CN117705816A - Unmanned aerial vehicle-based steel rail surface defect detection method, system, equipment and medium - Google Patents

Unmanned aerial vehicle-based steel rail surface defect detection method, system, equipment and medium Download PDF

Info

Publication number
CN117705816A
CN117705816A CN202410169968.0A CN202410169968A CN117705816A CN 117705816 A CN117705816 A CN 117705816A CN 202410169968 A CN202410169968 A CN 202410169968A CN 117705816 A CN117705816 A CN 117705816A
Authority
CN
China
Prior art keywords
steel rail
aerial vehicle
unmanned aerial
rail surface
rail
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202410169968.0A
Other languages
Chinese (zh)
Other versions
CN117705816B (en
Inventor
李祥
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Southwest Jiaotong University
Original Assignee
Southwest Jiaotong University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Southwest Jiaotong University filed Critical Southwest Jiaotong University
Priority to CN202410169968.0A priority Critical patent/CN117705816B/en
Publication of CN117705816A publication Critical patent/CN117705816A/en
Application granted granted Critical
Publication of CN117705816B publication Critical patent/CN117705816B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • 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
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S19/00—Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
    • G01S19/38—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system
    • G01S19/39—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system the satellite radio beacon positioning system transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
    • G01S19/42—Determining position
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • 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
    • G06V10/778—Active pattern-learning, e.g. online learning of image or video features
    • 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
    • G06V10/84—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using probabilistic graphical models from image or video features, e.g. Markov models or Bayesian networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes
    • G06V20/17—Terrestrial scenes taken from planes or by drones
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/70—Labelling scene content, e.g. deriving syntactic or semantic representations
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B64—AIRCRAFT; AVIATION; COSMONAUTICS
    • B64U—UNMANNED AERIAL VEHICLES [UAV]; EQUIPMENT THEREFOR
    • B64U2101/00—UAVs specially adapted for particular uses or applications
    • B64U2101/30—UAVs specially adapted for particular uses or applications for imaging, photography or videography
    • 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
    • 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
    • G01N2021/8887—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 based on image processing techniques
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2201/00—Features of devices classified in G01N21/00
    • G01N2201/02—Mechanical
    • G01N2201/021—Special mounting in general
    • G01N2201/0214—Airborne
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2201/00—Features of devices classified in G01N21/00
    • G01N2201/12—Circuits of general importance; Signal processing
    • G01N2201/129—Using chemometrical methods
    • G01N2201/1296—Using chemometrical methods using neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10004—Still image; Photographic image
    • G06T2207/10012—Stereo images
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Remote Sensing (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Databases & Information Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Medical Informatics (AREA)
  • Signal Processing (AREA)
  • Quality & Reliability (AREA)
  • Probability & Statistics with Applications (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Computational Linguistics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
  • Image Analysis (AREA)

Abstract

The invention provides a rail surface defect detection method, a system, equipment and a medium based on an unmanned aerial vehicle, which relate to the technical field of maintenance and detection of rail transit rails and comprise the steps of acquiring image information of the rail surface and first rail position information corresponding to the image information by utilizing the unmanned aerial vehicle; inputting the image information and the steel rail position information into a preset steel rail defect detection model for processing and judging according to a deep learning algorithm to obtain a steel rail surface identification result; and carrying out defect judgment on the rail surface recognition result, carrying out corresponding marking on the rail surface, and sending marking information and second rail position information corresponding to the marking information to a management platform so as to finish defect maintenance. The unmanned aerial vehicle inspection system has the beneficial effects that the unmanned aerial vehicle is used for carrying the camera for non-contact inspection, so that the unmanned aerial vehicle inspection system can fully exert the advantages of high automation degree, low energy consumption, obvious reduction of labor intensity of inspection workers, and the problems of low efficiency and high maintenance labor cost of traditional manual inspection are solved.

Description

Unmanned aerial vehicle-based steel rail surface defect detection method, system, equipment and medium
Technical Field
The invention relates to the technical field of maintenance and detection of rail transit steel rails, in particular to a steel rail surface defect detection method, system, equipment and medium based on an unmanned aerial vehicle.
Background
In recent years, with the rapid development of rail transit in China, the inspection labor cost is rapidly increased, and the maintenance pressure of the rail transit is obviously increased. The machine vision technology based on unmanned aerial vehicle assistance is developed along with the development of technology, and the intelligent identification of the surface defects of the steel rail is realized by combining a big data technology instead of the traditional manual inspection, and the principle is that the surface images of the steel rail are obtained through the machine vision technology, the images are processed by utilizing a cloud server, and finally the information of the surface defects of the steel rail for road detection is obtained.
Along with the increment of rail transit mileage of trunk railways (high-speed railways and common-speed railways), inter-city railways, urban (suburban) railways and urban rail transit, the requirements on automation and intellectualization of managing and maintaining trunk rails, emergency rescue and the like are higher and higher, and the traditional manual inspection mode is adopted for collecting rail information of the rails, so that the problems of low working efficiency, long time consumption, high subjectivity depending on personal operation experience of inspection staff, poor intellectualization and the like exist.
Disclosure of Invention
The invention aims to provide a rail surface defect detection method, system, equipment and medium based on an unmanned aerial vehicle so as to solve the problems. In order to achieve the above purpose, the technical scheme adopted by the invention is as follows:
in a first aspect, the present application provides a method for detecting a rail surface defect based on an unmanned aerial vehicle, including:
acquiring image information of the surface of the steel rail and first steel rail position information corresponding to the image information by using an unmanned aerial vehicle;
inputting the image information and the first steel rail position information into a preset steel rail defect detection model for processing and judging according to a deep learning algorithm to obtain a steel rail surface recognition result;
performing defect judgment on the steel rail surface identification result, and if the defect exists, performing corresponding marking on the steel rail surface according to the defect type to obtain marking information, wherein the defect comprises bulges, pits, cracks, deformation and displacement;
and sending the marking information and the second rail position information corresponding to the marking information to a management platform, thereby completing defect maintenance.
Preferably, the acquiring, by using the unmanned aerial vehicle, image information of a surface of the steel rail and first steel rail position information corresponding to the image information includes:
acquiring images to be detected of the surface of the steel rail in real time along a patrol track by using an unmanned aerial vehicle, wherein the unmanned aerial vehicle comprises a GPS (global positioning system), a 3D (three-dimensional) camera and a laser displacement sensor, the 3D camera is used for acquiring the images to be detected, and the laser displacement sensor is used for providing a light source for the 3D camera;
performing image processing on the image to be detected to obtain color image information of the surface of the steel rail, and recording the color image information as image information;
and acquiring the first rail position information corresponding to the image information in real time according to the GPS.
Preferably, the step of inputting the image information and the rail position information into a preset rail defect detection model according to a deep learning algorithm to process and judge to obtain a rail surface recognition result includes:
preprocessing the image information to obtain a preprocessing result, wherein the preprocessing comprises the step of adjusting the size of a steel rail picture in the image information to the size required by a preset steel rail defect detection model;
inputting the steel rail picture in the pretreatment result into a neural network of a convolution layer, an activation function and a pooling layer for training to obtain first characteristic information in the steel rail picture;
aiming at the first characteristic information, calculating the score condition of the steel rail picture, and further determining the position information of the target to be detected contained in the steel rail picture;
performing defect classification on the position information of the target to be detected based on the Softmax function and a preset steel rail defect detection model to obtain a probability distribution result;
and calculating a probability distribution result according to a non-maximum suppression algorithm to obtain a steel rail surface recognition result, wherein the steel rail surface recognition result comprises the category and position information of the steel rail surface defects.
Preferably, the calculating the probability distribution result according to the non-maximum suppression algorithm to obtain a rail surface recognition result includes:
calculating the scores of all candidate frames in the probability distribution result, and extracting a boundary frame with the highest score as a reference frame;
the cross ratio between each candidate frame except the reference frame and the reference frame is calculated, and the calculation formula is as follows:
in the method, in the process of the invention,is the firstiThe cross-ratios between the candidate frames and the reference frames,Ais a reference frame (I)>Is the firstiCandidate frames;
judging whether the value of the cross-over ratio exceeds a preset value, if so, predicting the cross-over ratio as the same target, and eliminating candidate frames of which the value of the cross-over ratio exceeds the preset value;
and selecting the highest score in all the remaining candidate frames as a new reference frame, repeating the judging step until all the candidate frames are removed, and marking the remaining candidate frames as a steel rail surface recognition result.
In a second aspect, the application further provides a rail surface defect detection system based on an unmanned aerial vehicle, which comprises an acquisition module, an identification module, a marking module and a detection module, wherein:
and the acquisition module is used for: the method comprises the steps of acquiring image information of the surface of a steel rail by using an unmanned aerial vehicle and first steel rail position information corresponding to the image information;
and an identification module: the method comprises the steps of inputting image information and first steel rail position information into a preset steel rail defect detection model according to a deep learning algorithm, and processing and judging to obtain a steel rail surface recognition result;
and a marking module: the method comprises the steps of performing defect judgment on a steel rail surface identification result, and if the defect exists, performing corresponding marking on the steel rail surface according to the defect type to obtain marking information, wherein the defect comprises protrusions, depressions, cracks, deformation and displacement;
and a detection module: and the second rail position information is used for sending the marking information and the second rail position information corresponding to the marking information to the management platform, so that defect maintenance is completed.
In a third aspect, the present application further provides a rail surface defect detection device based on an unmanned aerial vehicle, including:
a memory for storing a computer program;
and the processor is used for realizing the steps of the unmanned aerial vehicle-based steel rail surface defect detection method when executing the computer program.
In a fourth aspect, the present application further provides a readable storage medium, on which a computer program is stored, which when executed by a processor, implements the steps of the above-mentioned rail surface defect detection method based on an unmanned aerial vehicle.
The beneficial effects of the invention are as follows:
according to the invention, an unmanned aerial vehicle, a GPS, a laser displacement sensor, a 3D camera, a cloud server, a controller and a code spraying device are adopted, the unmanned aerial vehicle is provided with the GPS, the 3D camera and the laser displacement sensor, the 3D camera acquires image information of the surface of a rail along a patrol track in real time so as to acquire a color picture of the surface of the rail, the GPS acquires the position information of the rail in real time, and the laser displacement sensor provides a proper light source for the 3D camera; transmitting the acquired rail image information and position information to a cloud server, wherein the cloud server processes and judges the acquired image through a rail defect detection model based on deep learning, and if a defect (bulge, recess, fracture, deformation and bit shift) exists in the obtained identification result, if the defect exists, the cloud server sends different signals to a controller according to the defect type, and the code spraying device receives a corresponding control instruction and carries out corresponding code spraying mark on the surface of the defective rail, and the cloud server presents the rail surface defect image information and position information to maintenance staff, and the maintenance staff carries out code scanning maintenance to the defect-set position according to the given position; the unmanned aerial vehicle is used for carrying the camera to carry out non-contact inspection, so that the problems of low efficiency and high maintenance labor cost of traditional manual inspection are solved.
According to the invention, the unmanned aerial vehicle is used for carrying the camera for non-contact inspection, so that the unmanned aerial vehicle has the advantages of high automation degree, low energy consumption, obvious reduction of labor intensity of inspection workers, and the problems of low efficiency and high maintenance labor cost of traditional manual inspection are solved.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be apparent from the description, or may be learned by practice of the embodiments of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims thereof as well as the appended drawings.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the embodiments will be briefly described below, it being understood that the following drawings only illustrate some embodiments of the present invention and therefore should not be considered as limiting the scope, and other related drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a schematic flow chart of a rail surface defect detection method based on an unmanned aerial vehicle according to an embodiment of the invention;
fig. 2 is a schematic structural diagram of a rail surface defect detection system based on an unmanned aerial vehicle according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of a rail surface defect detecting device based on an unmanned aerial vehicle according to an embodiment of the present invention.
In the figure: 701. an acquisition module; 7011. a first acquisition unit; 7012. a processing unit; 7013. a second acquisition unit; 702. An identification module; 7021. a preprocessing unit; 7022. a training unit; 7023. a determination unit; 7024. a classification unit; 7025. a calculation unit; 70251. an extraction unit; 70252. calculating an intersection ratio unit; 70253. a judging unit; 70254. a circulation unit; 703. a marking module; 704. a detection module; 800. rail surface defect detection equipment based on unmanned aerial vehicle; 801. a processor; 802. a memory; 803. a multimedia component; 804. an I/O interface; 805. a communication component.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. The components of the embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the invention, as presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. 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: like reference numerals and letters denote like items in the following figures, and thus once an item is defined in one figure, no further definition or explanation thereof is necessary in the following figures. Meanwhile, in the description of the present invention, the terms "first", "second", and the like are used only to distinguish the description, and are not to be construed as indicating or implying relative importance.
Example 1:
the embodiment provides a rail surface defect detection method based on an unmanned aerial vehicle.
Referring to fig. 1, the method is shown to include step S100, step S200, step S300, and step S400.
The invention discloses a steel rail surface detection device, which comprises an unmanned aerial vehicle, a GPS (global positioning system), a laser displacement sensor, a 3D (three-dimensional) camera, a cloud server, a controller and a code sprayer, wherein the unmanned aerial vehicle hovers right above a rail, the camera, the GPS, the laser displacement sensor, the controller and the code sprayer are arranged below the unmanned aerial vehicle, the cloud server is used for receiving and processing a steel rail surface image and is connected with the cloud server through a camera wireless transmission module, a steel rail surface image processing result is input to the controller through the cloud server wireless transmission module, the controller is connected with the code sprayer, and the code sprayer is used for carrying out code spraying marking at a defect position of a connecting rail surface according to a received controller instruction.
S100, acquiring image information of the surface of the steel rail by using an unmanned aerial vehicle and first steel rail position information corresponding to the image information.
It will be appreciated that S101, S102 and S103 are included in this step S100, wherein:
s101, acquiring an image to be detected of the surface of a steel rail in real time along a patrol track by using an unmanned aerial vehicle, wherein the unmanned aerial vehicle comprises a GPS, a 3D camera and a laser displacement sensor, the 3D camera is used for acquiring the image to be detected, and the laser displacement sensor is used for providing a light source for the 3D camera;
it is to be noted that the unmanned aerial vehicle hovers over the rail for inspection, and the 3D camera, the GPS, the laser displacement sensor, the controller and the code spraying device are arranged below the unmanned aerial vehicle; the GPS is used for positioning the current geographic position of the unmanned aerial vehicle; the laser displacement sensor provides a light source for shooting by the 3D camera, provides initial coordinates for detecting the steel rail picture, and provides a basis for image calibration.
S102, performing image processing on an image to be detected to obtain color image information of the surface of the steel rail, and recording the color image information as image information;
s103, acquiring first steel rail position information corresponding to the image information in real time according to the GPS.
The 3D camera is used for shooting the surface image of the steel rail in real time and transmitting the surface image to the cloud server through the wireless transmission module; the cloud server is used for receiving GPS data in real time, processing the surface image of the steel rail by the 3D camera, and determining the surface defect of the steel rail.
And S200, inputting the image information and the first steel rail position information into a preset steel rail defect detection model for processing and judging according to a deep learning algorithm, and obtaining a steel rail surface recognition result.
It will be appreciated that S201, S202, S203, S204, and S205 are included in the present step S200, wherein:
s201, preprocessing the image information to obtain a preprocessing result, wherein the preprocessing comprises the step of adjusting the size of a steel rail picture in the image information to a size required by a preset steel rail defect detection model;
s202, inputting a steel rail picture in a preprocessing result into a neural network of a convolution layer, an activation function and a pooling layer for training to obtain first characteristic information in the steel rail picture;
it should be noted that the resized rail picture is transferred in the neural network through the convolution layer, the activation function and the pooling layer, so as to gradually extract the important features in the picture. The convolution calculation formula is as follows:
in the method, in the process of the invention,is the output characteristic diagram is->Is used as a reference to the value of (a),Kis a convolution kernel which is a convolution kernel,Iis an input signature, m and n vary over the height and width of the convolution kernel, respectively.
Wherein the activation function is:
where x is the input pixel value,for pixel output values calculated via the activation function, < >>Is the power of-x of the natural constant e.
Wherein, the pooling calculation formula is:
in the method, in the process of the invention,outputting the value at position (x, y) for pooling,/for pooling>For inputting the feature map at position->R represents the set of all input pixel positions.
S203, aiming at the first characteristic information, calculating the score of the steel rail picture so as to determine the position information of the target to be detected contained in the steel rail picture;
it should be noted that, the first characteristic information is that after the rail image is extracted through multi-layer characteristic circulation, the image is calculated and scored through a neural network algorithm, and the position area of the target to be detected contained in the rail image is determined, wherein the calculation formula is as follows:
in the method, in the process of the invention,is the activation function; />Object scores which are not calculated by the activating function for the prediction frame are more divergent in numerical value; />For object scores calculated via the activation function, the value is defined to be between 0 and 1.
When the score reaches or exceeds a preset value, the detection of the target is indicated, and meanwhile, the feature map coordinates are mapped back to the original image coordinates by using inverse convolution, and the calculation formula is as follows:
in the method, in the process of the invention,for the position information after the deconvolution operation, upsamples is the upsampling operation,Iscale_factor is the upsampled scale factor for the position information of the current feature map.
S204, carrying out defect classification on the position information of the target to be detected based on a Softmax function and a preset steel rail defect detection model to obtain a probability distribution result;
it should be noted that, based on the Softmax function target classification algorithm, each prediction frame is calculated, the neural network predicts the probability that the position information of the target to be detected belongs to a certain type of defect (bump, dent, fracture, deformation and bit shift) according to the trained steel rail defect detection model, and converts the output of the last full connection layer into probability distribution, and the calculation formula is as follows:
in the method, in the process of the invention,is a category ofiPrediction probability of +.>Is a category ofiK is the total number of categories,jfor an integer from 1 to K,is +.>To the power.
Therefore, when the prediction probability of a certain category is maximum, the model considers that the defect is detected, and a probability distribution result is obtained.
S205, calculating a probability distribution result according to a non-maximum suppression algorithm to obtain a steel rail surface recognition result, wherein the steel rail surface recognition result comprises the category and position information of the steel rail surface defects.
In step S205, S2051, S2052, S2053, and S2054 are included, in which:
s2051, calculating scores of all candidate frames in a probability distribution result, and extracting a boundary frame with the highest score as a reference frame;
s2052, calculating the cross ratio between each candidate frame except the reference frame and the reference frame, wherein the calculation formula is as follows:
in the method, in the process of the invention,is the firstiThe cross-ratios between the candidate frames and the reference frames,Ais a reference frame (I)>Is the firstiCandidate frames;
s2053, judging whether the value of the cross-over ratio exceeds a preset value, if so, predicting the cross-over ratio as the same target, and eliminating candidate frames with the value exceeding the preset value;
s2054, selecting the highest score of all the remaining candidate frames as a new reference frame, repeating the judging step until all the candidate frames are removed, and marking the remaining candidate frames as a rail surface recognition result.
It will be appreciated that Non-maximum suppression (Non-maximum Suppression, NMS) is used to remove duplicate or low probability detection frames to ensure that each target in the final result only generates one most accurate detection frame, and thus the type and location of the defect.
And S300, judging the defect of the steel rail surface identification result, and if the defect exists, correspondingly marking the steel rail surface according to the defect type to obtain marking information, wherein the defect comprises protrusions, depressions, cracks, deformation and displacement.
It can be understood that in this step, the controller controls the flight trajectory and the flight attitude of the unmanned aerial vehicle by sending out a signal instruction; and receiving defect information of the cloud server, and controlling the code spraying device to perform code spraying operation. The code spraying device can perform defect code spraying identification on the corresponding steel rail surface position according to the instruction of the controller.
It should be noted that, the judgment is made according to whether or not there is a defect (protrusion, depression, fracture, deformation, shift) in the rail surface recognition result obtained in S200. If the defect exists, the cloud server sends different signals to the controller according to the defect type, and the code spraying device receives the corresponding control instruction and carries out corresponding code spraying marks on the surface of the defective steel rail so as to be convenient for maintenance management personnel to identify and verify. If not, the cloud server continues to perform the above steps to perform the loop calculation.
S400, sending the marking information and the second rail position information corresponding to the marking information to a management platform, and further completing defect maintenance.
It can be understood that in the step, the cloud server presents the picture information and the position information of the defects on the surface of the steel rail to maintenance management personnel, and the maintenance personnel go to the defect set position according to the given position to carry out code scanning maintenance.
In summary, the cloud server is used for accepting and processing the steel rail surface image, is connected with the cloud server through the camera wireless transmission module, and the steel rail surface image processing result is input to the controller through the cloud server wireless transmission module, and the controller is connected with the code spraying device, the code spraying device sprays the code on the defect of the surface of the connecting track according to the received controller instruction, so that the unmanned aerial vehicle is high in automation degree, low in energy consumption, and capable of remarkably reducing labor intensity of inspection workers, and the problems of low traditional manual inspection efficiency and high maintenance labor cost are solved.
Example 2
As shown in fig. 2, the present embodiment provides a rail surface defect detection system based on an unmanned aerial vehicle, and the system described with reference to fig. 2 includes an acquisition module 701, an identification module 702, a marking module 703 and a detection module 704, wherein:
the acquisition module 701: the method comprises the steps of acquiring image information of the surface of a steel rail by using an unmanned aerial vehicle and first steel rail position information corresponding to the image information;
the identification module 702: the method comprises the steps of inputting image information and steel rail position information into a preset steel rail defect detection model for processing and judging according to a deep learning algorithm to obtain a steel rail surface recognition result;
marking module 703: the method comprises the steps of performing defect judgment on a steel rail surface identification result, and if the defect exists, performing corresponding marking on the steel rail surface according to the defect type to obtain marking information, wherein the defect comprises protrusions, depressions, cracks, deformation and displacement;
detection module 704: and the second rail position information is used for sending the marking information and the second rail position information corresponding to the marking information to the management platform, so that defect maintenance is completed.
Specifically, the acquisition module 701 includes a first acquisition unit 7011, a processing unit 7012, and a second acquisition unit 7013, wherein:
first acquisition unit 7011: the method comprises the steps that an unmanned aerial vehicle is used for collecting images to be detected of the surface of a steel rail in real time along a patrol track, wherein the unmanned aerial vehicle comprises a GPS, a 3D camera and a laser displacement sensor, the 3D camera is used for collecting the images to be detected, and the laser displacement sensor is used for providing a light source for the 3D camera;
processing unit 7012: the method comprises the steps of performing image processing on an image to be detected to obtain color image information of the surface of a steel rail, and recording the color image information as image information;
second acquisition unit 7013: and the first rail position information corresponding to the image information is acquired in real time according to the GPS.
Specifically, the recognition module 702 includes a preprocessing unit 7021, a training unit 7022, a determining unit 7023, a classifying unit 7024, and a calculating unit 7025, wherein:
preprocessing unit 7021: the method comprises the steps of preprocessing image information to obtain a preprocessing result, wherein the preprocessing comprises the step of adjusting the size of a steel rail picture in the image information to the size required by a preset steel rail defect detection model;
training unit 7022: the method comprises the steps of inputting a steel rail picture in a pretreatment result into a neural network of a convolution layer, an activation function and a pooling layer for training to obtain first characteristic information in the steel rail picture;
determination unit 7023: the method comprises the steps of calculating a steel rail picture score aiming at first characteristic information so as to determine position information of a target to be detected contained in the steel rail picture;
classification unit 7024: the method is used for carrying out defect classification on the position information of the target to be detected based on a Softmax function and a preset steel rail defect detection model to obtain a probability distribution result;
calculation unit 7025: and the method is used for calculating the probability distribution result according to the non-maximum suppression algorithm to obtain a steel rail surface recognition result, wherein the steel rail surface recognition result comprises the category and position information of the steel rail surface defects.
Specifically, the calculating unit 7025 includes an extracting unit 70251, a calculating-to-combining-ratio unit 70252, a judging unit 70253, and a circulating unit 70254, wherein:
extraction unit 70251: the method comprises the steps of calculating scores of all candidate frames in probability distribution results, and extracting a boundary frame with the highest score as a reference frame;
calculating an intersection ratio unit 70252: for calculating the intersection ratio between each candidate frame other than the reference frame and the reference frame, the calculation formula is as follows:
in the method, in the process of the invention,is the firstiThe cross-ratios between the candidate frames and the reference frames,Ais a reference frame (I)>Is the firstiCandidate frames；
The judgment unit 70253: judging whether the value of the cross-over ratio exceeds a preset value, if so, predicting the cross-over ratio as the same target, and eliminating candidate frames with the cross-over ratio exceeding the preset value;
circulation unit 70254: and selecting the highest score of all the remaining candidate frames as a new reference frame, repeating the judging step until all the candidate frames are removed, and marking the remaining candidate frames as a steel rail surface recognition result.
It should be noted that, regarding the system in the above embodiment, the specific manner in which the respective modules perform the operations has been described in detail in the embodiment regarding the method, and will not be described in detail herein.
Example 3
Corresponding to the above method embodiment, in this embodiment, a rail surface defect detecting device based on an unmanned aerial vehicle is further provided, and a rail surface defect detecting device based on an unmanned aerial vehicle described below and a rail surface defect detecting method based on an unmanned aerial vehicle described above may be referred to correspondingly.
Fig. 3 is a block diagram illustrating an unmanned-vehicle-based rail surface defect detection apparatus 800, according to an exemplary embodiment. As shown in fig. 3, the unmanned aerial vehicle-based rail surface defect detection apparatus 800 includes: a processor 801 and a memory 802. The unmanned aerial vehicle-based rail surface defect detection apparatus 800 further includes one or more of a multimedia component 803, an i/O interface 804, and a communication component 805.
The processor 801 is configured to control the overall operation of the rail surface defect detecting device 800 based on the unmanned aerial vehicle, so as to complete all or part of the steps in the rail surface defect detecting method based on the unmanned aerial vehicle. The memory 802 is used to store various types of data to support operation at the unmanned aerial vehicle-based rail surface defect detection apparatus 800, which may include, for example, instructions for any application or method operating on the unmanned aerial vehicle-based rail surface defect detection apparatus 800, as well as application-related data, such as contact data, messaging, pictures, audio, video, and the like. The Memory 802 may be implemented by any type or combination of volatile or non-volatile Memory devices, such as static random access Memory (Static Random Access Memory, SRAM for short), electrically erasable programmable Read-Only Memory (Electrically Erasable Programmable Read-Only Memory, EEPROM for short), erasable programmable Read-Only Memory (Erasable Programmable Read-Only Memory, EPROM for short), programmable Read-Only Memory (Programmable Read-Only Memory, PROM for short), read-Only Memory (ROM for short), magnetic Memory, flash Memory, magnetic disk, or optical disk. The multimedia component 803 may include a screen and an audio component. Wherein the screen may be, for example, a touch screen, the audio component being for outputting and/or inputting audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signals may be further stored in the memory 802 or transmitted through the communication component 805. The audio assembly further comprises at least one speaker for outputting audio signals. The I/O interface 804 provides an interface between the processor 801 and other interface modules, which may be a keyboard, mouse, or buttons, etc. These buttons may be virtual buttons or physical buttons. The communication component 805 is configured to perform wired or wireless communication between the unmanned aerial vehicle-based rail surface defect detection apparatus 800 and other apparatuses. Wireless communication, such as Wi-Fi, bluetooth, near field communication (Near FieldCommunication, NFC for short), 2G, 3G, or 4G, or a combination of one or more thereof, the respective communication component 805 may thus comprise: wi-Fi module, bluetooth module or NFC module.
In an exemplary embodiment, the unmanned aerial vehicle-based rail surface defect detection apparatus 800 may be implemented by one or more application specific integrated circuits (Application Specific Integrated Circuit, abbreviated as ASIC), digital signal processors (DigitalSignal Processor, abbreviated as DSP), digital signal processing apparatus (Digital Signal Processing Device, abbreviated as DSPD), programmable logic devices (Programmable Logic Device, abbreviated as PLD), field programmable gate arrays (Field Programmable Gate Array, abbreviated as FPGA), controllers, microcontrollers, microprocessors, or other electronic components for performing the unmanned aerial vehicle-based rail surface defect detection method described above.
In another exemplary embodiment, a computer readable storage medium is also provided, comprising program instructions which, when executed by a processor, implement the steps of the unmanned aerial vehicle-based rail surface defect detection method described above. For example, the computer readable storage medium may be the memory 802 described above including program instructions executable by the processor 801 of the unmanned aerial vehicle-based rail surface defect detection apparatus 800 to perform the unmanned aerial vehicle-based rail surface defect detection method described above.
Example 4
Corresponding to the above method embodiment, a readable storage medium is further provided in this embodiment, and a readable storage medium described below and a rail surface defect detection method based on an unmanned aerial vehicle described above may be referred to correspondingly.
The readable storage medium stores a computer program which when executed by a processor realizes the steps of the rail surface defect detection method based on the unmanned aerial vehicle in the method embodiment.
The readable storage medium may be a usb disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a magnetic disk, or an optical disk, and the like.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
The foregoing is merely illustrative of the present invention, and the present invention is not limited thereto, and any person skilled in the art will readily recognize that variations or substitutions are within the scope of the present invention. Therefore, the protection scope of the invention is subject to the protection scope of the claims.

Claims (10)

1. The method for detecting the surface defects of the steel rail based on the unmanned aerial vehicle is characterized by comprising the following steps of:
acquiring image information of the surface of the steel rail and first steel rail position information corresponding to the image information by using an unmanned aerial vehicle;
inputting the image information and the first steel rail position information into a preset steel rail defect detection model for processing and judging according to a deep learning algorithm to obtain a steel rail surface recognition result;
performing defect judgment on the steel rail surface identification result, and if the defect exists, performing corresponding marking on the steel rail surface according to defect types to obtain marking information, wherein the defect types comprise bulges, pits, cracks, deformation and displacement;
and sending the marking information and the second rail position information corresponding to the marking information to a management platform, thereby completing defect maintenance.
2. The unmanned aerial vehicle-based rail surface defect detection method according to claim 1, wherein the acquiring image information of the rail surface and first rail position information corresponding to the image information by using the unmanned aerial vehicle comprises:
acquiring images to be detected of the surface of the steel rail in real time along a patrol track by using an unmanned aerial vehicle, wherein the unmanned aerial vehicle comprises a GPS (global positioning system), a 3D (three-dimensional) camera and a laser displacement sensor, the 3D camera is used for acquiring the images to be detected, and the laser displacement sensor is used for providing a light source for the 3D camera;
performing image processing on the image to be detected to obtain color image information of the surface of the steel rail, and recording the color image information as image information;
and acquiring the first rail position information corresponding to the image information in real time according to the GPS.
3. The unmanned aerial vehicle-based rail surface defect detection method of claim 1, wherein the inputting the image information and the first rail position information into a preset rail defect detection model according to a deep learning algorithm for processing and discriminating, to obtain a rail surface recognition result, comprises:
preprocessing the image information to obtain a preprocessing result, wherein the preprocessing comprises the step of adjusting the size of a steel rail picture in the image information to the size required by a preset steel rail defect detection model;
inputting the steel rail picture in the pretreatment result into a neural network of a convolution layer, an activation function and a pooling layer for training to obtain first characteristic information in the steel rail picture;
aiming at the first characteristic information, calculating the score condition of the steel rail picture, and further determining the position information of the target to be detected contained in the steel rail picture;
performing defect classification on the position information of the target to be detected based on the Softmax function and a preset steel rail defect detection model to obtain a probability distribution result;
and calculating a probability distribution result according to a non-maximum suppression algorithm to obtain a steel rail surface recognition result, wherein the steel rail surface recognition result comprises the category and position information of the steel rail surface defects.
4. The unmanned aerial vehicle-based steel rail surface defect detection method of claim 3, wherein the calculating the probability distribution result according to the non-maximum suppression algorithm to obtain the steel rail surface recognition result comprises the following steps:
calculating the scores of all candidate frames in the probability distribution result, and extracting a boundary frame with the highest score as a reference frame;
the cross ratio between each candidate frame except the reference frame and the reference frame is calculated, and the calculation formula is as follows:
in (1) the->Is the firstiThe cross-ratios between the candidate frames and the reference frames,Ais a reference frame (I)>Is the firstiCandidate frames;
judging whether the value of the cross-over ratio exceeds a preset value, if so, predicting the cross-over ratio as the same target, and eliminating candidate frames of which the value of the cross-over ratio exceeds the preset value;
and selecting the highest score in all the remaining candidate frames as a new reference frame, repeating the judging step until all the candidate frames are removed, and marking the remaining candidate frames as a steel rail surface recognition result.
5. Rail surface defect detection system based on unmanned aerial vehicle, based on the rail surface defect detection method based on unmanned aerial vehicle according to any one of claims 1-4, characterized in that it comprises:
and the acquisition module is used for: the method comprises the steps of acquiring image information of the surface of a steel rail by using an unmanned aerial vehicle and first steel rail position information corresponding to the image information;
and an identification module: the method comprises the steps of inputting image information and first steel rail position information into a preset steel rail defect detection model according to a deep learning algorithm, and processing and judging to obtain a steel rail surface recognition result;
and a marking module: the method comprises the steps of performing defect judgment on a steel rail surface identification result, and if a defect exists, performing corresponding marking on the steel rail surface according to defect types to obtain marking information, wherein the defect types comprise bulges, pits, cracks, deformation and displacement;
and a detection module: and the second rail position information is used for sending the marking information and the second rail position information corresponding to the marking information to the management platform, so that defect maintenance is completed.
6. The unmanned aerial vehicle-based rail surface defect detection system of claim 5, wherein the acquisition module comprises:
the first acquisition unit: the method comprises the steps that an unmanned aerial vehicle is used for collecting images to be detected of the surface of a steel rail in real time along a patrol track, wherein the unmanned aerial vehicle comprises a GPS, a 3D camera and a laser displacement sensor, the 3D camera is used for collecting the images to be detected, and the laser displacement sensor is used for providing a light source for the 3D camera;
and a processing unit: the method comprises the steps of performing image processing on an image to be detected to obtain color image information of the surface of a steel rail, and recording the color image information as image information;
the second acquisition unit: and the first rail position information corresponding to the image information is acquired in real time according to the GPS.
7. The unmanned aerial vehicle-based rail surface defect detection system of claim 5, wherein the identification module comprises:
pretreatment unit: the method comprises the steps of preprocessing image information to obtain a preprocessing result, wherein the preprocessing comprises the step of adjusting the size of a steel rail picture in the image information to the size required by a preset steel rail defect detection model;
training unit: the method comprises the steps of inputting a steel rail picture in a pretreatment result into a neural network of a convolution layer, an activation function and a pooling layer for training to obtain first characteristic information in the steel rail picture;
a determination unit: the method comprises the steps of calculating the score condition of a steel rail picture aiming at first characteristic information, and further determining the position information of a target to be detected contained in the steel rail picture;
classification unit: the method is used for carrying out defect classification on the position information of the target to be detected based on a Softmax function and a preset steel rail defect detection model to obtain a probability distribution result;
a calculation unit: and the method is used for calculating the probability distribution result according to the non-maximum suppression algorithm to obtain a steel rail surface recognition result, wherein the steel rail surface recognition result comprises the category and position information of the steel rail surface defects.
8. The unmanned aerial vehicle-based rail surface defect detection system of claim 7, wherein the computing unit comprises:
extraction unit: the method comprises the steps of calculating scores of all candidate frames in probability distribution results, and extracting a boundary frame with the highest score as a reference frame;
calculating an intersection ratio unit: for calculating the intersection ratio between each candidate frame other than the reference frame and the reference frame, the calculation formula is as follows:
in (1) the->Is the firstiThe cross-ratios between the candidate frames and the reference frames,Ais a reference frame (I)>Is the firstiCandidate frames;
a judging unit: judging whether the value of the cross-over ratio exceeds a preset value, if so, predicting the cross-over ratio as the same target, and eliminating candidate frames with the cross-over ratio exceeding the preset value;
and a circulation unit: and selecting the highest score of all the remaining candidate frames as a new reference frame, repeating the judging step until all the candidate frames are removed, and marking the remaining candidate frames as a steel rail surface recognition result.
9. Rail surface defect detection equipment based on unmanned aerial vehicle, characterized by, include:
a memory for storing a computer program;
a processor for implementing the unmanned aerial vehicle-based rail surface defect detection method according to any one of claims 1 to 4 when executing the computer program.
10. A readable storage medium, characterized by: the readable storage medium has stored thereon a computer program which, when executed by a processor, implements the unmanned aerial vehicle-based rail surface defect detection method according to any one of claims 1 to 4.
CN202410169968.0A 2024-02-06 2024-02-06 Unmanned aerial vehicle-based steel rail surface defect detection method, system, equipment and medium Active CN117705816B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202410169968.0A CN117705816B (en) 2024-02-06 2024-02-06 Unmanned aerial vehicle-based steel rail surface defect detection method, system, equipment and medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202410169968.0A CN117705816B (en) 2024-02-06 2024-02-06 Unmanned aerial vehicle-based steel rail surface defect detection method, system, equipment and medium

Publications (2)

Publication Number Publication Date
CN117705816A true CN117705816A (en) 2024-03-15
CN117705816B CN117705816B (en) 2024-04-26

Family

ID=90162941

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202410169968.0A Active CN117705816B (en) 2024-02-06 2024-02-06 Unmanned aerial vehicle-based steel rail surface defect detection method, system, equipment and medium

Country Status (1)

Country Link
CN (1) CN117705816B (en)

Citations (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1999054717A2 (en) * 1998-04-17 1999-10-28 Cae Electronic Ltd. Cae Electronique Ltee Method and apparatus for identification of probable defects in a workpiece
CN105501248A (en) * 2016-02-16 2016-04-20 株洲时代电子技术有限公司 Railway line inspection system
US20180024076A1 (en) * 2016-07-20 2018-01-25 Ball Corporation System and method for aligning an inker of a decorator
US20180157938A1 (en) * 2016-12-07 2018-06-07 Samsung Electronics Co., Ltd. Target detection method and apparatus
CN108921846A (en) * 2018-07-17 2018-11-30 北京航空航天大学 A kind of rail tread defect identification method combined based on gray level image and depth image
CN108960134A (en) * 2018-07-02 2018-12-07 广东容祺智能科技有限公司 A kind of patrol UAV image mark and intelligent identification Method
CN109409404A (en) * 2018-09-13 2019-03-01 西南交通大学 A kind of high iron catenary radix saposhnikoviae bracing wire fault detection method based on deep learning
CN110111328A (en) * 2019-05-16 2019-08-09 上海中认尚科新能源技术有限公司 A kind of blade crack of wind driven generator detection method based on convolutional neural networks
CN111126287A (en) * 2019-12-25 2020-05-08 武汉大学 Remote sensing image dense target deep learning detection method
CN111242891A (en) * 2019-12-27 2020-06-05 北京交通大学 Method for identifying and classifying surface defects of steel rails
CN111724342A (en) * 2020-05-18 2020-09-29 天津大学 Method for detecting thyroid nodule in ultrasonic image
CN112067632A (en) * 2020-06-04 2020-12-11 西南交通大学 Power equipment detection cloud platform and detection method
CN113112501A (en) * 2021-05-11 2021-07-13 上海市东方海事工程技术有限公司 Vehicle-mounted track inspection device and method based on deep learning
US20210291980A1 (en) * 2020-03-18 2021-09-23 Coretronic Corporation Unmanned aerial vehicle and image recognition method thereof
CN114384075A (en) * 2021-12-06 2022-04-22 西安理工大学 Track slab defect online detection system and detection method based on three-dimensional laser scanning
CN114693617A (en) * 2022-03-17 2022-07-01 中山大学 Method, system, equipment and storage medium for detecting defects of railway steel rails
CN115731210A (en) * 2022-11-29 2023-03-03 国网浙江省电力有限公司信息通信分公司 Small sample target detection method based on multi-angle information and related equipment
CN115937492A (en) * 2022-12-12 2023-04-07 国网四川省电力公司电力科学研究院 Transformer equipment infrared image identification method based on feature identification
CN116228780A (en) * 2023-05-10 2023-06-06 江苏森标科技有限公司 Silicon wafer defect detection method and system based on computer vision

Patent Citations (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1999054717A2 (en) * 1998-04-17 1999-10-28 Cae Electronic Ltd. Cae Electronique Ltee Method and apparatus for identification of probable defects in a workpiece
CN105501248A (en) * 2016-02-16 2016-04-20 株洲时代电子技术有限公司 Railway line inspection system
US20180024076A1 (en) * 2016-07-20 2018-01-25 Ball Corporation System and method for aligning an inker of a decorator
US20180157938A1 (en) * 2016-12-07 2018-06-07 Samsung Electronics Co., Ltd. Target detection method and apparatus
CN108960134A (en) * 2018-07-02 2018-12-07 广东容祺智能科技有限公司 A kind of patrol UAV image mark and intelligent identification Method
CN108921846A (en) * 2018-07-17 2018-11-30 北京航空航天大学 A kind of rail tread defect identification method combined based on gray level image and depth image
CN109409404A (en) * 2018-09-13 2019-03-01 西南交通大学 A kind of high iron catenary radix saposhnikoviae bracing wire fault detection method based on deep learning
CN110111328A (en) * 2019-05-16 2019-08-09 上海中认尚科新能源技术有限公司 A kind of blade crack of wind driven generator detection method based on convolutional neural networks
CN111126287A (en) * 2019-12-25 2020-05-08 武汉大学 Remote sensing image dense target deep learning detection method
CN111242891A (en) * 2019-12-27 2020-06-05 北京交通大学 Method for identifying and classifying surface defects of steel rails
US20210291980A1 (en) * 2020-03-18 2021-09-23 Coretronic Corporation Unmanned aerial vehicle and image recognition method thereof
CN111724342A (en) * 2020-05-18 2020-09-29 天津大学 Method for detecting thyroid nodule in ultrasonic image
CN112067632A (en) * 2020-06-04 2020-12-11 西南交通大学 Power equipment detection cloud platform and detection method
CN113112501A (en) * 2021-05-11 2021-07-13 上海市东方海事工程技术有限公司 Vehicle-mounted track inspection device and method based on deep learning
CN114384075A (en) * 2021-12-06 2022-04-22 西安理工大学 Track slab defect online detection system and detection method based on three-dimensional laser scanning
CN114693617A (en) * 2022-03-17 2022-07-01 中山大学 Method, system, equipment and storage medium for detecting defects of railway steel rails
CN115731210A (en) * 2022-11-29 2023-03-03 国网浙江省电力有限公司信息通信分公司 Small sample target detection method based on multi-angle information and related equipment
CN115937492A (en) * 2022-12-12 2023-04-07 国网四川省电力公司电力科学研究院 Transformer equipment infrared image identification method based on feature identification
CN116228780A (en) * 2023-05-10 2023-06-06 江苏森标科技有限公司 Silicon wafer defect detection method and system based on computer vision

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
JIA, Z 等: "Fabric defect detection based on transfer learning and imoproved faster R-CNN", 《JOURNAL OF ENGINEERED FIBERS AND FABRICS》, vol. 17, 6 July 2022 (2022-07-06), pages 1177 *
吴铭权 等: "改进型CenterNet的高铁无砟轨道板表面裂缝检测算法", 《电子测量技术》, vol. 46, no. 10, 17 April 2023 (2023-04-17), pages 123 - 128 *
韦若禹 等: "基于改进YOLO V3算法的轨道扣件缺陷检测", 《铁道标准设计》, vol. 64, no. 12, 25 May 2020 (2020-05-25), pages 30 - 36 *

Also Published As

Publication number Publication date
CN117705816B (en) 2024-04-26

Similar Documents

Publication Publication Date Title
US20230014874A1 (en) Obstacle detection method and apparatus, computer device, and storage medium
CN113378686B (en) Two-stage remote sensing target detection method based on target center point estimation
CN112528979B (en) Transformer substation inspection robot obstacle distinguishing method and system
KR102346676B1 (en) Method for creating damage figure using the deep learning-based damage image classification of facility
CN114596500A (en) Remote sensing image semantic segmentation method based on channel-space attention and DeeplabV3plus
CN111462140B (en) Real-time image instance segmentation method based on block stitching
CN115439458A (en) Industrial image defect target detection algorithm based on depth map attention
CN110909657A (en) Method for identifying apparent tunnel disease image
CN116152863B (en) Personnel information identification method and device, electronic equipment and storage medium
CN115995056A (en) Automatic bridge disease identification method based on deep learning
CN115457044B (en) Pavement crack segmentation method based on class activation mapping
CN114972258A (en) Battery surface defect detection method and system based on machine vision and related equipment
CN113111875A (en) Seamless steel rail weld defect identification device and method based on deep learning
CN116229292A (en) Inspection system and method based on unmanned aerial vehicle road surface inspection disease
CN115719475A (en) Three-stage trackside equipment fault automatic detection method based on deep learning
CN116824399A (en) Pavement crack identification method based on improved YOLOv5 neural network
CN113763423B (en) Systematic target recognition and tracking method based on multi-mode data
CN117705816B (en) Unmanned aerial vehicle-based steel rail surface defect detection method, system, equipment and medium
CN112528903A (en) Face image acquisition method and device, electronic equipment and medium
CN116935174A (en) Multi-mode fusion method and system for detecting surface defects of metal workpiece
CN116091976A (en) Station room defect identification detection method, system, device and storage medium
CN114463685B (en) Behavior recognition method, behavior recognition device, electronic equipment and storage medium
CN115984647A (en) Remote sensing distributed collaborative reasoning method, device, medium and satellite for constellation
CN115984646A (en) Distributed target detection method and device for remote sensing cross-satellite observation and satellite
CN114627371A (en) Bridge health monitoring method based on attention mechanism

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant