WO2017113805A1 - 列车车号和车型识别方法和系统及安全检查方法和系统 - Google Patents

列车车号和车型识别方法和系统及安全检查方法和系统 Download PDF

Info

Publication number
WO2017113805A1
WO2017113805A1 PCT/CN2016/094207 CN2016094207W WO2017113805A1 WO 2017113805 A1 WO2017113805 A1 WO 2017113805A1 CN 2016094207 W CN2016094207 W CN 2016094207W WO 2017113805 A1 WO2017113805 A1 WO 2017113805A1
Authority
WO
WIPO (PCT)
Prior art keywords
train
image
inspected
car
vehicle
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.)
Ceased
Application number
PCT/CN2016/094207
Other languages
English (en)
French (fr)
Inventor
许艳伟
喻卫丰
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.)
Nuctech Co Ltd
Original Assignee
Nuctech Co Ltd
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 Nuctech Co Ltd filed Critical Nuctech Co Ltd
Priority to EA201792342A priority Critical patent/EA035267B1/ru
Priority to EP16880600.8A priority patent/EP3321860B1/en
Priority to RU2017140638A priority patent/RU2682007C1/ru
Priority to PL16880600T priority patent/PL3321860T3/pl
Priority to UAA201711435A priority patent/UA122334C2/uk
Publication of WO2017113805A1 publication Critical patent/WO2017113805A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
    • G06V30/10—Character recognition
    • G06V30/16—Image preprocessing
    • G06V30/168—Smoothing or thinning of the pattern; Skeletonisation
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • 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
    • G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
    • G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
    • G06V10/751—Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T3/00—Geometric image transformations in the plane of the image
    • G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
    • G06T3/4038—Image mosaicing, e.g. composing plane images from plane sub-images
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/80—Geometric correction
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0004—Industrial image inspection
    • 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
    • G06V10/34—Smoothing or thinning of the pattern; Morphological operations; Skeletonisation
    • 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
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
    • H04N1/0035—User-machine interface; Control console
    • H04N1/00405—Output means
    • H04N1/00408—Display of information to the user, e.g. menus
    • H04N1/0044—Display of information to the user, e.g. menus for image preview or review, e.g. to help the user position a sheet
    • H04N1/00442—Simultaneous viewing of a plurality of images, e.g. using a mosaic display arrangement of thumbnails
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/20—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using video object coding
    • H04N19/23—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using video object coding with coding of regions that are present throughout a whole video segment, e.g. sprites, background or mosaic
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2200/00—Indexing scheme for image data processing or generation, in general
    • G06T2200/32—Indexing scheme for image data processing or generation, in general involving image mosaicing
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V2201/00—Indexing scheme relating to image or video recognition or understanding
    • G06V2201/06—Recognition of objects for industrial automation
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
    • G06V30/10—Character recognition
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N2209/00—Details of colour television systems
    • H04N2209/04—Picture signal generators
    • H04N2209/041—Picture signal generators using solid-state devices
    • H04N2209/042—Picture signal generators using solid-state devices having a single pick-up sensor
    • H04N2209/045—Picture signal generators using solid-state devices having a single pick-up sensor using mosaic colour filter

Definitions

  • the present invention relates to the field of train inspection, and in particular to a train number and a vehicle type identification method and system, and a train safety inspection method and system.
  • the present application discloses a train model identification method and system, and a train safety inspection method and system, which can realize automatic identification of a train model.
  • a train number identification method including: continuously taking a picture of a train being inspected by a line camera that has a relative motion with a train being inspected to generate a plurality of train sub-images; and splicing the plurality of trains The image is divided to obtain a stitched image; the stitched image is subjected to distortion correction to obtain a corrected image; and the vehicle number is identified from the corrected image.
  • Distortion correction of the stitched image includes: extracting a wheel profile from the stitched image; obtaining a ratio of horizontal and vertical diameters of the wheel from the wheel profile; if the ratio is greater than a first predetermined threshold, The stitched image is subjected to horizontal compression processing; if the ratio is less than a second predetermined threshold, the stitched image is subjected to horizontal stretching processing.
  • the image acquisition module acquires multiple generated by the line camera according to a preset photographing frequency.
  • the train is divided into images.
  • recognizing the car number from the corrected image comprises: dividing a car number region on the corrected image to obtain a car number region image; performing smoothing denoising and binarization processing on the car number region image; The character recognition engine is used to identify the car number; and the identified car number is automatically corrected according to the car number definition rule.
  • a train number identification system including: an image acquisition module, configured to acquire a plurality of trains generated by continuously taking pictures of a train under test using a line camera that moves relative to the train being inspected. a sub-image; an image splicing module for splicing the plurality of train sub-images to obtain a spliced image; an image correction module for performing distortion correction on the spliced image to obtain a corrected image; a car number recognition module for Correct the image to identify the car number.
  • Distortion correction of the stitched image by the image correction module includes: extracting a wheel profile from the stitched image; obtaining a ratio of a horizontal and vertical diameter of the wheel from the wheel profile; and if the ratio is greater than a first predetermined threshold, according to the ratio Horizontally compressing the stitched image; if the ratio is less than a second predetermined threshold, performing horizontal stretching processing on the stitched image.
  • the image acquisition module is configured to acquire a plurality of train sub-images generated by the line camera according to a preset photographing frequency.
  • the car number identification module recognizing the car number from the corrected image includes: segmenting a car number region on the corrected image to obtain a car number region image; performing smoothing denoising on the car number region image and Value processing; using the character recognition engine for car number identification; and automatically correcting the identified car number according to the car number definition rule.
  • a train vehicle identification method including: identifying a train number; and determining a vehicle type based on the train number.
  • identifying the train number includes performing vehicle number identification using the aforementioned one of the train number identification methods.
  • determining the vehicle type based on the train number includes using the train number to perform a vehicle model lookup from a database or data table.
  • determining the vehicle type based on the train number includes directly determining the train model from the train number based on the train number definition rule.
  • a train vehicle identification system including: a vehicle number identification module for identifying a train number; and a vehicle type determination module for performing vehicle type determination based on a train number.
  • the car number identification module is any one of the aforementioned train car number identification systems.
  • determining the vehicle type based on the train number includes using the train number to perform a vehicle model lookup from a database or data table.
  • determining the vehicle type based on the train number includes directly determining the train model from the train number based on the train number definition rule.
  • the train model is one of a locomotive, a passenger car, and a truck.
  • the train vehicle identification system further includes: a behavior recognition module for identifying a location of the train; and an image segmentation module for dividing the train image into a plurality of train sub-images according to the location of the train.
  • a train safety inspection method including: utilizing any of the foregoing train vehicles
  • the type identification method identifies the vehicle type of the train to be inspected entering the inspection area; if the model of the train to be inspected is a locomotive or a passenger vehicle, the train to be inspected is irradiated with a low dose or the train to be inspected is not irradiated; if the model of the train to be inspected is a truck, The train to be inspected is illuminated at a high dose.
  • a train safety inspection system comprising: any one of the foregoing train vehicle identification systems; and a radiation control module for determining a vehicle type of the train to be inspected by using the train vehicle identification system.
  • the control ray source illuminates the train to be inspected with the first dose or does not illuminate the train to be inspected.
  • the radiation source is controlled to be irradiated with the second dose. The train is inspected, wherein the first dose is less than the second dose.
  • train model identification method and system and the train safety inspection method and system of the present disclosure it is possible to realize automatic identification of train models and safety inspection of trains, and has the advantages of high efficiency, good usability, and the like. Moreover, automatic recognition of the modified vehicle can be achieved.
  • FIG. 1A schematically illustrates a line camera that can be used in a vehicle number and a vehicle type identification device according to some example embodiments of the present disclosure
  • FIG. 1B illustrates a schematic diagram of car number and vehicle type identification in accordance with some example embodiments of the present disclosure
  • FIG. 2 illustrates a vehicle number identification method in accordance with some embodiments of the present disclosure
  • FIG. 3 illustrates a car number identification system in accordance with some embodiments of the present disclosure
  • FIG. 5 illustrates a vehicle type identification system in accordance with some embodiments of the present disclosure
  • FIG. 6 illustrates a train safety inspection method in accordance with some embodiments of the present disclosure
  • FIG. 7 illustrates a train safety inspection system in accordance with some embodiments of the present disclosure.
  • the present disclosure provides a real-time plotting system and a security check system and method for large-scale targets, which enable security personnel to understand the progress of the scan through real-time images, and can also make preliminary judgments on the inspected objects through real-time images.
  • FIG. 1A schematically illustrates a line camera 120 that may be used in a vehicle number and vehicle type identification device in accordance with some example embodiments of the present disclosure.
  • FIG. 1B illustrates a schematic diagram of vehicle number and vehicle type identification in accordance with some example embodiments of the present disclosure.
  • the train 110 can be photographed using the line camera 120 for car number and vehicle type identification.
  • the train image obtained by the line camera 120 can be used for vehicle type recognition.
  • FIG. 2 illustrates a vehicle number identification method in accordance with some embodiments of the present disclosure.
  • the car number identification method continuously photographs the train to be inspected using a line camera that moves relative to the train being inspected to generate a plurality of train sub-images.
  • a line camera can be placed on the side of the train to obtain a side image of the train.
  • the present disclosure is not limited thereto, and for example, a line camera may be disposed above the train to acquire a bird's-eye view image of the train, as needed.
  • a line camera can be placed in multiple orientations to obtain train images from multiple orientations.
  • the vehicle number referred to in the present disclosure should be understood broadly, and is not limited to only the official train number.
  • the car number here may also be a mark for train identification at any part of the vehicle body.
  • the plurality of train sub-images are spliced to obtain a spliced image.
  • a plurality of train sub-images may be stitched using a computer image processing system to obtain a train image.
  • the image can be processed during image stitching. It is also possible to process the image according to the situation after the splicing is completed.
  • the stitched image obtained by the stitching is subjected to distortion correction.
  • Distortion of the train image will affect the car number identification.
  • the distortion of the train image is related to the speed of the car and the camera's camera frequency.
  • the line camera can be controlled to continuously take a picture of the train under test according to the photographing frequency calculated according to the relative speed of the train being inspected.
  • the ratio of the relative speed of the train being inspected to the number of train sub-images for that time period is determined by the width of the object identified per second per imaging element of the line camera.
  • the line camera have a focal length f of 35 mm and an imaging element width d (the width of a pixel is usually referred to in a line camera.
  • the width of the imaging element corresponds to one pixel.
  • the width of the line is 14 ⁇ m
  • the object distance h is 2.5 m
  • the multiple h/f 7143 times.
  • the photographing frequency of the line camera is 10 kHz
  • the line camera set in this example can obtain an image in the same proportion as the real object at the photographing frequency (ie, uncompressed or stretched). If the train speed is 18km/h and the photographing area is passed, the original 10,000-point image/s is adjusted to 5000-minute image/s by the algorithm, then the relative speed of the train and the time in each time period.
  • the stitched train image can be an undistorted train image.
  • the speed of the train can be directly measured by setting a speed sensor such as a speed measuring radar near the line camera.
  • the position of the train passing through the two position sensors can also be determined by two position sensors and/or photoelectric switches and/or electronic light curtains, and the train speed can be determined based on the distance between the two position sensors.
  • the line camera may be controlled to take a photo immediately or take a certain time delay.
  • a position sensor such as a photoelectric switch or an electronic light curtain blocks the sensor after the arrival of the train, so that the sensor senses the arrival of the train.
  • the speed sensor such as a speed radar
  • the speed of the train can also be fed back in real time, and the frequency of the train image of the train taken by the line camera can be adjusted by the speed of the feedback, so that the train image generation frequency in each time period of the line camera is equal to the train average speed in the time period. In proportion to avoid image distortion.
  • the sensor can be in close proximity to the line camera. When the train reaches the sensing range of the sensor, the sensor sends a command in real time to inform the line camera to take a picture immediately. It is also possible to set the sensor to a predetermined distance in front of the line camera. When the sensor detects the arrival of the train, the line camera is notified to take a picture immediately, or the line camera is notified to take a picture of the train after a certain delay.
  • the line camera can also be controlled to continuously take photos of the train to be inspected according to a preset photographing frequency (for example, if the photographing frequency of the line camera is preset to 10 kHz, that is, 10,000 images/s, then the mileage is extremely high.
  • a preset photographing frequency for example, if the photographing frequency of the line camera is preset to 10 kHz, that is, 10,000 images/s, then the mileage is extremely high.
  • One train produces one train sub-image
  • a continuous time period may be set, and according to the relative speed of the train to be inspected for each time period, the number of train sub-images of the time period is adjusted based on the principle described above, so that the relative speed of the train to be inspected for each time period The ratio of the number of train sub-images in this time period is kept consistent.
  • At least one train minute image may be extracted from the train sub-image obtained in the time period according to a predetermined rule. If the relative speed of the train to be inspected is higher than the relative speed corresponding to the preset photographing frequency, the interpolation method may be used to add at least one train minute image to the train sub-image obtained in the time period.
  • the time period may be calculated according to the number of train sub-images and/or the train speed setting or artificial setting, and may be 1 s or 10 s.
  • more than one train sub-image is typically produced over a period of time.
  • the train speed in each time period may be the average speed of the train during that time period or the speed of the train at the beginning or end of the time period.
  • the maximum number of trains can be obtained for the maximum camera frequency of the line camera during the train detection time (ie, the time the entire train passes). If the speed of the train being inspected is below the maximum speed for a certain period of time, then The magnitude of the difference between the actual speed and the maximum speed adopts the extraction method to reduce the number of train sub-images in the time period, so that the train speed in each time period is roughly proportional to the number of train sub-images generated to avoid the train. Image distortion caused by speed.
  • the maximum speed of the train is 30km/h during the detection time, corresponding to the photographing frequency of 50 images/s (that is, the frequency at which the train sub-image is generated), if the measured speed of a certain period of time is 24km/h. Then, the corresponding number of images in the time period becomes 40 images/s, that is, one image is taken out every 5 images.
  • the image can be extracted according to a preset rule, for example, the third of every 5 images is taken out. In this way, it is ensured that the train speed is proportional to the number of generated train sub-images, so that the train image generated by the train sub-images obtained by each line period of the line camera is not distorted.
  • the average camera frequency of a line camera can be taken to the train, corresponding to the most frequently occurring train speed during the test time, or the average train speed. If the train speed in a certain period of time is higher than the most frequently occurring train speed, or the average train speed, a method of complementing the value is adopted for the purpose of not distorting, for example, by successive two positions of the appropriate position during the time period.
  • the method of image fitting and smoothing complements the two train sub-images to form a new train sub-image.
  • the image thus obtained has a resolution that is not as good as the train image obtained by photographing, but the distance relationship in the image and the train contour are equal to the actual train condition.
  • the number of train sub-images is reduced by the above-mentioned extraction method. For example, if the average speed of the train is 30km/h, corresponding to the photographing frequency of 50 images/s (that is, the frequency at which the train sub-image is generated), if the vehicle speed of a certain period of time is 36km/h, the corresponding The number of images in the time period becomes 60 images/s, that is, one image is added every 5 images. For example, a newly added train image obtained by fitting or image averaging between the second train sub-image and the third train sub-image in each time period can be used to ensure the train speed and the generated train points. The number of images is proportional.
  • the stitched image can be corrected for distortion using various methods.
  • image correction is performed using a wheel profile as a reference.
  • This correction method is simple and effective, which can improve processing efficiency and reduce processing cost.
  • the wheel profile can be extracted from the stitched image.
  • the ratio of the horizontal and vertical diameters of the wheel is then obtained from the wheel profile. If the ratio is greater than the first predetermined threshold, the stitched image is horizontally compressed according to the ratio; if the ratio is less than the second predetermined threshold, the stitched image is horizontally stretched. Finally, a less distorted train image can be obtained to facilitate subsequent operations.
  • car number identification is performed.
  • the vehicle number is identified by the conventional car number identification method using the obtained train image.
  • the car number area can be segmented on the train image, and then the car number area image is smoothed and denoised and binarized, and sent to the character recognition engine for car number identification.
  • the car number recognized by the character recognition engine can be automatically corrected according to the car number definition rule. Since the conventional car number identification method is well known to those skilled in the art, the details thereof will not be described again.
  • the location of the train may be identified in the obtained train image, and then the train image is divided into a plurality of train sub-images according to the location of the train.
  • the train number identifying method according to the present disclosure has been described above. The following describes the train number that can realize the above method. Do not system.
  • a train number identification system may include an image acquisition module 502, an image stitching module 504, a distortion correction module 506, and a vehicle number identification module 508.
  • the image acquisition module 502 is configured to acquire a plurality of train sub-images that are continuously photographed by the line camera that is in relative motion with the train to be inspected.
  • the system may further comprise a camera control module 512 for continuously taking pictures of the train under test based on the camera frequency calculated from the relative speed of the train being inspected.
  • the image acquisition module 502 acquires a plurality of train sub-images generated by the line camera according to a preset photographing frequency.
  • the image splicing module 504 is configured to splicing the plurality of train sub-images to obtain a spliced image.
  • the image splicing module 504 is further configured to set a continuous time period, and adjust the number of train sub-images of the time period according to the relative speed of the detected trains in each time period, so that each time period is The relative speed of the detected train is consistent with the ratio of the number of train sub-images of the time period. If the relative speed of the train to be inspected is lower than the relative speed corresponding to the preset photographing frequency, at least one train minute image is extracted from the train sub-image obtained from the time period according to a predetermined rule. If the relative speed of the train to be inspected is higher than the relative speed corresponding to the preset photographing frequency, at least one train minute image is added to the train sub-image obtained during the time period by interpolation.
  • the distortion correction module 506 is configured to perform distortion correction on the stitched image.
  • distortion correction module 506 can be used to extract a wheel profile from the stitched image and obtain a ratio of wheel horizontal and vertical diameters from the wheel profile. If the ratio is greater than the first predetermined threshold, the stitched image is horizontally compressed according to the ratio. If the ratio is less than a second predetermined threshold, the stitched image is subjected to a horizontal stretching process.
  • the car number identification module 508 is used to identify the car number from the stitched image, and will not be described again.
  • Model identification can be performed based on the car number.
  • Train type identification has many applications in practice, such as in the field of train safety inspection or train maintenance.
  • In the field of safety inspection of trains it is necessary to distinguish different vehicle models to determine whether the vehicle to be inspected is loaded, thereby setting different X-ray doses, or only performing X-ray scanning on the cargo compartment without scanning the passenger compartment. This requires first identifying the vehicle to determine if the vehicle entering the inspection area is a manned locomotive or passenger car.
  • Train models can be roughly classified into locomotives, buses, trucks, etc., and trucks are divided into container trucks, vans, tankers, and carts. Different models have different wheelbases, heights, and types of objects to be carried (for example, one or more of the objects transported by different models may be human, solid cargo, liquid cargo, etc.).
  • a vehicle identification method for a train a plurality of train detection points are arranged along the rail, magnetic steel is set on the track of each train detection point, and the relative speed of the train and the position of the train axle are detected by the magnetic steel to determine the wheelbase.
  • Passenger cars and trucks are distinguished by the difference in wheelbase.
  • the standards of train cars vary from country to country, so that the trains of the same model have different wheelbases.
  • the method of detecting the wheelbase cannot identify the same type of car in different countries.
  • the car modification (such as changing the passenger car to the car, changing the passenger to the cargo) causes the carrying object to change, but the axle is usually unchanged, which makes the detection axle unable to distinguish the carrier of the car.
  • FIG. 4 illustrates a vehicle type identification method according to an embodiment of the present disclosure.
  • the embodiment shown in Fig. 4 utilizes the method shown in Fig. 2, mainly by adding a vehicle type judging operation S210. Only the operation of S210 will be described below.
  • the vehicle type search is performed in the database or the data table using the identified vehicle number.
  • the correspondence between the car number and the model can be recorded in the database or the data table, and the train model can be determined by searching by using the car number.
  • the vehicle number can be directly judged from the vehicle number according to the vehicle number definition rule.
  • the car number definition rule can be the first two characters representing the model.
  • vehicle type identification method according to the present disclosure is not limited to the vehicle number obtained by the method according to the present disclosure.
  • FIG. 5 illustrates a vehicle type identification system in accordance with an embodiment of the present disclosure.
  • the embodiment shown in Fig. 5 utilizes the system shown in Fig. 3, primarily with the addition of the vehicle type determination module 510.
  • the vehicle type determination module 510 can perform vehicle type search in a database or a data table using the identified vehicle number. For example, the correspondence between the car number and the model can be recorded in the database or the data table, and the train model can be determined by searching by using the car number. In a variant embodiment, the vehicle type determination module 510 may not have to perform a search, but may directly determine the vehicle type from the vehicle number according to the vehicle number definition rule.
  • the system may further include an activity recognition module 514 and an image segmentation module 516.
  • the activity identification module 514 is used to identify the location of the train.
  • the image segmentation module 516 is configured to divide the train image into a plurality of train sub-images according to the location of the tour.
  • vehicle type identification system is not limited to the use of the vehicle number recognition system according to the present disclosure.
  • the train safety inspection can be realized, as shown in FIG. 6.
  • step 602 the vehicle type of the train to be inspected entering the inspection area is identified by the aforementioned train vehicle identification method.
  • a corresponding illumination check is performed based on the vehicle type. If the model of the train to be inspected is a locomotive or a passenger car, the train to be inspected is irradiated with a low dose or the train to be inspected is not irradiated; if the model of the train to be inspected is a truck, the train to be inspected is irradiated with a high dose.
  • FIG. 7 illustrates a train safety inspection system that implements the aforementioned train safety inspection method, in accordance with some embodiments of the present disclosure.
  • the train safety inspection system may include the aforementioned train vehicle identification system 702 and radiation control module 704.
  • the radiation control module 704 is configured to determine the vehicle type of the train to be inspected by using the train vehicle identification system, and control the radiation source to illuminate the train to be inspected with the first dose or not to be inspected when the model of the train to be inspected is a locomotive or a passenger vehicle.
  • the control radiation source illuminates the train to be inspected with a second dose, wherein the first dose is smaller than the second dose.
  • the train model identification system 702 is as described above and will not be described again here.
  • the wheel profile is used for distortion correction, and the car number is recognized, and the operation is simple and quick.
  • Modules can be implemented by software or by partial software hardening (for example, via an FPGA). Therefore, the technical solution of the embodiments of the present disclosure may be embodied in the form of a software product, which may be stored in a non-volatile storage medium (which may be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including a plurality of instructions.
  • a computing device which may be a personal computer, server, mobile terminal, or network device, etc.
  • modules may be distributed in the device according to the description of the embodiments, or the corresponding changes may be located in one or more devices different from the embodiment.
  • the modules of the above embodiments may be combined into one module, or may be further split into multiple sub-modules.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Signal Processing (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Quality & Reliability (AREA)
  • Human Computer Interaction (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Health & Medical Sciences (AREA)
  • Evolutionary Biology (AREA)
  • Databases & Information Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Train Traffic Observation, Control, And Security (AREA)
  • Image Processing (AREA)
  • Image Analysis (AREA)

Abstract

列车车号和车型识别方法和系统及列车安全检查方法及系统。一种列车车号识别方法包括：利用与被检列车发生相对运动的线阵相机对被检列车连续拍照从而产生多个列车分图像（S202）；拼接所述多个列车分图像从而得到拼接图像（S204）；对拼接图像进行失真校正，从而得到校正图像（S206）；从所述校正图像识别车号（S208）。对拼接图像进行失真校正包括：从所述拼接图像提取车轮轮廓；从所述车轮轮廓获得车轮水平和竖直直径的比值；如果所述比值大于第一预定阈值，则根据所述比值对所述拼接图像进行水平压缩处理；如果所述比值小于第二预定阈值，则对所述拼接图像进行水平拉伸处理。根据本方法及系统能够实现列车车型的自动识别和列车的安全检查，效率高，易用性好。

Description

列车车号和车型识别方法和系统及安全检查方法和系统
本申请基于申请号为201511016979.2、申请日为2015年12月29日的中国专利申请提出，并要求该中国专利申请的优先权，该中国专利申请的全部内容在此引入本申请作为参考。
技术领域
本发明涉及列车检查领域，具体而言，涉及列车车号和车型识别方法和系统及列车安全检查方法及系统。
背景技术
在列车运输管理中，经常需要记录、检查、核对车号。人工进行车号处理是一件费时费力且容易出错的工作。一种自动处理方式是采用RFID(射频识别)标签，但该方式成本较高，且有些情况下无法使用。另一种处理方式是利用数字图像处理进行车号自动识别。数字图像处理由于可通过复杂的算法进行识别运算、处理精度高，目前已成为常规的识别方式。通过数字图像处理识别车号需要获取车身(侧面)图像。如果图像存在失真，则会对车号识别的准确度造成影响。
因此，需要一种能够判断并消除车身图像失真的车号识别方法。
在所述背景技术部分公开的上述信息仅用于加强对本公开的背景的理解，因此它可以包括不构成对本领域普通技术人员已知的现有技术的信息。
发明内容
本申请公开一种列车车型识别方法和系统及列车安全检查方法及系统，能够实现列车车型的自动识别。
本公开的其他特性和优点将通过下面的详细描述变得显然，或部分地通过本公开的实践而习得。
根据本公开的一个方面，提供一种列车车号识别方法，包括：利用与被检列车发生相对运动的线阵相机对被检列车连续拍照从而产生多个列车分图像；拼接所述多个列车分图像从而得到拼接图像；对拼接图像进行失真校正，从而得到校正图像；从所述校正图像识别车号。对拼接图像进行失真校正包括：从所述拼接图像提取车轮轮廓；从所述车轮轮廓获得车轮水平和竖直直径的比值；如果所述比值大于第一预定阈值，则根据所述比值对所述拼接图像进行水平压缩处理；如果所述比值小于第二预定阈值，则对所述拼接图像进行水平拉伸处理。
根据一些实施例，图像获取模块获取所述线阵相机按照预设的拍照频率而产生的多个 列车分图像。
根据一些实施例，从所述校正图像识别车号包括：在所述校正图像上分割出车号区域从而得到车号区域图像；对所述车号区域图像进行平滑去噪和二值化处理；利用字符识别引擎进行车号识别；及根据车号定义规则对识别出的车号进行自动修正。
根据本公开的另一方面，提供一种列车车号识别系统，包括：图像获取模块，用于获取利用与被检列车发生相对运动的线阵相机对被检列车连续拍照而产生的多个列车分图像；图像拼接模块，用于拼接所述多个列车分图像从而得到拼接图像；图像校正模块，用于对拼接图像进行失真校正，从而得到校正图像；车号识别模块，用于从所述校正图像识别车号。图像校正模块对拼接图像进行失真校正包括：从所述拼接图像提取车轮轮廓；从所述车轮轮廓获得车轮水平和竖直直径的比值；如果所述比值大于第一预定阈值，则根据所述比值对所述拼接图像进行水平压缩处理；如果所述比值小于第二预定阈值，则对所述拼接图像进行水平拉伸处理。
根据一些实施例，图像获取模块用于获取所述线阵相机按照预设的拍照频率而产生的多个列车分图像。
根据一些实施例，车号识别模块从所述校正图像识别车号包括：在所述校正图像上分割出车号区域从而得到车号区域图像；对所述车号区域图像进行平滑去噪和二值化处理；利用字符识别引擎进行车号识别；及根据车号定义规则对识别出的车号进行自动修正。
根据本公开的另一方面，提供一种列车车型识别方法，包括：识别列车车号；根据列车车号进行车型判断。
根据一些实施例，识别列车车号包括利用前述一项列车车号识别方法进行车号识别。
根据一些实施例，根据列车车号进行车型判断包括利用列车车号从数据库或数据表中进行车型查找。
根据一些实施例，根据列车车号进行车型判断包括根据列车车号定义规则从所述列车车号直接判断列车车型。
根据本公开的另一方面，提供一种列车车型识别系统，包括：车号识别模块，用于识别列车车号；车型判断模块，用于根据列车车号进行车型判断。
根据一些实施例，所述车号识别模块为前述任一项列车车号识别系统。
根据一些实施例，根据列车车号进行车型判断包括利用列车车号从数据库或数据表中进行车型查找。
根据一些实施例，根据列车车号进行车型判断包括根据列车车号定义规则从所述列车车号直接判断列车车型。
根据一些实施例，所述列车车型为机车、客车和货车中的一种。
根据一些实施例，列车车型识别系统还包括：勾当识别模块，用于识别列车的勾当位置；图像切分模块，用于将所述列车图像按勾当位置分为多张列车子图像。
根据本公开的另一方面，提供一种列车安全检查方法，包括：利用前述任一项列车车 型识别方法识别进入检查区域的被检列车的车型；如果被检列车的车型是机车或客车，则以低剂量照射被检列车或不照射被检列车；如果被检列车的车型是货车，则以高剂量照射被检列车。
根据本公开的另一方面，提供一种列车安全检查系统，包括：前述任一项列车车型识别系统；辐射控制模块，用于利用所述列车车型识别系统判断被检列车的车型，在被检列车的车型是机车或客车的情况下，控制射线源以第一剂量照射被检列车或不照射被检列车，在被检列车的车型是货车的情况下，控制射线源以第二剂量照射被检列车，其中第一剂量小于第二剂量。
根据本公开的列车车型识别方法和系统及列车安全检查方法及系统，能够实现列车车型的自动识别和列车的安全检查，具有效率高、易用性好等优点。而且，能够实现对改装车的自动识别。
附图说明
通过参照附图详细描述其示例实施方式，本公开的上述和其它特征及优点将变得更加明显。
图1A示意性示出可用于根据本公开一些示例实施方式的车号及车型识别装置的线阵相机；
图1B示出根据本公开一些示例实施方式进行车号及车型识别的示意图；
图2示出根据本公开一些实施方式的车号识别方法；
图3示出根据本公开一些实施方式的车号识别系统；
图4示出根据本公开一些实施方式的车型识别方法；
图5示出根据本公开一些实施方式的车型识别系统；
图6示出根据本公开一些实施方式的列车安全检查方法；及
图7示出根据本公开一些实施方式的列车安全检查系统。
具体实施方式
现在将参考附图更全面地描述示例实施例。然而，示例实施例能够以多种形式实施，且不应被理解为限于在此阐述的实施例；相反，提供这些实施例使得本公开将全面和完整，并将示例实施例的构思全面地传达给本领域的技术人员。在图中相同的附图标记表示相同或类似的部分，因而将省略对它们的重复描述。
此外，所描述的特征、结构或特性可以以任何合适的方式结合在一个或更多实施例中。在下面的描述中，提供许多具体细节从而给出对本公开的实施例的充分理解。然而，本领域技术人员将意识到，可以实践本公开的技术方案而没有所述特定细节中的一个或更多，或者可以采用其它的方法、组元、材料、装置、步骤等。在其它情况下，不详细示出或描述公知结构、方法、装置、实现、材料或者操作以避免模糊本公开的各方面。
附图中所示的方框图仅仅是功能实体，不一定必须与物理上独立的实体相对应。即，可以采用软件形式来实现这些功能实体，或在一个或多个软件硬化的模块中实现这些功能实体或功能实体的一部分，或在不同网络和/或处理器装置和/或微控制器装置中实现这些功能实体。
本公开提供一种用于大型目标的实时出图系统和安全检查系统及方法，可以使安检人员通过实时图像了解扫描进度，并也可以通过实时图像对被检查目标进行初步判断。
图1A示意性示出可用于根据本公开一些示例实施方式的车号及车型识别装置的线阵相机120。图1B示出根据本公开一些示例实施方式进行车号及车型识别的示意图。
如图1A和1B所示，可利用线阵相机120对列车110进行拍照，以用于车号及车型识别。
根据本公开的车号及车型识别方法，可利用线阵相机120获得的列车图像进行车型识别。
图2示出根据本公开一些实施方式的车号识别方法。
参见图2，在S202，根据本公开的车号识别方法利用与被检列车发生相对运动的线阵相机对被检列车连续拍照从而产生多个列车分图像。线阵相机的原理和使用已是公知，在此不再赘述。在利用线阵相机对被检列车拍照时，线阵相机可设置在列车侧方以获取列车侧面图像。但本公开不限于此，例如，根据需要，线阵相机也可设置在列车上方，以获取列车的俯视图像。或者，也可以在多个方位设置线阵相机以从多个方位获取列车图像。另外，本公开中所称的车号应进行广义理解，而不是仅仅局限于正式的列车编号。例如，这里的车号也可以是位于车身任何部位的用于进行列车识别的标记。
在S204，拼接所述多个列车分图像从而得到拼接图像。例如，可以利用计算机图像处理系统对多个列车分图像进行拼接以获取列车图像。如后面所描述的，在图像拼接过程中，可以对图像进行处理。也可以在拼接全部完成之后，再根据情况对图像进行处理。
在S206，对拼接得到的拼接图像进行失真校正。
列车图像的失真会影响车号识别。而列车图像的失真与车速及相机的拍照频率相关。例如，可以控制线阵相机按照根据被检列车的相对速度计算的拍照频率对所述被检列车进行连续拍照。
易于理解，每个时间段被检列车的相对速度与该时间段的列车分图像的数量的比例由线阵相机的每个成像元件每秒识别的物体宽度确定。
例如，设线阵相机的焦距f为35mm，成像元件宽d(在线阵相机中通常指一个像素点的宽度，当然如果线阵相机使用了n个像素，成像元件的宽度相应为一个像素点的宽度的n倍)为14μm，物距h为2.5m，物宽(即每个成像元件识别的物体宽度)W＝d*h/f＝1mm，则此线阵相机在这种情况下的放大倍数h/f＝7143倍。设线阵相机的拍照频率为10kHz，而每个成像元件每秒识别物宽为D＝10000张×1mm＝10m，即每个成像元件识别速度为10m/s＝36km/h。此时，每个时间段内列车的相对速度与该时间段内产生的列车分图像的数 量的比值为36km/h÷10000张/s＝1mm/张(此处取时间段为1s)，该比值即为上述每个成像元件识别的物体宽度w。若列车以速度36km/h经过拍照区域，则以该例设置的线阵相机以该拍照频率正好能得到与实物同比例的图像(即未经压缩或拉伸)。如果此时列车速度为18km/h驶过拍照区域，则通过算法将原来的1万张分图像/s调整为5000张分图像/s，则时每个时间段内列车的相对速度与该时间段内产生的列车分图像的数量的比值为为18km/h÷5000张/s＝1mm/张(此处取时间段为1s)，也为上述相同的比值。由上述理论推导可知，可以通过实际测量的列车速度及上述比值，确定实际使用的列车分图像的数量。这样，拼接成的列车图像可为不失真的列车图像。
列车车速的测量有多种方式，可以通过设置在线阵相机附近的测速雷达等速度传感器直接测定列车车速。也可以通过两个地感线圈和/或光电开关和/或电子光幕等位置传感器测定列车经过该两个位置传感器的时刻，并基于该两个位置传感器的距离测定列车车速。根据一些实施方式，还可以在通过雷达或传感器检测到被检列车后，控制所述线阵相机立即拍照或经过一定延时拍照。例如，通过光电开关或电子光幕等位置传感器，在列车到来后阻断传感器，使得传感器感知列车的到来。或者，通过测速雷达等速度传感器，即可知道列车的到来。也可实时反馈列车的速度，通过反馈的速度调整线阵相机拍照的列车分图像的产生频率，使得线阵相机每一时间段内的列车分图像产生频率与该时间段内的列车平均速度成正比，避免图像失真。传感器可以与线阵相机紧邻，在列车到达传感器的感知范围时由传感器实时发出指令通知线阵相机立即拍照。也可以将传感器设置在线阵相机前的预定距离，当传感器检测到列车到达，通知线阵相机立即拍照，或通知线阵相机经一定延时之后再对列车拍照。
易于理解，也可以控制线阵相机按照预设的拍照频率对所述被检列车进行连续拍照(例如，如果线阵相机的拍照频率预设为10kHz，即1万张图像/s，则万分之一秒产生1张列车分图像)，并对分图像数量进行调整。例如，可以设定连续的时间段，根据每个时间段被检列车的相对速度，基于前面描述的原理，调整该时间段的列车分图像的数量，使得每个时间段被检列车的相对速度与该时间段的列车分图像的数量的比例保持一致。
如果被检列车的相对速度低于对应于预设的拍照频率的相对速度，则可从该时间段获得的列车分图像中按照预定规则抽掉至少一个列车分图像。如果被检列车的相对速度高于对应于预设的拍照频率的相对速度，则可利用插值法在该时间段获得的列车分图像中增加至少一个列车分图像。
例如，时间段可根据列车分图像的数量和/或列车速度计算设定或人为设定，可以是1s，也可以是10s。一般地，列车速度越快、线阵相机拍照频率越高，则时间段设定得越小。但在一个时间段内一般产生多于1张列车分图像。每个时间段内的列车速度可以是该时间段内列车的平均速度或列车在该时间段内的起始时刻或终了时刻的速度。
例如，可以线阵相机的最大拍照频率在列车检测时间内(即检测整个列车通过的时间)获得最多数量的列车分图像。如果被检列车在某个时间段内的速度低于最大速度，则根据 其实际速度与最大速度的差值的大小采取抽值法减少该时间段内列车分图像的数量，使得每个时间段内列车速度与产生的列车分图像的数量大致成正比，以避免由列车速度引起的图像失真。例如，如果在检测时间内列车的最大时速为30km/h，对应于50张图像/s的拍照频率(即产生列车分图像的频率)，则如果测得某个时间段的车速为24km/h，则相应的该时间段内的图像数量变为40张图像/s，即每5张图像抽掉1张图像。图像可以依据预先设定的规则抽掉，例如，将每5张图像中的第3张抽掉。这样，就保证了列车车速与产生的列车分图像的数量成正比，从而使得由线阵相机每个时间段得到的列车分图像拼接而产生的列车图像不失真。
例如，可以线阵相机的平均拍照频率对列车拍照，对应于检测时间内最常出现的列车速度，或列车平均速度。若某个时间段内的列车速度高于该最常出现的列车速度，或列车平均速度，为了不失真则采取补值的方法，例如，通过将该时间段内拍摄的适当位置的连续两张图像进行拟合和平滑化的方法在该两张列车分图像之间补值形成新的列车分图像。这样得到的图像虽然分辨率不如拍照得到的列车分图像，但图像中的距离关系和列车轮廓与实际中列车的情况是等比例的。若某个时间段内的列车速度低于该最常出现的列车速度，则通过上述抽值法减少列车分图像数量。例如，如果列车的平均时速为30km/h，对应于50张图像/s的拍照频率(即产生列车分图像的频率)，则如果测得某时间段的车速为36km/h，则相应的该时间段内的图像数量变为60张图像/s，即每5张图像增加1张图像。例如，可以将每个时间段内第2张列车分图像与第3张列车分图像之间通过拟合或图像均值法得到的新添加列车分图像，这样就保证了列车车速与产生的列车分图像的数量成正比。
可以采用各种方法对拼接图像进行失真校正。
根据本公开的示例实施例，利用车轮轮廓作为基准进行图像校正。这种校正方式简单有效，可以提高处理效率，降低处理成本。
例如，得到拼接图像之后，可以从所述拼接图像提取车轮轮廓。然后，从所述车轮轮廓获得车轮水平和竖直直径的比值。如果所述比值大于第一预定阈值，则根据所述比值对所述拼接图像进行水平压缩处理；如果所述比值小于第二预定阈值，则对所述拼接图像进行水平拉伸处理。最终，可以得到失真较小的列车图像以利于进行后续操作。
在S208，进行车号识别。利用得到的列车图像通过常规的车号识别方法进行车号识别。例如，可在列车图像上分割出车号区域，然后对车号区域图像进行平滑去噪和二值化处理等操作，并送入字符识别引擎进行车号识别。最后，可对字符识别引擎识别出的车号根据车号定义规则进行自动修正。由于常规的车号识别方法为本领域人员所熟知，其具体内容不再赘述。
为了便于操作员检视图像，根据一些实施方式，可以在得到的列车图像中识别列车的勾当位置，然后将所述列车图像按勾当位置分为多张列车子图像。
以上描述了根据本公开的列车车号识别方法。下面描述可实现上述方法的列车车号识 别系统。
如图3所示，根据本公开一些实施方式的列车车号识别系统可包括图像获取模块502、图像拼接模块504、失真校正模块506、车号识别模块508。
图像获取模块502用于获取利用与被检列车发生相对运动的线阵相机对被检列车连续拍照而产生的多个列车分图像。
在变型实施例中，系统还可包括拍照控制模块512，用于根据被检列车的相对速度计算的拍照频率对所述被检列车进行连续拍照。替代地，图像获取模块502获取所述线阵相机按照预设的拍照频率而产生的多个列车分图像。
图像拼接模块504用于拼接所述多个列车分图像从而得到拼接图像。
在变型实施例中，图像拼接模块504还可用于，设定连续的时间段，根据每个时间段被检列车的相对速度，调整该时间段的列车分图像的数量，使得每个时间段被检列车的相对速度与该时间段的列车分图像的数量的比例保持一致。如果被检列车的相对速度低于对应于预设的拍照频率的相对速度，则从该时间段获得的列车分图像中按照预定规则抽掉至少一个列车分图像。如果被检列车的相对速度高于对应于预设的拍照频率的相对速度，则利用插值法在该时间段获得的列车分图像中增加至少一个列车分图像。
失真校正模块506用于对拼接图像进行失真校正。
例如，失真校正模块506可用于从所述拼接图像提取车轮轮廓，并从所述车轮轮廓获得车轮水平和竖直直径的比值。如果所述比值大于第一预定阈值，则根据所述比值对所述拼接图像进行水平压缩处理。如果所述比值小于第二预定阈值，则对所述拼接图像进行水平拉伸处理。
车号识别模块508用于从拼接图像识别车号，不再赘述。
可以根据车号进行车型识别。
列车车型识别在实际中有很多应用，例如在列车的安全检查领域或列车维修领域。在列车的安全检查领域，需要分辨不同的车型以确定被检查车型是否载人，从而设定不同的X射线剂量，或者仅对载物车厢进行X射线扫描而不对载人车厢扫描。这就需要先对车型进行识别以判断进入检查区域的车型是否为载人的机车或客车。
列车车型大致可分为机车、客车、货车等类型，而货车又分为集装箱货车、厢车、油罐车、板车等。不同的车型其轴距、高度、运载对象类型(例如，不同车型运输的对象可以是人、固体货物、液体货物等其中的一种或几种)通常会有所不同。在一种对于列车的车型识别方法中，在铁轨沿线设置多个列车检测点，在每个列车检测点的轨道上设置磁钢，通过磁钢检测列车相对速度及列车车轴位置以确定轴距，通过轴距的不同对客车和货车加以识别区分。但是各个国家对于列车车厢的标准不同，使得同样车型的列车其轴距也不尽相同。因此，通过检测轴距的方法无法识别不同国家的同一类型车厢。而且，有的时候车厢改装(比如客车改为保温车，由载客变为载货)使得运载对象发生变化，但车轴通常不变，这样就使得检测车轴无法辨别车厢的运载对象。
图4示出根据本公开实施方式的车型识别方法。图4所示实施方式利用了图2所示的方法，主要是增加了车型判断操作S210。下面仅描述S210的操作。
在S210，利用识别出的车号在数据库或数据表中进行车型查找。例如，数据库或数据表中可记录车号与车型的对应关系，利用车号即可通过查找确定列车车型。在变型实施例中，可不必进行查找，而是根据车号定义规则，从车号可直接判断车型。例如，车号定义规则可以是前两个字符表示车型。
易于理解，根据本公开的车型识别方法不限于利用根据本公开的方法获得的车号。
图5示出根据本公开实施方式的车型识别系统。图5所示实施方式利用了图3所示的系统，主要是增加了车型判断模块510。
车型判断模块510可利用识别出的车号在数据库或数据表中进行车型查找。例如，数据库或数据表中可记录车号与车型的对应关系，利用车号即可通过查找确定列车车型。在变型实施例中，车型判断模块510可不必进行查找，而是根据车号定义规则，从车号可直接判断车型。
另外，在变型实施例中，系统还可包括勾当识别模块514和图像切分模块516。勾当识别模块514用于识别列车的勾当位置。图像切分模块516用于将所述列车图像按勾当位置分为多张列车子图像。
易于理解，根据本公开的车型识别系统不限于利用根据本公开的车号识别系统。
利用本公开的列车车型识别方法，可实现列车安全检查，如图6所示。
参见图6，在步骤602，利用前述的列车车型识别方法识别进入检查区域的被检列车的车型。
在步骤604，根据车型执行相应的照射检查。如果被检列车的车型是机车或客车，则以低剂量照射被检列车或不照射被检列车；如果被检列车的车型是货车，则以高剂量照射被检列车。
图7示出根据本公开的一些实施方式的列车安全检查系统，可实现前述列车安全检查方法。
如图7所示，列车安全检查系统可包括前述的列车车型识别系统702以及辐射控制模块704。辐射控制模块704用于利用所述列车车型识别系统判断被检列车的车型，在被检列车的车型是机车或客车的情况下，控制射线源以第一剂量照射被检列车或不照射被检列车，在被检列车的车型是货车的情况下，控制射线源以第二剂量照射被检列车，其中第一剂量小于第二剂量。列车车型识别系统702如前所述，此处不再赘述。
通过以上的详细描述，本领域的技术人员易于理解，根据本发明实施例的系统和方法具有以下优点中的一个或多个。
利用车轮轮廓进行失真校正，进而进行车号识别，操作简单而快捷。
利用车号进行车型判断，简单易行，识别准确率高。
通过以上的实施例的描述，本领域的技术人员易于理解，本公开实施例的方法和相应 模块可以通过软件或部分软件硬化(例如，通过FPGA)的方式来实现。因此，本公开实施例的技术方案可以以软件产品的形式体现出来，该软件产品可以存储在一个非易失性存储介质(可以是CD-ROM，U盘，移动硬盘等)中，包括若干指令用以使得一台计算设备(可以是个人计算机、服务器、移动终端、或者网络设备等)执行根据本公开实施例的方法。
本领域技术人员可以理解，附图只是示例实施例的示意图，附图中的模块或流程并不一定是实施本公开所必须的，因此不能用于限制本公开的保护范围。
本领域技术人员可以理解上述各模块可以按照实施例的描述分布于装置中，也可以进行相应变化位于不同于本实施例的一个或多个装置中。上述实施例的模块可以合并为一个模块，也可以进一步拆分成多个子模块。
以上具体地示出和描述了本公开的示例性实施例。应该理解，本公开不限于所公开的实施例，相反，本公开意图涵盖包含在所附权利要求的精神和范围内的各种修改和等效布置。

Claims (18)

  1. 一种列车车号识别方法，其特征在于，包括：
    利用与被检列车发生相对运动的线阵相机对被检列车连续拍照从而产生多个列车分图像；
    拼接所述多个列车分图像从而得到拼接图像；
    对拼接图像进行失真校正，从而得到校正图像；
    从所述校正图像识别车号，
    其中对拼接图像进行失真校正包括：
    从所述拼接图像提取车轮轮廓；
    从所述车轮轮廓获得车轮水平和竖直直径的比值；
    如果所述比值大于第一预定阈值，则根据所述比值对所述拼接图像进行水平压缩处理；如果所述比值小于第二预定阈值，则对所述拼接图像进行水平拉伸处理。
  2. 如权利要求1所述的列车车号识别方法，其中图像获取模块获取所述线阵相机按照预设的拍照频率而产生的多个列车分图像。
  3. 如权利要求1所述的列车车号识别方法，其中从所述校正图像识别车号包括：
    在所述校正图像上分割出车号区域从而得到车号区域图像；
    对所述车号区域图像进行平滑去噪和二值化处理；
    利用字符识别引擎进行车号识别；及
    根据车号定义规则对识别出的车号进行自动修正。
  4. 一种列车车号识别系统，其特征在于，包括：
    图像获取模块，用于获取利用与被检列车发生相对运动的线阵相机对被检列车连续拍照而产生的多个列车分图像；
    图像拼接模块，用于拼接所述多个列车分图像从而得到拼接图像；
    图像校正模块，用于对拼接图像进行失真校正，从而得到校正图像；
    车号识别模块，用于从所述校正图像识别车号，
    其中图像校正模块配置为：
    从所述拼接图像提取车轮轮廓；
    从所述车轮轮廓获得车轮水平和竖直直径的比值；
    如果所述比值大于第一预定阈值，则根据所述比值对所述拼接图像进行水平压缩处理；如果所述比值小于第二预定阈值，则对所述拼接图像进行水平拉伸处理。
  5. 如权利要求4所述的列车车号识别系统，其中图像获取模块用于获取所述线阵相机按照预设的拍照频率而产生的多个列车分图像。
  6. 如权利要求4所述的列车车号识别系统，其中车号识别模块配置为：
    在所述校正图像上分割出车号区域从而得到车号区域图像；
    对所述车号区域图像进行平滑去噪和二值化处理；
    利用字符识别引擎进行车号识别；及
    根据车号定义规则对识别出的车号进行自动修正。
  7. 一种列车车型识别方法，其特征在于，包括：
    识别列车车号；
    根据列车车号进行车型判断。
  8. 如权利要求7所述的列车车型识别方法，其中所述识别列车车号包括利用如权利要求1-3中任一项所述的列车车号识别方法进行车号识别。
  9. 如权利要求7所述的列车车型识别方法，其中根据列车车号进行车型判断包括利用列车车号从数据库或数据表中进行车型查找。
  10. 如权利要求7所述的列车车型识别方法，其中根据列车车号进行车型判断包括根据列车车号定义规则从所述列车车号直接判断列车车型。
  11. 一种列车车型识别系统，其特征在于，包括：
    车号识别模块，用于识别列车车号；
    车型判断模块，用于根据列车车号进行车型判断。
  12. 如权利要求11所述的列车车型识别系统，其中所述车号识别模块为如权利要求4-6中任一项所述的列车车号识别系统。
  13. 如权利要求11所述的列车车型识别系统，其中所述车型判断模块配置为利用列车车号从数据库或数据表中进行车型查找。
  14. 如权利要求11所述的列车车型识别系统，其中所述车型判断模块配置为根据列车车号定义规则从所述列车车号直接判断列车车型。
  15. 如权利要求11所述的列车车型识别系统，其中所述列车车型为机车、客车和货车中的一种。
  16. 如权利要求11所述的列车车型识别系统，还包括：
    勾当识别模块，用于识别列车的勾当位置；
    图像切分模块，用于将所述列车图像按勾当位置分为多张列车子图像。
  17. 一种列车安全检查方法，包括：
    利用如权利要求7-10中任一项所述的列车车型识别方法识别进入检查区域的被检列车的车型；
    如果被检列车的车型是机车或客车，则以低剂量照射被检列车或不照射被检列车；如果被检列车的车型是货车，则以高剂量照射被检列车。
  18. 一种列车安全检查系统，包括
    如权利要求11-16中任一项所述的列车车型识别系统；
    辐射控制模块，用于利用所述列车车型识别系统判断被检列车的车型，在被检列车的车型是机车或客车的情况下，控制射线源以第一剂量照射被检列车或不照射被检列车，在被检列车的车型是货车的情况下，控制射线源以第二剂量照射被检列车，其中第一剂量小 于第二剂量。
PCT/CN2016/094207 2015-12-29 2016-08-09 列车车号和车型识别方法和系统及安全检查方法和系统 Ceased WO2017113805A1 (zh)

Priority Applications (5)

Application Number Priority Date Filing Date Title
EA201792342A EA035267B1 (ru) 2015-12-29 2016-08-09 Способ и система для идентификации номера и типа вагона и способ и система для инспекции безопасности
EP16880600.8A EP3321860B1 (en) 2015-12-29 2016-08-09 Method and system for identifying train number and model, and safety check method and system
RU2017140638A RU2682007C1 (ru) 2015-12-29 2016-08-09 Способ и система для идентификации номера и типа вагона и способ и система для инспекции безопасности
PL16880600T PL3321860T3 (pl) 2015-12-29 2016-08-09 Sposób i system identyfikacji numeru pociągu i modelu oraz sposób i system kontroli bezpieczeństwa
UAA201711435A UA122334C2 (uk) 2015-12-29 2016-08-09 Спосіб і система ідентифікації номера і типу поїзда та спосіб і система перевірки безпеки

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201511016979.2A CN105447528B (zh) 2015-12-29 2015-12-29 列车车号和车型识别方法和系统及安全检查方法和系统
CN201511016979.2 2015-12-29

Publications (1)

Publication Number Publication Date
WO2017113805A1 true WO2017113805A1 (zh) 2017-07-06

Family

ID=55557687

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/094207 Ceased WO2017113805A1 (zh) 2015-12-29 2016-08-09 列车车号和车型识别方法和系统及安全检查方法和系统

Country Status (8)

Country Link
US (1) US10204268B2 (zh)
EP (1) EP3321860B1 (zh)
CN (1) CN105447528B (zh)
EA (1) EA035267B1 (zh)
PL (1) PL3321860T3 (zh)
RU (1) RU2682007C1 (zh)
UA (1) UA122334C2 (zh)
WO (1) WO2017113805A1 (zh)

Families Citing this family (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105447528B (zh) 2015-12-29 2018-11-27 同方威视技术股份有限公司 列车车号和车型识别方法和系统及安全检查方法和系统
CN106054271B (zh) * 2016-07-22 2020-02-07 同方威视技术股份有限公司 安全检查方法和系统
CN120298994A (zh) * 2017-11-21 2025-07-11 同方威视技术股份有限公司 车辆图像校正方法和装置、检查系统
CN107985347A (zh) * 2017-12-11 2018-05-04 同方威视技术股份有限公司 列车车厢自动扫描设备及方法
CN108108710A (zh) * 2017-12-29 2018-06-01 重庆微标科技股份有限公司 通过获取有效过车图像来快速识别过车机车上的车号
CN109178040A (zh) 2018-11-01 2019-01-11 同方威视技术股份有限公司 列车识别系统及其方法、列车安全检查系统及其方法
US12530636B2 (en) 2019-04-30 2026-01-20 Transportation Ip Holdings, Llc System and method for vehicle defect detection
US11755965B2 (en) 2019-04-30 2023-09-12 Transportation Ip Holdings, Llc Asset identification and tracking system
CN112132053B (zh) * 2020-09-24 2024-08-13 杭州海康威视数字技术股份有限公司 车厢标识识别方法、装置及计算机存储介质
EP4047550B1 (en) * 2021-02-23 2024-11-13 Hitachi, Ltd. Train inspection system
WO2022226159A1 (en) * 2021-04-21 2022-10-27 Siemens Mobility GmbH Automated selection and semantic connection of images
CN113495009B (zh) * 2021-05-24 2022-11-04 柳州龙燊汽车部件有限公司 一种车厢配套制造的质量检测方法及系统
CN114157808B (zh) * 2021-12-13 2022-11-29 北京国泰星云科技有限公司 一种高效的集装箱闸口图像采集系统及方法
CN114827463B (zh) * 2022-04-19 2023-09-22 中车青岛四方车辆研究所有限公司 一种图像采集控制系统及方法
CN115294299A (zh) * 2022-08-01 2022-11-04 成都唐源智控技术有限责任公司 基于雷达相机结合的列车停车状态识别装置、系统及方法
CN115031640B (zh) * 2022-08-12 2022-11-01 广州运达智能科技有限公司 一种列车轮对在线检测方法、系统、设备及存储介质
CN115856926A (zh) * 2022-11-24 2023-03-28 深圳市平方科技股份有限公司 集装箱识别方法、系统、装置、电子设备及介质

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050169500A1 (en) * 2004-01-30 2005-08-04 Fujitsu Limited Method of and apparatus for setting image-capturing conditions, and computer program
CN103294997A (zh) * 2013-05-20 2013-09-11 北京康拓红外技术股份有限公司 一种基于静态图像的列车车号识别方法
CN103646381A (zh) * 2013-11-22 2014-03-19 西安理工大学 一种线阵ccd的行进畸变校正方法
CN103679140A (zh) * 2013-11-26 2014-03-26 天津光电高斯通信工程技术股份有限公司 列车车号修正算法
CN104391339A (zh) * 2014-12-17 2015-03-04 同方威视技术股份有限公司 车型识别方法以及利用该方法的车辆快速检查系统
CN104463886A (zh) * 2014-12-18 2015-03-25 西南交通大学 一种线阵相机拍摄图像的处理方法及装置
CN105447528A (zh) * 2015-12-29 2016-03-30 同方威视技术股份有限公司 列车车号和车型识别方法和系统及安全检查方法和系统

Family Cites Families (20)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4876597A (en) * 1987-09-04 1989-10-24 Adt Security Systems, Inc. Video observation systems
US7618011B2 (en) * 2001-06-21 2009-11-17 General Electric Company Consist manager for managing two or more locomotives of a consist
JP2004212056A (ja) * 2002-12-26 2004-07-29 Aisin Aw Co Ltd ナビゲーションシステム及び経路探索方法のプログラム
IL162921A0 (en) * 2004-07-08 2005-11-20 Hi Tech Solutions Ltd Character recognition system and method
CN101468651B (zh) * 2007-12-27 2011-03-23 同方威视技术股份有限公司 火车车辆信息自动识别方法和系统
RU111936U1 (ru) * 2011-06-24 2011-12-27 НПП "ХАРТРОН-ЭКСПРЕСС Лтд" (ООО) Система удаленного мониторинга железнодорожных поездов
JP6075942B2 (ja) * 2011-09-30 2017-02-08 日本信号株式会社 列車案内放送システム
US8914168B2 (en) * 2012-04-05 2014-12-16 Union Pacific Railroad Company System and method for automated locomotive startup and shutdown recommendations
CN102708694B (zh) * 2012-06-19 2014-07-09 公安部第三研究所 基于高速脉冲激光扫描的车型自动识别系统及方法
US20140088865A1 (en) * 2012-09-27 2014-03-27 Siemens Industry, Inc. Apparatus and method for predicting arrival times in a transportation network
US9446776B2 (en) * 2012-12-02 2016-09-20 General Electric Company Inspection system and method
RU133082U1 (ru) * 2013-04-08 2013-10-10 Денис Сергеевич Гвоздев Устройство идентификации подвижных единиц железнодорожного транспорта
CN103279998B (zh) * 2013-05-06 2016-01-06 华南理工大学 一种收费公路收费站出入口车道车辆信息采集与识别系统
CN203311473U (zh) * 2013-05-06 2013-11-27 华南理工大学 一种收费公路收费站出入口车道车辆信息采集与识别系统
WO2015135015A1 (en) * 2014-03-11 2015-09-17 3D Image Automation Pty Ltd Train wagon 3d profiler
CN103839299A (zh) * 2014-03-17 2014-06-04 山东高速股份有限公司 无人值守自动发卡车道防作弊方法及实现该方法的系统
US9669850B2 (en) * 2014-09-08 2017-06-06 Genscape Intangible Holding, Inc. Method and system for monitoring rail operations and transport of commodities via rail
CN104374785B (zh) * 2014-11-14 2017-12-05 北京君和信达科技有限公司 一种连续通过式辐射扫描系统和方法
CN104408431B (zh) * 2014-11-28 2017-12-26 江苏物联网研究发展中心 交通监控下的车款识别方法
CN105059327B (zh) * 2015-07-30 2017-03-08 苏州华兴致远电子科技有限公司 一种车号识别方法及装置

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050169500A1 (en) * 2004-01-30 2005-08-04 Fujitsu Limited Method of and apparatus for setting image-capturing conditions, and computer program
CN103294997A (zh) * 2013-05-20 2013-09-11 北京康拓红外技术股份有限公司 一种基于静态图像的列车车号识别方法
CN103646381A (zh) * 2013-11-22 2014-03-19 西安理工大学 一种线阵ccd的行进畸变校正方法
CN103679140A (zh) * 2013-11-26 2014-03-26 天津光电高斯通信工程技术股份有限公司 列车车号修正算法
CN104391339A (zh) * 2014-12-17 2015-03-04 同方威视技术股份有限公司 车型识别方法以及利用该方法的车辆快速检查系统
CN104463886A (zh) * 2014-12-18 2015-03-25 西南交通大学 一种线阵相机拍摄图像的处理方法及装置
CN105447528A (zh) * 2015-12-29 2016-03-30 同方威视技术股份有限公司 列车车号和车型识别方法和系统及安全检查方法和系统

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3321860A4 *

Also Published As

Publication number Publication date
EP3321860A4 (en) 2019-04-10
EA201792342A1 (ru) 2018-05-31
EP3321860A1 (en) 2018-05-16
CN105447528B (zh) 2018-11-27
UA122334C2 (uk) 2020-10-26
US20170186172A1 (en) 2017-06-29
US10204268B2 (en) 2019-02-12
EA035267B1 (ru) 2020-05-22
EP3321860B1 (en) 2020-09-23
PL3321860T3 (pl) 2021-03-08
RU2682007C1 (ru) 2019-03-14
CN105447528A (zh) 2016-03-30

Similar Documents

Publication Publication Date Title
WO2017113805A1 (zh) 列车车号和车型识别方法和系统及安全检查方法和系统
CN105426922A (zh) 列车车型识别方法和系统及安全检查方法和系统
CN105203552A (zh) 一种360°踏面图像检测系统及其检测方法
JP6602595B2 (ja) 車種判別装置、及び車種判別方法
JPWO2019111976A1 (ja) 対象物検出装置、予測モデル作成装置、対象物検出方法及びプログラム
CN105809655A (zh) 车辆检查方法和系统
JP2014009975A (ja) ステレオカメラ
CN105891220A (zh) 一种路面标志线缺陷检测装置及其检测方法
CN103294997A (zh) 一种基于静态图像的列车车号识别方法
Ryu et al. Feature-based pothole detection in two-dimensional images
Kim et al. System and method for detecting potholes based on video data
CN102706883B (zh) 识别隧道已铺设段防水板孔洞的系统和方法
CN205365645U (zh) 列车安全检查系统
KR102540323B1 (ko) 영상분석 기반의 차종정보 자동추출장치 및 그 장치의 구동방법
WO2012014972A1 (ja) 車両挙動解析装置及び車両挙動解析プログラム
CN110796099A (zh) 一种车辆超限检测方法及装置
JP4674179B2 (ja) 影認識方法及び影境界抽出方法
CN108629226B (zh) 一种基于图像分层技术的车辆检测方法及系统
CN105118054B (zh) 一种基于ccd单目测距的驾驶考试系统
CN104766047B (zh) 基于车辆长度检测的高速收费站车辆识别方法及装置
CN104978562A (zh) 用于检测铁路车辆所行驶的轨道的方法
CN114399671A (zh) 目标识别方法及装置
CN109850596B (zh) 一种列车集装箱多目标定位方法、系统及运输车
JP2002342880A (ja) 車番認識装置
JP6233165B2 (ja) 形鋼測定方法、形鋼測定システム、及びプログラム

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16880600

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 201792342

Country of ref document: EA

Ref document number: A201711435

Country of ref document: UA

Ref document number: 2017140638

Country of ref document: RU

NENP Non-entry into the national phase

Ref country code: DE