and recording the best camera return picture and related parameters of the current hanging component after the counting of the frame counter reaches the threshold value of the counting range. Drawing selection algorithm: in the process of high-speed running of the train, a camera shoots hanging pictures at a frequency not lower than 25HZ, simultaneously carries out YOLO-POSE key point identification on each picture, selects the best picture shot by the same suspension according to the proportion of hanging pictures and the number of detected key points in the pictures, then sets a certain threshold according to the position of an object in the suspension to carry out picture scaling and transformation, and the position of a hanging part shot by the best picture of each selected hanging part is fixed within a certain range through the steps 2.
The contact net suspension part position detection module adopts a YOLO-POSE key point detection model to run by a method of additionally opening threads, positions the positions of all parts of the suspension part of the contact net, and detects in real time. Because the detection of the YOLO-POSE key point detection model takes a long time, but the interval between trains passing through each suspension part is about 1s, the real-time detection is realized by running the key point detection flow in an additional thread opening mode; key point correction algorithm: and (3) for the obtained optimal picture of the suspension component and the position of each key point on the picture, dividing the picture with a certain size by taking each key point in the picture as the center, repositioning the center point of each connecting piece by using a filtering algorithm, an edge extraction algorithm, hough straight line, skeleton extraction and other methods, and evaluating the positioned center point to fulfill the aim of correcting the YOLO-POSE key point, so that the key point is the center point of the connecting piece as much as possible, and the relative length and angle of each connecting piece can be calculated accurately. Because the pixel position of the key point detected by the YOLO-POSE model and the actual position possibly generate certain deviation, the position detection module of the overhead line system suspension part corrects the offset pixel position by a local binarization and skeleton extraction method, so that the detection precision is improved, and particularly as shown in fig. 2, the original picture, the local binarization picture and the skeleton extraction picture returned by the camera are sequentially from left to right.
In order to determine whether the detected individual components of the suspension components are displaced, it is necessary to record and archive each suspension component data on the track line. The algorithm supports the comparison of the current single-pass suspension component detection data with the manually created historical data, and also supports the comparison between suspension component detection data of certain two passes. Because the detection data are affected by factors such as weather, illumination intensity and train speed, the number of the suspension components detected by each time of data is different, aiming at the problem, a dp dynamic programming algorithm is adopted, the type sequence of the suspension components is used as a matching basis, the maximum matching scheme in the two times of line data is calculated, the accurate matching of the two times of data is realized, when the difference value of the coordinates of the pixels of the return picture of the same suspension component in the two times of line data is larger than a preset displacement threshold value, the suspension component is judged to displace, and the current suspension component image, suspension component number information and alarm information are recorded. dp dynamic programming algorithm: and for the type identified by each suspension component, after confidence level screening, classifying data between stations and establishing the data as a sequence, comparing actual sequence data with the established standard inter-station suspension component type data sequence, acquiring the optimal matching position of the actual sequence and the confidence level of matching through a matching algorithm, and obtaining the most reliable matching of the actual sequence after confidence level threshold screening again, thereby obtaining the standard data of all suspension components between the stations so as to realize the difference comparison of the standard data and the actual data.
Any embodiment of the invention can be used as an independent technical scheme or can be combined with other embodiments. All patents and publications mentioned in the specification are indicative of those of ordinary skill in the art to which this invention pertains and which may be applied. All patents and publications cited herein are hereby incorporated by reference to the same extent as if each individual publication were specifically and individually indicated to be incorporated by reference. The invention may be practiced without any element or elements, limitation or limitations, which are not expressly described herein. The terms and expressions which have been employed herein are used as terms of description and not of limitation, and there is no intention in the use of such terms and expressions of excluding any equivalents of the features shown and described, but it is recognized that various modifications are possible within the scope of the invention and of the claims. It is to be understood that the embodiments described herein are illustrative of the embodiments and features disclosed herein and that modifications and variations may be resorted to by those skilled in the art without departing from the spirit of this invention as those aspects are considered to fall within the scope of this invention as defined by the independent claims and the appended claims.
Claims (9)
1. The contact net suspension part displacement detection method based on the monocular camera is characterized by comprising the following steps of:
s1, a total control module stores configuration information, wherein the configuration information comprises camera configuration information and inertial navigation configuration information, and the total control module judges the running state of a train according to the inertial navigation configuration information and controls the operation of an exposure adjustment module, an integral detection module of a contact net suspension part and a position detection module of the contact net suspension part, and controls a camera to shoot and return pictures;
s2, the exposure adjustment module automatically adjusts the exposure of the camera, and the brightness average value of the pixels of the brightness detection area in the returned picture of the camera is analyzed in real time and judged by setting the brightness detection area, so that the exposure of the camera is adjusted to enable the brightness average value of the pixels of the returned picture of the camera to be within a preset threshold range;
s3, automatically grabbing and recording an optimal image of the suspension component by the integral detection module of the suspension component of the overhead line system;
s4, positioning the positions of all parts of the suspension component through a key point detection model by using a contact net suspension component position detection module;
s5, comparing the detected positioning information of the positions of all parts of the suspension component, judging whether displacement occurs according to the comparison result, and if so, recording information and giving an alarm.
2. The method for detecting displacement of a suspension member of a contact net based on a monocular camera according to claim 1, wherein the step S1 specifically further comprises: the general control module adopts json format file to store configuration information, and judges the running state of the train according to inertial navigation configuration information, when the train is detected to start, the general control module initializes and self-checks, after self-checking, the camera is controlled according to frequency to shoot and return pictures, when the train is detected to be temporarily idle, the operation is continuously maintained, and when the train is detected to be idle for a long time, the general control module is automatically closed. .
3. The method for detecting displacement of a suspension member of a contact net based on a monocular camera according to claim 1, wherein the step S2 specifically further comprises: calculating the pixel brightness average value of the brightness detection area, if the pixel brightness average value exceeds the pixel brightness threshold range, calling an exposure adjustment interface of the camera to dynamically adjust the exposure value of the camera by the exposure adjustment module, reserving the time of 3 frames of images as exposure adjustment time by the exposure adjustment module, and then detecting and judging the pixel brightness average value of the brightness detection area again until the pixel brightness average value of the brightness detection area is within the pixel brightness threshold range.
4. A method for detecting displacement of a suspension member of a contact net based on a monocular camera according to claim 3, wherein the specific method for calculating the pixel brightness average value of the brightness detection area is as follows:
wherein x, y represents the coordinates of the pixels of the brightness detection area, g (x, y) is the gray value of the coordinates of the pixels of the brightness detection area, th is the gray threshold of the brightness of the pixels, result is the calculation result, and the result is compared with the threshold range of the brightness of the pixels to judge whether the exposure needs to be adjusted.
5. The method for detecting displacement of a suspension member of a contact net based on a monocular camera according to claim 1, wherein the step S3 specifically further comprises: and the whole detection module of the overhead line system hanging component adopts a YOLO series target detection model to detect each frame of image of the camera return image in real time, a frame counter is arranged and enters a picture selection mode after the hanging component is detected and identified, whether the image is the best image of the current hanging component is comprehensively judged by detecting the size information of the identified hanging component within the counting range of the frame counter, and the best camera return image and related parameters of the current hanging component are recorded after the counting of the frame counter reaches the threshold value of the counting range.
6. The method for detecting displacement of a suspension member of a contact net based on a monocular camera according to claim 1, wherein the step S4 specifically further comprises: the contact net suspension part position detection module adopts a YOLO-POSE key point detection model to run by a method of additionally opening threads, positions the positions of all parts of the suspension part of the contact net, and detects in real time.
7. The method for detecting displacement of a suspension member of a contact net based on a monocular camera according to claim 1, wherein the step S4 specifically further comprises: the contact net suspension part position detection module corrects the offset pixel position by a local binarization and skeleton extraction method, so that the detection precision is improved.
8. The method for detecting the displacement of the suspension member of the overhead line system based on the monocular camera according to claim 1, wherein the specific method for comparing the positioning information of the positions of the detected respective parts of the suspension member in S5 is as follows: and (3) calculating the maximum matching scheme in the two-pass line data by using a dp dynamic programming algorithm and taking the type sequence of the hanging part as a matching basis, realizing the accurate matching of the two-pass line data, judging that the hanging part is displaced when the difference value of the coordinates of the pixels of the return picture of the same hanging part in the two-pass line data is larger than a preset displacement threshold value, and recording the current hanging part image, hanging part number information and alarm information.
9. The method for detecting displacement of a suspension member of a catenary based on a monocular camera according to claim 1, wherein the comparing the positioning information of the positions of the detected respective parts of the suspension member in S5 further specifically comprises: the sensed data of the suspension elements using the current single pass line is compared with the sensed data of the manually created historical suspension elements.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310789664.XA CN116993814A (en) | 2023-06-29 | 2023-06-29 | Contact net suspension part displacement detection method based on monocular camera |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310789664.XA CN116993814A (en) | 2023-06-29 | 2023-06-29 | Contact net suspension part displacement detection method based on monocular camera |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116993814A true CN116993814A (en) | 2023-11-03 |
Family
ID=88520407
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310789664.XA Pending CN116993814A (en) | 2023-06-29 | 2023-06-29 | Contact net suspension part displacement detection method based on monocular camera |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116993814A (en) |
-
2023
- 2023-06-29 CN CN202310789664.XA patent/CN116993814A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2017080102A1 (en) | Flying device, flying control system and method | |
| CN108470356B (en) | Target object rapid ranging method based on binocular vision | |
| EP2426642B1 (en) | Method, device and system for motion detection | |
| KR101411668B1 (en) | A calibration apparatus, a distance measurement system, a calibration method, and a computer readable medium recording a calibration program | |
| CN105674880B (en) | Contact net geometric parameter measurement method and system based on binocular principle | |
| CN105346706A (en) | Flight device, and flight control system and method | |
| KR101095529B1 (en) | A traffic data collecting system in high speed using image processing skill and a traffic data collecting method using the same | |
| CN103454285A (en) | Transmission chain quality detection system based on machine vision | |
| JP2003083742A (en) | Distance correction apparatus and method of monitoring system | |
| CN108174111B (en) | Crusing robot target image grasping means | |
| KR20160100788A (en) | Apparatus and method for measuring speed of moving object | |
| KR101874427B1 (en) | licence plate recognizing method | |
| CN111723778B (en) | Vehicle distance measuring system and method based on MobileNet-SSD | |
| CN107016353A (en) | A kind of method and system of variable resolution target detection and identification integration | |
| CN116989881A (en) | Contact line vibration measurement system and method based on line laser target identification | |
| KR102062579B1 (en) | Vehicle license-plate recognition system that recognition of Vehicle license-plate damaged by shadow and light reflection through the correction | |
| CN115546738A (en) | Rail foreign matter detection method | |
| JPH1144533A (en) | Preceding vehicle detector | |
| CN112419364A (en) | Target tracking method and system based on image feature matching | |
| CN116993814A (en) | Contact net suspension part displacement detection method based on monocular camera | |
| CN112508995A (en) | Coal flow real-time dynamic metering method based on TOF camera | |
| JP2002367077A (en) | Device and method for deciding traffic congestion | |
| JP3879874B2 (en) | Logistics measuring device | |
| CN110322508A (en) | A kind of assisted location method based on computer vision | |
| KR20110077465A (en) | The apparatus and method of moving object tracking with shadow removal moudule in camera position and time |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication |