CN119379973A - A train pantograph detection system and method based on artificial intelligence - Google Patents
A train pantograph detection system and method based on artificial intelligence Download PDFInfo
- Publication number
- CN119379973A CN119379973A CN202411335647.XA CN202411335647A CN119379973A CN 119379973 A CN119379973 A CN 119379973A CN 202411335647 A CN202411335647 A CN 202411335647A CN 119379973 A CN119379973 A CN 119379973A
- Authority
- CN
- China
- Prior art keywords
- pantograph
- train
- analysis
- target object
- analysis result
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
- G06V10/225—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition based on a marking or identifier characterising the area
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8854—Grading and classifying of flaws
- G01N2021/888—Marking defects
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8883—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges involving the calculation of gauges, generating models
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Medical Informatics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Pathology (AREA)
- Chemical & Material Sciences (AREA)
- Biochemistry (AREA)
- Immunology (AREA)
- Analytical Chemistry (AREA)
- Life Sciences & Earth Sciences (AREA)
- Signal Processing (AREA)
- Train Traffic Observation, Control, And Security (AREA)
Abstract
The invention provides a system and a method for detecting a pantograph of a train based on artificial intelligence, wherein a plurality of camera devices are arranged in a train detection room in a scattered manner and are used for collecting image data of a train identification part and a pantograph part on the train entering the train detection room, an edge analysis device is connected with the camera devices and is used for processing the image data based on a pre-trained image analysis model, identifying a train identification, obtaining a pantograph analysis result and sending the result to a cloud platform, and the cloud platform is connected with the edge analysis device and is used for storing the pantograph analysis result according to the train identification classification and generating alarm information when the pantograph is abnormal in the pantograph analysis result. The detection efficiency is improved, the subjectivity of manual detection is reduced, the accuracy and the reliability of detection results are improved, the standardization of the detection flow is realized, and the situations of manual misjudgment, missing detection and the like are reduced.
Description
Technical Field
The invention relates to the technical field of train pantograph and fault detection, in particular to a train pantograph detection system and method based on artificial intelligence.
Background
The accuracy and efficiency of the detection of locomotive pantographs, as a key component to ensure the safety of power supply and locomotive operation, is critical to the safety of the entire railway system. However, the conventional manual detection method has the following problems:
(1) The manual detection has many links, cannot meet the requirement of mass detection, has low efficiency, consumes time and labor, has high labor intensity, and is easy to cause missed detection and misjudgment;
(2) The detection accuracy is influenced by experience and working state of detection personnel, the detection result is deviated from subjective judgment of personnel, and consistency and reliability are difficult to ensure;
(3) Manual detection often has incomplete recording, and data recording and history tracing are difficult;
(4) The detection standards of all staff are not uniform, and scientific and standardized detection means are lacked.
Disclosure of Invention
Based on the problems, the invention provides a system and a method for detecting a train pantograph based on artificial intelligence, and aims to solve the technical problems that in the prior art, the artificial detection of the pantograph is easy to miss judgment, misjudgment and the like.
A train pantograph detection system based on artificial intelligence, comprising:
the plurality of camera devices are arranged in the train detection room in a scattered way and are used for collecting image data of a train identification part and a pantograph part on a train entering the train detection room;
The edge analysis device is connected with the camera device and is used for processing image data based on a pre-trained image analysis model, identifying a train identifier, obtaining a pantograph analysis result and sending the pantograph analysis result to the cloud platform;
cloud platform, connect edge analysis device for:
Classifying and storing pantograph analysis results according to train identification;
when the pantograph analysis result shows that the pantograph is abnormal, alarm information is generated.
Further, the edge analysis device includes:
The storage server is used for storing the image data acquired by the image pickup device;
The edge analysis device is connected with the storage server and is used for processing the image data stored in the storage server based on a pre-trained image analysis model, identifying a train identifier, obtaining a pantograph analysis result and sending the pantograph analysis result to the cloud platform.
Further, the edge analysis device includes:
The target detection module is used for identifying a target object in the image data based on the target identification model and segmenting an image of the target object;
the anomaly analysis module is connected with the target detection module and is used for carrying out anomaly analysis on the target object based on an anomaly analysis model corresponding to the target object, and carrying out anomaly labeling in an image of the target object when the anomaly of the target object is analyzed, so as to obtain an analysis result of the target object;
and the result sending module is connected with the abnormality analysis module and is used for integrating analysis results of all target objects on the pantograph into a pantograph analysis result and sending the pantograph analysis result to the cloud platform.
Further, the cloud platform includes:
the cloud storage module is used for storing analysis results of all target objects of the pantograph according to train identification classification and time sequence;
The alarm generation module is connected with the cloud storage module and is used for generating alarm information aiming at the analysis result of the target object with the abnormality, wherein the alarm information comprises an abnormality labeling image of the target object with the abnormality;
and the display module is connected with the alarm generation module and used for displaying alarm information.
Further, the cloud platform includes:
The cloud storage module is used for storing analysis results of all the target objects of the pantograph according to the train identification classification and the time sequence to form an analysis result time sequence of all the target objects;
The risk prediction module is connected with the cloud storage module and is used for performing risk prediction on the analysis result time sequence of each target object by using a pre-trained risk prediction model to obtain abnormal risk probability of each target object, and obtaining the abnormal risk probability of the pantograph based on the abnormal risk probability of each target object;
The display module is connected with the risk prediction module and is used for the abnormal risk probability of each target object and the abnormal risk probability of the pantograph.
The utility model provides a train pantograph detecting method based on artificial intelligence, uses aforesaid train pantograph detecting system based on artificial intelligence, includes:
Step A1, a camera device collects image data of a train identification part and a pantograph part on a train entering a train detection room;
Step A2, the edge analysis device processes image data based on a pre-trained image analysis model, identifies a train identifier, obtains a pantograph analysis result and sends the pantograph analysis result to the cloud platform;
and A3, the cloud platform stores the pantograph analysis results according to the train identification classification, and generates alarm information when the pantograph abnormality exists in the pantograph analysis results.
Further, the edge analysis device includes a storage server and an edge analysis device, and step A2 includes:
Step A21, a storage server stores image data acquired by an image pickup device;
And step A22, the edge analysis equipment processes the image data stored in the storage server based on the pre-trained image analysis model, recognizes the train identification, obtains a pantograph analysis result and sends the pantograph analysis result to the cloud platform.
Further, step a22 includes:
Step A221, identifying a target object in the image data based on the target identification model and segmenting out an image of the target object;
Step A222, carrying out anomaly analysis on the target object based on an anomaly analysis model corresponding to the target object, and carrying out anomaly labeling in an image of the target object when the target object is analyzed to be anomalous, so as to obtain an analysis result of the target object;
and step A223, combining analysis results of all target objects on the pantograph into a pantograph analysis result, and sending the pantograph analysis result to the cloud platform.
Further, step A3 includes:
Step A31, storing analysis results of all target objects of the pantograph according to train identification classification and time sequence;
Step A32, generating alarm information aiming at the analysis result of the target object with the abnormality, wherein the alarm information comprises an abnormality labeling image of the target object with the abnormality;
and step A33, displaying alarm information.
Further, in step a31, the analysis results of the objects of the pantograph are stored according to the train identification classification and the time sequence, so as to form an analysis result time sequence of each object;
Step a31 is followed by:
Step B1, performing risk prediction on an analysis result time sequence of each target object by using a pre-trained risk prediction model to obtain abnormal risk probability of each target object, and obtaining the abnormal risk probability of the pantograph based on the abnormal risk probability of each target object;
in step a33, the abnormal risk probability of each target object and the abnormal risk probability of the pantograph are displayed.
The invention has the beneficial technical effects that the pantograph fault detection is carried out based on the artificial intelligence algorithm to replace manual detection, so that the detection efficiency is improved, the subjectivity of the manual detection is reduced, the accuracy and the reliability of the detection result are improved, the standardization of the detection flow is realized, and the situations of manual misjudgment, missed detection and the like are reduced.
Drawings
FIGS. 1-3 are schematic block diagrams of a train pantograph detection system based on artificial intelligence in accordance with the present invention;
FIGS. 4-8 are flow charts of steps of a method for detecting a pantograph of a train based on artificial intelligence in accordance with the present invention;
fig. 9 is a diagram of the installation position of the camera device of the train pantograph detection system based on artificial intelligence.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
It should be noted that, without conflict, the embodiments of the present invention and features of the embodiments may be combined with each other.
The invention is further described below with reference to the drawings and specific examples, which are not intended to be limiting.
Referring to fig. 1, the present invention provides an artificial intelligence based train pantograph detection system, comprising:
The plurality of camera devices (1) are arranged in a scattered manner in the train detection room and are used for collecting image data of a train identification part and a pantograph part on a train entering the train detection room;
The edge analysis device (2) is connected with the camera device (1) and is used for processing image data based on a pre-trained image analysis model, identifying a train identifier, obtaining a pantograph analysis result and sending the pantograph analysis result to the cloud platform;
cloud platform (3), connect edge analysis device (2) for:
Classifying and storing pantograph analysis results according to train identification;
when the pantograph analysis result shows that the pantograph is abnormal, alarm information is generated.
The implementation of the invention can obviously improve the efficiency and accuracy of the detection of the pantograph of the train, and can effectively and accurately judge the train number and the pantograph when the train passes through a train detection greenhouse (train detection room) at the speed of 3-5km per hour, thereby meeting the identification requirements of the train number and the pantograph of a dynamic detection system of the pantograph.
The detection time is greatly shortened by automatic image acquisition and advanced image recognition technology, so that the production efficiency is improved. The deep learning algorithm adopted by the system reduces the interference of artificial factors and improves the reliability of detection results. In addition, the unified detection standard and the data-driven decision flow realize the standardization of the detection flow and optimize the overhaul flow. Through the functions of real-time data recording and history tracing, the data management capability is enhanced. The automation of the system reduces the labor cost, reduces the economic loss caused by detection errors, and provides powerful guarantee for the safety and reliability of railway transportation.
Referring to fig. 2, further, the edge analysis device includes:
A storage server (21) for storing image data acquired by the image pickup device;
and the edge analysis device (22) is connected with the storage server (21) and is used for processing the image data stored in the storage server based on the pre-trained image analysis model, identifying the train identification, obtaining the pantograph analysis result and sending the pantograph analysis result to the cloud platform.
The system adopts a layered design, and consists of three main parts, namely a data acquisition layer, an edge analysis layer and a cloud service layer, so that the high efficiency and the order of the detection flow are ensured.
And the data acquisition layer is provided with a plurality of camera devices in the train detection room and is responsible for capturing the image data of the train number (train identification) and the pantograph thereof in real time and transmitting the data to the edge analysis layer in real time.
The imaging device is, for example, a network high-definition camera. May be an infrared camera or a visible camera.
The plurality of camera devices are arranged in the train detection room and can be respectively arranged on three sections of the front section, the middle section and the rear section of the train detection room, and the arrangement positions of the camera devices of the three sections are the same.
Referring to fig. 9, specifically, on each section, two imaging dedicated lines are respectively disposed on two inclined surfaces of a ceiling in a train detection room, one for each of the left side wall and the right side wall, 4 for each of the left side wall and the right side wall, two imaging devices of the ceiling monitor the upper portion of the pantograph, and imaging devices of the left side wall and the right side wall monitor the side surface and the bottom of the pantograph. It can be seen that 4 front and middle three sections are arranged on each section, 12 camera devices are arranged in total between train detection rooms,
In order to ensure that the camera device can clearly shoot a train mark (namely a train number) and a pantograph part in motion, a sufficient light source is needed in a train detection room, so that the camera device is deployed, a supplementary light source is needed to be arranged in the detection room, the light source is arranged beside each camera, the illumination of the top position of the train in the train detection room is ensured to be more than 300lx, and particularly, under the condition of insufficient light at night, each camera device can clearly generate a shooting image.
Each position camera device needs to be subjected to angle adjustment, and focal length, angle and the like are adjusted, so that different position camera devices have clear detection key points, such as a top camera device shoots a car number, a charcoal sliding plate, a side camera device shoots screws at each pantograph position and the like. The shooting and imaging of the camera device requires that the shooting target is clear in a train motion state, the focus is clear, and high-definition pictures with positive target positions and high picture quality can be obtained from continuous shooting video streams.
The visual AI edge analysis device is responsible for carrying out depth calculation and analysis on the image from the data acquisition layer, and then transmitting an analysis result to the cloud service layer.
The cloud service layer is used for deploying a cloud platform for evaluating, monitoring and early warning the damage quality of the pantograph in a control room, and the cloud platform is responsible for receiving the analysis result sent by the edge analysis layer and executing key functions such as abnormal alarm, notification of responsible personnel, risk prediction, model update, strategy issuing, false alarm library maintenance, configuration management of edge equipment and the like.
Referring to fig. 2, further, the edge analysis device (22) includes:
a target detection module (221) for identifying a target object in the image data based on the target identification model and segmenting an image of the target object;
The anomaly analysis module (222) is connected with the target detection module (221) and is used for carrying out anomaly analysis on the target object based on an anomaly analysis model corresponding to the target object, and carrying out anomaly labeling in an image of the target object when the target object is analyzed to be abnormal, so as to obtain an analysis result of the target object;
And the result sending module (223) is connected with the abnormality analysis module (222) and is used for combining the analysis results of all the target objects on the pantograph into a pantograph analysis result and sending the pantograph analysis result to the cloud platform.
The edge analysis device (22) uses deep learning techniques to identify specific objects in the image and to determine their locations, and the detected specific objects are classified to identify which predefined class they belong to. Such as whether the carbon sled is a bolt area or a wire rope area. And carrying out abnormality analysis according to the specific type of the identification book, wherein the type is a carbon slide plate, and analyzing whether the carbon slide plate is abnormal such as cracks, damages and the like, for example, whether a bolt area is abnormal such as falling of a bolt, and whether a wire rope area is abnormal such as strand breakage of a wire rope.
Specifically, the target object is a target detection object of the pantograph, such as a carbon slide plate, a wire rope, a bolt fastening area, and the like. The object detection module (221) is also used for identifying the train identification.
The cloud platform is also used for:
acquiring images of a plurality of pantographs as training samples;
firstly, constructing an image analysis model;
Training the constructed image analysis model by using a training sample so as to initially establish the image analysis model.
The image analysis model (comprising a target recognition model and an anomaly analysis model) is combined with basic models such as image classification, object detection, image comparison, image segmentation and the like to perform primary construction. And collecting a large number of positive and negative training samples, and utilizing a characteristic extraction technology to obtain the characteristic of abnormal phenomenon of the monitoring target object from the samples so as to train the model. The training purpose is to perform secondary development and optimization of a model algorithm according to the characteristics of the monitored target object, and perform a large amount of operation and training of the model through a training sample set, so that parameters are continuously adjusted to improve the recognition accuracy. The main parameters of the adjustment include an estimated amount, a minimum sample division number, a learning rate, a loss amount, a maximum depth, an iteration number, a characteristic score, a threshold value, a type, a range, a weight, a bias and the like. And continuously iterating until the model reaches allowable accuracy and precision in the training sample set range.
After the model is initially established, the accuracy and precision of the visual AI safety detection model are tested by utilizing the test sample sets. Once the deviation is found, optimization and parameter adjustment of the model algorithm are performed until optimal performance is achieved within the range of the test sample set.
And in the pre-online stage, loading the trained image analysis model onto edge analysis equipment for online testing. Meanwhile, a false alarm library is established, and samples of false alarms, false alarms and deviation are collected. And the model is adjusted and optimized on the cloud platform, so that the problems of false alarm, missing report and deviation are solved. And carrying out online upgrade and update on the model after the adjustment and optimization, and loading the model to the edge analysis equipment again for online operation test. Through repeated iteration, the model can be continuously optimized according to the actual production condition.
Referring to fig. 3, further, the cloud platform (3) includes:
The cloud storage module (31) stores analysis results of all target objects of the pantograph according to train identification classification and time sequence;
The alarm generation module (32) is connected with the cloud storage module (31) and is used for generating alarm information aiming at the analysis result of the target object with the abnormality, wherein the alarm information comprises an abnormality labeling image of the target object with the abnormality;
and the display module (33) is connected with the alarm generation module (32) and is used for displaying alarm information.
The display module (33) is also connected with the cloud storage module (31), so that not only the alarm information is displayed, but also the analysis result of each target object of the pantograph is displayed, the analysis result of each target object comprises images of each target object, and if the analysis result is abnormal, the displayed images are images of the target objects which are marked with the abnormality. The abnormal marking can be to outline the abnormal area by using a picture frame, for example, when the carbon slide plate is cracked or damaged, the cracked area or the damaged area is marked, for example, when the bolt is fallen, the bolt is marked as the fallen area.
As an embodiment of abnormality detection of the present invention, for example, a target detection module (221) of an edge analysis device performs preliminary classification on an image acquired by an imaging device to obtain a target object of a carbon slide plate, then enters a carbon slide plate breakage detection branch, an abnormality analysis module (222) locates the carbon slide plate, and uses a corresponding abnormality analysis model (a target detection algorithm) to identify a breakage condition on a carbon slide plate component, if a breakage position is detected and located on the carbon slide plate, the breakage position is marked, a cloud platform takes a breakage mark picture as an alarm picture, forms alarm information and pushes the alarm information to related personnel and a display module. The carbon slide plate is a part which is easy to damage in the pantograph part, is easy to crack and damage, and is easy to clearly shoot from a top camera device.
As another embodiment of abnormality detection of the present invention, for example, a target detection module (221) of an edge analysis device performs preliminary classification on an image acquired by an imaging device to obtain a target object, which is a bolt fastening region of a pantograph, then enters a bolt fastening region detection branch, an abnormality analysis module (222) locates a position of the bolt fastening region, first detects a bolt and a nut located at the position of the bolt fastening region of the pantograph with an abnormality analysis model (a machine vision image detection algorithm), then performs comparison and judgment according to preset position information, and judges whether the number and the position of the bolts preset in the bolt fastening region are met, and if the number of the bolts is not matched, it is judged that the bolts fall off or are missing. There may be a plurality of bolt fastening areas depending on the connection member.
The pantograph bolts and nuts are key points for connecting all parts, the number of the pantograph bolts and nuts is large, the number of the pantograph bolts and nuts is small, the pantograph bolts and nuts are easy to be shielded and shot in an unclear mode, the falling and loosening of screws are difficult in the pantograph detection item of the whole running train, and the pantograph bolts and nuts are difficult to cover the whole running train. Based on the actual situation, firstly, bolts and nuts of key parts which can be clearly shot at the installation position of the camera are required to be ensured, and the key is arranged. Instead of pursuing the wide-angle shooting bolt and nut positions, the key positions are not detected, the shot image effect is poor, the detection requirement is not met, and the detection is not lost.
Further, the cloud platform (3) includes:
the cloud storage module (31) stores analysis results of all the target objects of the pantograph according to the train identification classification and the time sequence to form an analysis result time sequence of all the target objects;
The risk prediction module (34) is connected with the cloud storage module (31) and is used for performing risk prediction on the analysis result time sequence of each target object by using a pre-trained risk prediction model to obtain abnormal risk probability of each target object and obtaining the abnormal risk probability of the pantograph based on the abnormal risk probability of each target object;
The display module (33) is connected with the risk prediction module (34) and is used for the abnormal risk probability of each target object and the abnormal risk probability of the pantograph.
Preventive maintenance and risk management functions of the system are helpful for finding and solving potential problems in time, and guaranteeing train operation safety.
The cloud platform is mainly used for forming alarm information, carrying out risk analysis prediction on one hand, constructing and training an image analysis model, updating and maintaining the image analysis model, and configuring and managing edge equipment on the other hand.
Referring to fig. 4, the invention further provides a method for detecting a train pantograph based on artificial intelligence, which uses the system for detecting a train pantograph based on artificial intelligence, comprising:
Step A1, a camera device collects image data of a train identification part and a pantograph part on a train entering a train detection room;
Step A2, the edge analysis device processes image data based on a pre-trained image analysis model, identifies a train identifier, obtains a pantograph analysis result and sends the pantograph analysis result to the cloud platform;
and A3, the cloud platform stores the pantograph analysis results according to the train identification classification, and generates alarm information when the pantograph abnormality exists in the pantograph analysis results.
The detection time is greatly shortened by automatic image acquisition and advanced image recognition technology, so that the production efficiency is improved. The deep learning algorithm adopted by the system reduces the interference of artificial factors and improves the reliability of detection results. In addition, the unified detection standard and the data-driven decision flow realize the standardization of the detection flow and optimize the overhaul flow. Through the functions of real-time data recording and history tracing, the data management capability is enhanced. The automation of the system reduces the labor cost, reduces the economic loss caused by detection errors, and provides powerful guarantee for the safety and reliability of railway transportation.
Specifically, the imaging device is, for example, a network high-definition camera. May be an infrared camera or a visible camera.
Specifically, the plurality of camera devices are deployed in the train detection room and can be deployed on three sections of the front section, the middle section and the rear section of the train detection room respectively, and the deployment positions of the camera devices of the three sections are the same.
Specifically, on each section, two ceiling slopes in the train detection room are dedicated to shooting, one for each of the left side wall and the right side wall, and 4 for each of the left side wall and the right side wall are arranged.
Specifically, a light source is installed beside each camera.
Referring to fig. 5, further, the edge analysis apparatus includes a storage server and an edge analysis device, and step A2 includes:
Step A21, a storage server stores image data acquired by an image pickup device;
And step A22, the edge analysis equipment processes the image data stored in the storage server based on the pre-trained image analysis model, recognizes the train identification, obtains a pantograph analysis result and sends the pantograph analysis result to the cloud platform.
Referring to fig. 6, further, step a22 includes:
Step A221, identifying a target object in the image data based on the target identification model and segmenting out an image of the target object;
Step A222, carrying out anomaly analysis on the target object based on an anomaly analysis model corresponding to the target object, and carrying out anomaly labeling in an image of the target object when the target object is analyzed to be anomalous, so as to obtain an analysis result of the target object;
and step A223, combining analysis results of all target objects on the pantograph into a pantograph analysis result, and sending the pantograph analysis result to the cloud platform.
Specifically, the target object is a target detection object of the pantograph, such as a carbon slide plate, a wire rope, a bolt fastening area, and the like.
Referring to fig. 7, further, step A3 includes:
Step A31, storing analysis results of all target objects of the pantograph according to train identification classification and time sequence;
Step A32, generating alarm information aiming at the analysis result of the target object with the abnormality, wherein the alarm information comprises an abnormality labeling image of the target object with the abnormality;
and step A33, displaying alarm information.
Specifically, in step a33, not only the alarm information but also the analysis result of each target object of the pantograph is displayed, the analysis result of each target object includes the image of each target object, and if the analysis result is abnormal, the displayed image is the image of the target object with the abnormal label. The abnormal marking can be to outline the abnormal area by using a picture frame, for example, when the carbon slide plate is cracked or damaged, the cracked area or the damaged area is marked, for example, when the bolt is fallen, the bolt is marked as the fallen area.
Referring to fig. 8, further, in step a31, the analysis results of the objects of the pantograph are stored according to the train identification classification and the time sequence, so as to form a time sequence of the analysis results of the objects;
Step a31 is followed by:
Step B1, performing risk prediction on an analysis result time sequence of each target object by using a pre-trained risk prediction model to obtain abnormal risk probability of each target object, and obtaining the abnormal risk probability of the pantograph based on the abnormal risk probability of each target object;
in step a33, the abnormal risk probability of each target object and the abnormal risk probability of the pantograph are displayed.
Preventive maintenance and risk management functions of the system are helpful for finding and solving potential problems in time, and guaranteeing train operation safety.
The introduction of the image recognition technology provides an efficient and accurate solution for intelligent detection of the locomotive pantograph. Through automatic image acquisition and advanced image processing algorithm, the intelligent detection system can realize rapid identification and analysis of the defects of the pantograph, so that the detection efficiency is greatly improved, personal errors are reduced, and the consistency and traceability of the detection result are ensured. The method is not only beneficial to improving the safety and reliability of railway transportation, but also provides important technical support for railway maintenance and operation, and promotes the technical progress and modern development of railway industry.
The foregoing is merely illustrative of the preferred embodiments of the present invention and is not intended to limit the embodiments and scope of the present invention, and it should be appreciated by those skilled in the art that equivalent substitutions and obvious variations may be made using the description and illustrations of the present invention, and are intended to be included in the scope of the present invention.
Claims (10)
1. A train pantograph detection system based on artificial intelligence, comprising:
The plurality of camera devices are arranged in a scattered manner in the train detection room and are used for collecting image data of a train identification part and a pantograph part on a train entering the train detection room;
The edge analysis device is connected with the camera device and is used for processing the image data based on a pre-trained image analysis model, identifying a train identifier, obtaining a pantograph analysis result and sending the pantograph analysis result to the cloud platform;
the cloud platform is connected with the edge analysis device and is used for:
classifying and storing the pantograph analysis result according to the train identification;
And when the pantograph abnormality exists in the pantograph analysis result, generating alarm information.
2. The artificial intelligence based train pantograph detection system according to claim 1, wherein the edge analysis device includes:
the storage server is used for storing the image data acquired by the image pickup device;
And the edge analysis equipment is connected with the storage server and is used for processing the image data stored in the storage server based on a pre-trained image analysis model, identifying a train identifier, obtaining a pantograph analysis result and sending the pantograph analysis result to the cloud platform.
3. The artificial intelligence based train pantograph detection system according to claim 2, wherein the edge analysis device includes:
The target detection module is used for identifying a target object in the image data based on a target identification model and segmenting out an image of the target object;
The abnormality analysis module is connected with the target detection module and is used for carrying out abnormality analysis on the target object based on an abnormality analysis model corresponding to the target object, and carrying out abnormality labeling in an image of the target object when the abnormality of the target object is analyzed, so as to obtain an analysis result of the target object;
and the result sending module is connected with the abnormality analysis module and is used for combining analysis results of the target objects on the pantograph into a pantograph analysis result and sending the pantograph analysis result to the cloud platform.
4. The artificial intelligence based train pantograph detection system of claim 3, wherein the cloud platform includes:
the cloud storage module is used for storing analysis results of all the target objects of the pantograph according to the train identification classification and the time sequence;
The alarm generation module is connected with the cloud storage module and is used for generating alarm information aiming at an analysis result of a target object with an abnormality, wherein the alarm information comprises an abnormality labeling image of the target object with the abnormality;
And the display module is connected with the alarm generation module and used for displaying the alarm information.
5. The artificial intelligence based train pantograph detection system of claim 3, wherein the cloud platform includes:
The cloud storage module is used for storing analysis results of all the target objects of the pantograph according to the train identification classification and the time sequence to form an analysis result time sequence of all the target objects;
The risk prediction module is connected with the cloud storage module and is used for performing risk prediction on the analysis result time sequence of each target object by using a pre-trained risk prediction model to obtain abnormal risk probability of each target object, and obtaining the abnormal risk probability of the pantograph based on the abnormal risk probability of each target object;
The display module is connected with the risk prediction module and is used for the abnormal risk probability of each target object and the abnormal risk probability of the pantograph.
6. A method for detecting a pantograph of a train based on artificial intelligence, wherein the system for detecting a pantograph of a train based on artificial intelligence according to any one of claims 1 to 5 is used, comprising:
step A1, a camera device collects image data of a train identification part and a pantograph part on a train entering a train detection room;
step A2, the edge analysis device processes the image data based on a pre-trained image analysis model, recognizes a train identifier, obtains a pantograph analysis result and sends the pantograph analysis result to a cloud platform;
and A3, the cloud platform stores the pantograph analysis results according to the train identification classification, and generates alarm information when the pantograph abnormality exists in the pantograph analysis results.
7. The method for detecting a pantograph of a train based on artificial intelligence according to claim 6, wherein the edge analysis device includes a storage server and an edge analysis device, and the step A2 includes:
step A21, the storage server stores the image data acquired by the camera device;
And step A22, the edge analysis equipment processes the image data stored in the storage server based on a pre-trained image analysis model, recognizes a train identification, obtains a pantograph analysis result and sends the pantograph analysis result to a cloud platform.
8. The method for detecting a pantograph of a train based on artificial intelligence according to claim 7, wherein the step a22 includes:
Step A221, identifying a target object in the image data based on a target identification model and segmenting out an image of the target object;
step A222, carrying out anomaly analysis on the target object based on an anomaly analysis model corresponding to the target object, and carrying out anomaly labeling in an image of the target object when the target object is analyzed to be anomalous, so as to obtain an analysis result of the target object;
and step A223, combining analysis results of the target objects on the pantograph into a pantograph analysis result, and sending the pantograph analysis result to the cloud platform.
9. The method for detecting a pantograph of a train based on artificial intelligence according to claim 8, wherein the step A3 includes:
step A31, storing analysis results of all target objects of the pantograph according to the train identification classification and time sequence;
step A32, generating alarm information aiming at an analysis result of a target object with an abnormality, wherein the alarm information comprises an abnormality labeling image of the target object with the abnormality;
And step A33, displaying the alarm information.
10. The method for detecting a pantograph of a train based on artificial intelligence according to claim 9, wherein in the step a31, analysis results of the respective target objects of the pantograph are stored in time sequence according to the train identification classification to form a time series of the analysis results of the respective target objects;
The step A31 further comprises the following steps:
Step B1, performing risk prediction on the analysis result time sequence of each target object by using a pre-trained risk prediction model to obtain the abnormal risk probability of each target object, and obtaining the abnormal risk probability of the pantograph based on the abnormal risk probability of each target object;
In the step a33, an abnormal risk probability of each target object and an abnormal risk probability of the pantograph are displayed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411335647.XA CN119379973A (en) | 2024-09-24 | 2024-09-24 | A train pantograph detection system and method based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411335647.XA CN119379973A (en) | 2024-09-24 | 2024-09-24 | A train pantograph detection system and method based on artificial intelligence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN119379973A true CN119379973A (en) | 2025-01-28 |
Family
ID=94331433
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202411335647.XA Pending CN119379973A (en) | 2024-09-24 | 2024-09-24 | A train pantograph detection system and method based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN119379973A (en) |
-
2024
- 2024-09-24 CN CN202411335647.XA patent/CN119379973A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR102017870B1 (en) | Real-time line defect detection system | |
| CN111507147A (en) | Intelligent inspection method and device, computer equipment and storage medium | |
| CN108037770A (en) | Unmanned plane power transmission line polling system and method based on artificial intelligence | |
| CN113726606A (en) | Abnormality detection method and apparatus, electronic device, and storage medium | |
| CN117557570B (en) | Rail vehicle abnormality detection method and system | |
| CN112560816A (en) | Equipment indicator lamp identification method and system based on YOLOv4 | |
| CN117214617B (en) | Smart power grid fault real-time monitoring and positioning system and method | |
| CN112083000B (en) | An intelligent identification method and system for appearance defects of substation equipment | |
| CN112508911A (en) | Rail joint touch net suspension support component crack detection system based on inspection robot and detection method thereof | |
| CN113780594A (en) | A server intelligent identification and inspection system | |
| CN113898537A (en) | Remote intelligent inspection system for wind power plant | |
| CN112115770A (en) | Method and system for identifying autonomous inspection defects of unmanned aerial vehicle of overhead line | |
| Hamey et al. | Pancam: In-service inspection of locomotive pantographs | |
| CN119810037A (en) | A method and system for detecting early cracks in asphalt concrete pavement | |
| CN117498225B (en) | Unmanned aerial vehicle intelligent power line inspection system | |
| CN119379973A (en) | A train pantograph detection system and method based on artificial intelligence | |
| CN111126771A (en) | Safety inspector image recognition quality supervision and guarantee system and method based on regional attention prediction | |
| CN115880638A (en) | Safety production management method based on artificial intelligence driving block chain | |
| CN120725975A (en) | Steel structure defect recognition system and method based on image analysis | |
| CN118034354B (en) | A power inspection control method and system based on unmanned aerial vehicle | |
| CN119722610A (en) | A method and system for identifying abnormality of chain bucket of continuous ship unloader | |
| CN119783974A (en) | A remote real-time monitoring system for on-site construction scenes | |
| CN119251997A (en) | A laboratory intelligent inspection method and system based on multiple data sources | |
| CN113131611B (en) | Valve hall equipment inspection early warning method and device, electronic equipment and storage medium | |
| CN115187880A (en) | Communication optical cable defect detection method and system based on image recognition and storage medium |
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 |