WO2019095588A1 - 基于多摄像机的路侧停车管理方法、装置和系统 - Google Patents

基于多摄像机的路侧停车管理方法、装置和系统 Download PDF

Info

Publication number
WO2019095588A1
WO2019095588A1 PCT/CN2018/078300 CN2018078300W WO2019095588A1 WO 2019095588 A1 WO2019095588 A1 WO 2019095588A1 CN 2018078300 W CN2018078300 W CN 2018078300W WO 2019095588 A1 WO2019095588 A1 WO 2019095588A1
Authority
WO
WIPO (PCT)
Prior art keywords
vehicle
license plate
parking
area
image
Prior art date
Application number
PCT/CN2018/078300
Other languages
English (en)
French (fr)
Inventor
闫军
项炎平
杨学明
Original Assignee
智慧互通科技有限公司
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 智慧互通科技有限公司 filed Critical 智慧互通科技有限公司
Priority to US16/762,508 priority Critical patent/US10930151B2/en
Publication of WO2019095588A1 publication Critical patent/WO2019095588A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0108—Measuring and analyzing of parameters relative to traffic conditions based on the source of data
    • G08G1/0116—Measuring and analyzing of parameters relative to traffic conditions based on the source of data from roadside infrastructure, e.g. beacons
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/017—Detecting movement of traffic to be counted or controlled identifying vehicles
    • G08G1/0175—Detecting movement of traffic to be counted or controlled identifying vehicles by photographing vehicles, e.g. when violating traffic rules
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/40—Scenes; Scene-specific elements in video content
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
    • G06V20/54—Surveillance or monitoring of activities, e.g. for recognising suspicious objects of traffic, e.g. cars on the road, trains or boats
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0125—Traffic data processing
    • G08G1/0129—Traffic data processing for creating historical data or processing based on historical data
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
    • G08G1/0141—Measuring and analyzing of parameters relative to traffic conditions for specific applications for traffic information dissemination
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/04—Detecting movement of traffic to be counted or controlled using optical or ultrasonic detectors
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/065—Traffic control systems for road vehicles by counting the vehicles in a section of the road or in a parking area, i.e. comparing incoming count with outgoing count
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/14—Traffic control systems for road vehicles indicating individual free spaces in parking areas
    • G08G1/141—Traffic control systems for road vehicles indicating individual free spaces in parking areas with means giving the indication of available parking spaces
    • G08G1/142—Traffic control systems for road vehicles indicating individual free spaces in parking areas with means giving the indication of available parking spaces external to the vehicles
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/14—Traffic control systems for road vehicles indicating individual free spaces in parking areas
    • G08G1/145—Traffic control systems for road vehicles indicating individual free spaces in parking areas where the indication depends on the parking areas
    • G08G1/147—Traffic control systems for road vehicles indicating individual free spaces in parking areas where the indication depends on the parking areas where the parking area is within an open public zone, e.g. city centre
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N7/00—Television systems
    • H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
    • H04N7/181—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast for receiving images from a plurality of remote sources
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/40—Scenes; Scene-specific elements in video content
    • G06V20/44—Event detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/60—Type of objects
    • G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
    • G06V20/625—License plates

Definitions

  • the present disclosure relates to the field of target detection technologies, and in particular, to a parking management method, a parking management device, a parking management system, and a computer readable storage medium.
  • roadside parking also called roadside parking or on-street parking
  • Parallel parking lots in the roadside parking that is, parking spaces that are connected end to end to form a row of parking lots, hereinafter referred to as roadside parallel parking lots
  • This management method has irregular parking billing and low timing accuracy. Leakage charges, random charges, insufficient evidence for parking evidence, and difficult to trace, can not be managed 24 hours a day, and the human work management has high work intensity, low work efficiency and high security risks. Therefore, the current roadside parking has always been a difficult problem in urban management.
  • the management scheme based on the linkage of the gun ball camera is mainly through the use of a common master-slave gun-ball linkage camera, for example, including a group of guns (ie, a plurality of camera sets with fixed installation angles and focal lengths, hereinafter referred to as guns) and A billiard machine (a camera that automatically adjusts the angle of capture and focal length, hereinafter referred to as a ball machine).
  • a group of guns ie, a plurality of camera sets with fixed installation angles and focal lengths, hereinafter referred to as guns
  • a billiard machine a camera that automatically adjusts the angle of capture and focal length
  • the linkage between the gun and the ball machine on the L-shaped monitoring rod (the gun is adjusted to a fixed focal length and a snap angle to detect the parking event, and the ball machine adjusts the focal length and angle according to the detection result of the gun to capture the image of the target vehicle, and completes Vehicle parking forensics.), while taking into account the event detection and capturing the vehicle image and license plate, to achieve a wide range of parking spaces to capture the parking event.
  • the inventors of the present disclosure have found that the above-mentioned related art has a problem in that the angle of view of the gun and the dome are limited by the position at which the pole position is mounted, and the same target vehicle can only be captured from a single direction. Moreover, when the dome is called, it takes a period of time (ie, response time) to implement the capture task.
  • the above situation limits the ball machine to only capture one shot at a time when performing the capture task, which can only capture one at a time, thus affecting the shooting efficiency of the ball machine, resulting in subsequent series of such as leak capture, low recognition efficiency and concurrent parking events. Complex, low management efficiency issues.
  • the present disclosure proposes a parking management technical solution that can improve parking management efficiency.
  • a parking management method including: receiving a monitoring area image captured by a camera; and dividing the monitoring area image according to a preset image dividing rule to obtain the monitoring area image.
  • Each tracking detection area monitoring at least one of a vehicle and a license plate in the image of the monitoring area, and determining parking event information according to a tracking detection area in which at least one of the vehicle and the license plate is located.
  • the tracking detection area image includes a parking space area, an entrance detection area, and an exit detection area;
  • the parking event information includes a license plate number of the vehicle, a type of the parking event, and the vehicle where the parking event occurs. Parking space.
  • the entrance detection area includes an area adjacent to a boundary line of the parking space area; and the exit detection area is an area outside the parking space area.
  • a monitoring area image is acquired from the monitoring area image captured by the camera as a first monitoring area image; coordinates of the tracking detection area are set in the first monitoring area image; The image of the surveillance area that is subsequently captured by the camera is divided.
  • performing vehicle detection in the tracking detection area determining a vehicle in each tracking detection area; performing license plate detection in the tracking detection area, and determining a license plate in each tracking detection area; Determining the license plate, determining a license plate number of the license plate; performing motion detection on the vehicle and the license plate, determining a motion state of the vehicle and the license plate, the motion state including stationary and moving; a state of motion, determining parking event information of the vehicle; and determining parking event information of the vehicle of the license plate identification according to the motion state of the license plate.
  • the result of the vehicle detection is that only the license plate and/or the license plate number identifying the license plate is recognized, but the vehicle of the license plate identification is not detected; in the case where the vehicle area of the vehicle does not include the license plate area of any license plate
  • the result of determining the license plate detection and the vehicle detection is that only the vehicle is detected, but the license plate of the vehicle and the license plate number of the license plate are not detected.
  • determining parking event information of the vehicle according to the tracking detection area where the vehicle is located in a case where the motion state of the license plate is stationary, determining parking event information of the vehicle according to the tracking detection area where the vehicle is located; and in a case where the motion state of the license plate is stationary, according to a tracking detection area where the license plate is located, determining parking event information of the vehicle with the license plate identifier; and if the motion state of the vehicle is motion, tracking the vehicle to determine parking event information of the vehicle; In a case where the movement state of the license plate is motion, the license plate is tracked to determine parking event information of the vehicle identified by the license plate.
  • determining the parking event information of the license plate according to the tracking detection area where the license plate is located including: when the license plate is located in the parking space area, and identifying the license plate number of the license plate, Describe the license plate number, and find whether the license plate number is recorded in the presence vehicle information table; if the vehicle license plate number is not recorded in the presence vehicle information table, determine that the vehicle with the license plate identifier is an entry vehicle, and determine and Recording a license plate number of the vehicle with the license plate identification, an entry time of the vehicle, and a parking space where the vehicle is located; and calculating the license plate staying in the case where the license plate is located in an entrance detection area or an exit detection area The time of the parking area or the parking area; if the time exceeds the first threshold, the vehicle that identifies the license plate is determined to be a violating vehicle, and the parking event of the vehicle identified by the license plate is an illegal parking event.
  • determining the parking event information of the vehicle according to the tracking detection area where the vehicle is located including: when the vehicle is located in the parking space area, and the license plate number of the vehicle is not recognized, according to The license plate, searching whether the presence vehicle information table records whether the vehicle is recorded; if the vehicle is not recorded in the presence vehicle information table, determining that the vehicle is an entrance vehicle, and determining and recording the vehicle The time of entry of the vehicle, the parking space where the vehicle is located; and the time when the vehicle is located in the entrance detection area or the exit detection area, calculating the time when the vehicle stays in the entrance parking area or the exit parking area; If the time exceeds the first threshold, it is determined that the vehicle is a violating vehicle, and the parking event of the vehicle is an illegal parking event.
  • the vehicle area is intercepted as an initial target vehicle area for tracking, And intercepting the license plate area and the area around the license plate as an initial target license plate area for tracking; and detecting the vehicle in the tracking detection area, but not detecting the license plate of the vehicle, intercepting the The vehicle area is tracked as an initial target vehicle area; if the vehicle in the tracking detection area image is not detected, but the license plate of the vehicle is detected or the license plate number is recognized, the license plate image is intercepted and The area around the license plate image is tracked as the initial target license plate area.
  • the latest image of the vehicle is intercepted, and the target vehicle area is updated; the latest license plate image of the vehicle and an image of the surrounding area of the license plate image are intercepted, and the target license plate area is updated.
  • calculating a time at which the vehicle stays on the parking space In the case where the entrance detection area enters the parking space in the parking space area over the parking line, the time at which the license plate stays on the parking space is calculated.
  • the parking parking event information includes a license plate number, an admission time, and the vehicle when the admission parking event occurs.
  • the motion detection of the vehicle continues; the license plate is in the parking space
  • the time of staying above exceeds the second threshold, the tracking of the license plate is stopped, and the motion detection of the license plate is continued.
  • the entry key image includes: an image that is first detected by the vehicle in the entrance detection area, a license plate image that is first detected by the vehicle in the entrance detection area, the vehicle from The entrance detection area crosses the parking line to enter the image of the parking space area, and the image of the vehicle parked on the parking space.
  • the admission key timing image is a process image selected from the admission tracking video.
  • the entrance tracking video is a video segment in which the vehicle is first detected in the entrance detection area until the vehicle stops on the parking space.
  • the vehicle in a case where the vehicle enters the exit detection area from a parking space in the parking space area, determining that the vehicle is an exit vehicle, and determining an exit parking event information of the vehicle; Where the parking space in the parking space area enters the exit detection area, the vehicle that identifies the license plate identification is an exit vehicle, and the parking parking event information of the vehicle that determines the license plate identification; the exit parking event information includes a license plate No., playing time, the parking space where the vehicle is located, the key point image of the exit, the key timing image of the appearance, and the appearance tracking video.
  • the exit key image includes an image of the vehicle parked on the parking space before leaving the vehicle, an image that the vehicle is first detected in the exit detection area, and the vehicle is first detected in the exit detection area.
  • the exit key time series image is a process image selected from the exit tracking video.
  • the exit tracking video includes a first exit tracking video and a second exit tracking video, where the first exit tracking video is a video segment in which the vehicle enters the exit detection area from the parking space area until disappears from the exit detection area.
  • the second exit tracking video is a video segment in which the vehicle is detected for the first time in the exit detection area until the vehicle disappears in the exit detection area.
  • the license plate is tracked in the exit detection area and the license plate number of the license plate is recognized, whether the license plate number has been recorded in the presence vehicle information table is found according to the license plate number;
  • the vehicle that identifies the license plate identification is the exit vehicle, and the license plate number of the vehicle in which the license plate identification is determined, the time of the exit, and the parking space where the vehicle of the license plate identification is located.
  • a parking management apparatus including: a processor, configured to receive a monitoring area image captured by a camera; and dividing the monitoring area image to obtain tracking in the monitoring area image a detection area; and performing vehicle and license plate monitoring on the tracking detection area to determine parking event information; and a memory for storing the monitoring area image and parking event information.
  • the tracking detection area image includes a parking space area, an entrance detection area, and an exit detection area;
  • the parking event information includes a license plate number of the vehicle, a type of a parking event, and the vehicle when the parking event occurs.
  • the parking space where you are.
  • the parking management device further includes a display for displaying an operation management interface, the operation management interface is configured to receive a first query instruction input by the user, and the first query instruction is a user selecting to query the camera.
  • An operation of the monitoring information the operation management interface acquires an image of the monitoring area captured by the camera according to the first query instruction, and displays the image of the monitoring area; the operation management interface receives an image of the user in the monitoring area
  • the tracking detection area coordinates selected in the tracking; the coordinates of the tracking detection area are the basis for dividing the image of the monitoring area.
  • the processor performs vehicle detection in each of the tracking detection areas, determines a vehicle in each of the tracking detection areas, and the processor performs vehicle license plate detection in each tracking detection area to determine the Tracking the license plate in the detection area, the processor identifying the license plate, determining a license plate number of the license plate, the processor performing motion detection on the vehicle and the license plate, determining the vehicle, the a motion state of the license plate, the motion state includes a rest and a motion, and the processor determines the parking event information of the vehicle and the license plate according to a motion state of the vehicle and the license plate.
  • the processor acquires a positional relationship between the license plate and the vehicle, and the processor determines the license plate detection and location if a license plate area of the license plate is located in a vehicle area of the vehicle.
  • the result of the vehicle detection is that both the vehicle is detected and the license plate number of the license plate and/or the license plate is detected, and the processor is not located in the vehicle area of any vehicle in the license plate area of the license plate.
  • the processor is in the vehicle of the vehicle In the case where the area does not include the license plate area of any license plate, the result of determining the license plate detection and the vehicle detection is that only the vehicle is detected, but the license plate of the vehicle and the license plate number of the license plate are not detected.
  • the processor determines, when the motion state of the vehicle is stationary, determining parking event information of the vehicle according to the tracking detection area where the vehicle is located, and the motion of the processor in the license plate
  • the parking event information of the vehicle of the license plate identification is determined according to the tracking detection area where the license plate is located, and the processor is in the case that the motion state of the vehicle is motion.
  • Tracking is performed to determine parking event information of the vehicle, and the processor tracks the license plate when the movement state of the license plate is motion, and determines parking event information of the vehicle identified by the license plate.
  • the processor searches for the license plate number in the presence vehicle information table according to the license plate number when the license plate is located in the parking space area and the license plate number of the license plate is recognized.
  • the processor determines that the vehicle of the license plate identification is an entrance vehicle, and determines and records the license plate number of the vehicle with the license plate identification, the vehicle, in a case where the vehicle information table does not record the license plate number.
  • the processor searches whether the vehicle is recorded in the presence vehicle information table according to the vehicle; If the presence vehicle information table does not record the vehicle, determining that the vehicle is an entrance vehicle, and determining and recording admission parking event information of the vehicle; the parking parking event information of the vehicle includes a vehicle An entry time, a parking space where the vehicle is located; the processor calculates a time when the vehicle is stuck in an entrance parking area or an exit parking area, in a case where the vehicle is located in an entrance detection area or an exit detection area, If the time exceeds the first threshold, it is determined that the vehicle is a violating vehicle, and the parking event of the vehicle is an illegal parking event.
  • the processor intercepts the vehicle area as an initial target vehicle area to track when detecting the vehicle in the tracking detection area image and detecting the license plate or license plate number of the vehicle. And intercepting the license plate area and the area around the license plate as an initial target license plate area for tracking, the processor detecting the vehicle in the tracking detection area, but not detecting the license plate of the vehicle And intercepting the vehicle area as an initial target vehicle area, wherein the processor does not detect the vehicle in the tracking detection area image, but detects the license plate or license plate number of the vehicle, The license plate image and an area around the license plate image are intercepted for tracking as the initial target license plate area.
  • the processor intercepts the latest image of the vehicle to update the target vehicle area; the processor intercepts an image of a latest license plate image of the vehicle and a surrounding area of the license plate image, and the target license plate The area is updated.
  • the processor calculates a time for the vehicle to stay on the parking space in a case where the vehicle enters a parking space in the parking space area from the entrance detection area over a parking line; the processing Calculating a time for the license plate to stay on the parking space when the license plate enters a parking space in the parking space area from the entrance detection area over a parking line; the processor is at the vehicle If the time of staying on the parking space exceeds the second threshold, determining that the vehicle is an entrance vehicle, a parking event of the vehicle is an entrance parking event, and determining an admission parking event information, and parking the admission The event information is recorded in the presence parking information table; if the time when the license plate stays on the parking space exceeds a second threshold, the processor determines that the vehicle identified by the license plate is an entry vehicle, the license plate identifier The parking event of the vehicle is an admission parking event, and the information of the admission parking event is determined, and the information of the admission parking event is recorded in the on-site parking information table; Information including the license plate number, time of admission, time
  • the processor stops tracking the vehicle if the time that the vehicle stays on the parking space exceeds the second threshold, and the motion detection of the vehicle continues; the processor In the case where the license plate stays on the parking space for more than the second threshold, the tracking of the license plate is stopped, and the motion detection of the license plate continues.
  • the entry key image includes: an image that is first detected by the vehicle in the entrance detection area, a license plate image that is first detected by the vehicle in the entrance detection area, the vehicle from The entrance detection area crosses the parking line to enter the image of the parking space area and the process image of the vehicle parked on the parking space.
  • the admission key timing image is an image selected from the admission tracking video.
  • the entrance tracking video is a video segment in which the vehicle is first detected in the entrance detection area until the vehicle stops on the parking space.
  • the processor determines, when the vehicle enters the exit detection area from a parking space in the parking space area, determines that the vehicle is an exit vehicle, and determines an exit parking event information of the vehicle;
  • the vehicle that determines the license plate identification is an exit vehicle, and the parking parking event information of the vehicle that determines the license plate identification;
  • the exit parking event information includes a license plate number, an exit time, a parking space where the vehicle is located when the exit parking event occurs, an image of the key point of the exit, a key timing image of the exit, and an exit tracking video.
  • the exit key image includes an image of the vehicle parked on the parking space before leaving the vehicle, an image that the vehicle is first detected in the exit detection area, and the vehicle is first detected in the exit detection area.
  • the exit key time series image is a process image selected from the exit tracking video.
  • the appearance tracking video includes a first appearance tracking video and a second appearance tracking video;
  • the first appearance tracking video is a video segment in which the vehicle enters the exit detection area from the parking space area until disappearing from the exit detection area;
  • the second exit tracking video is a video segment in which the vehicle is detected for the first time in the exit detection area until the vehicle stops in the exit detection area.
  • the processor tracks the license plate in the exit detection area and identifies the license plate number of the license plate, according to the license plate number, it is found whether the license plate number has been recorded in the presence vehicle information table.
  • the vehicle that identifies the license plate identification is an exit vehicle, and the vehicle number, the exit time, and the license plate identification vehicle of the vehicle with the license plate identification determined The parking space where you are.
  • a parking management apparatus including: an input module, configured to receive a monitoring area image captured by a camera; and an image dividing module, configured to: according to a preset image dividing rule, the monitoring area The image is divided to obtain a tracking detection area in the image of the monitoring area; a detection and identification tracking module is configured to perform vehicle and license plate monitoring on the tracking detection area, and determine parking event information.
  • the tracking detection area image includes a parking space area, an entrance detection area, and an exit detection area;
  • the parking event information includes a license plate number of the vehicle, a type of a parking event, and the vehicle when the parking event occurs.
  • the parking space where you are.
  • the entrance detection area includes a neighboring area on both sides of the boundary line of the parking space area; and the exit detection area is an area outside the parking space area.
  • the parking management device further includes an operation management interface, configured to receive a first query instruction input by the user, where the first query instruction is an operation of the user selecting to query the monitoring information of the camera, according to the first Querying an instruction, acquiring an image of the first monitoring area captured by the camera, and displaying the image of the first monitoring area, and receiving coordinates of the tracking detection area selected by the user in the image of the first monitoring area;
  • the coordinates are the image division rules of the camera.
  • the coordinates of the tracking detection area include coordinates of a parking space area, coordinates of an entrance detection area, and coordinates of an exit detection area.
  • the detection and identification tracking module comprises: a vehicle detection module, configured to perform vehicle detection on the tracking detection area by using a vehicle detection algorithm, determine a vehicle in the tracking detection area; and a license plate detection module, configured to adopt a license plate detection algorithm Performing license plate detection on the tracking detection area to determine a license plate in the tracking detection area; a license plate recognition module for identifying the license plate by using a license plate recognition algorithm to determine a license plate number of the license plate; and a motion detection module for adopting a motion detection algorithm performs motion detection on the vehicle and the license plate, determines a motion state of the vehicle and the license plate, and a motion state of the target vehicle and the target license plate includes a stationary and a motion; and a tracking processing module, configured to: Determining the parking event information of the vehicle and the license plate according to the motion state of the vehicle and the license plate.
  • the parking management apparatus further includes: an operation determining module, configured to acquire a positional relationship between the license plate and the vehicle, and the operation determining module is located in a vehicle area of the vehicle in a license plate area of the license plate Determining, the result of the license plate detection and the vehicle detection is that the vehicle is detected, and the license plate number of the vehicle and/or the license plate number of the license plate is detected; the operation judgment module is in the license plate area of the license plate In the case where there is no vehicle area located in any vehicle, it is determined that the result of the license plate detection and the vehicle detection is a vehicle that detects only the license plate number of the license plate and/or the license plate, but does not detect the license plate identification; The operation judging module determines that the vehicle license plate detection and the vehicle detection result are that only the vehicle is detected but the vehicle license plate is not detected, in a case where the vehicle region of the vehicle does not include a license plate region of any license plate. And the license plate number of the license plate.
  • an operation determining module configured to acquire a
  • the tracking processing module determines parking event information of the vehicle according to a tracking detection area where the vehicle is located, where the motion state of the vehicle is stationary, and the tracking processing module is in the license plate
  • the parking event information of the vehicle with the license plate identification is determined according to the tracking detection area where the license plate is located, and the tracking processing module is in the case that the motion state of the vehicle is motion
  • the vehicle performs tracking to determine parking event information of the vehicle, and the tracking processing module tracks the license plate when the movement state of the license plate is motion, and determines a parking event of the vehicle with the license plate identification information.
  • the tracking processing module searches whether the recorded vehicle information table is recorded according to the license plate number.
  • the license plate number in a case where the vehicle number information table does not record the license plate number, the vehicle that identifies the license plate identification is an entrance vehicle, and the license plate number of the vehicle that identifies and records the license plate identification, The entry time of the vehicle and the parking space where the vehicle is located.
  • the tracking processing module calculates a time when the license plate stays in an entrance parking area or an exit parking area, where the time exceeds the first time.
  • the vehicle that identifies the license plate identification is a violating vehicle, and the parking event of the vehicle identified by the license plate is an illegal parking event.
  • the tracking processing module searches whether the presence vehicle information table is recorded according to the license plate. There is the vehicle, in the case where the vehicle information table does not record the vehicle, determining that the vehicle is an entrance vehicle, and determining and recording an entry parking event information of the vehicle, the admission of the vehicle
  • the parking event information includes the entry time of the vehicle and the parking space where the vehicle is located.
  • the tracking processing module calculates a time when the vehicle stays in an entrance parking area or an exit parking area, where the time exceeds the In the case of the first threshold, it is determined that the vehicle is a violating vehicle, and the parking event of the vehicle is an illegal parking event.
  • the tracking processing module starts the vehicle tracking module to intercept the vehicle area as an initial target vehicle area for tracking, and activates the license plate tracking module to intercept the license plate area and the area around the license plate as an initial target license plate area for tracking;
  • the tracking processing module starts the vehicle tracking module to intercept the vehicle area as an initial Tracking of the target vehicle area;
  • the vehicle in the tracking detection area image is not detected by the vehicle detection module, but the license plate detection module detects that the license plate of the vehicle or the license plate recognition module recognizes the license plate number
  • the tracking processing module starts the license plate tracking module to intercept the car The card image and the area around the license plate image are tracked as the initial target license plate area.
  • the vehicle tracking module intercepts a latest image of the vehicle, and updates the target vehicle area;
  • the license plate tracking module intercepts an image of a latest license plate image of the vehicle and an image of a peripheral area of the license plate image, The target license plate area is updated.
  • the tracking processing module calculates a time for the vehicle to stay on the parking space in a case where the vehicle enters a parking space in the parking space area from the entrance detection area over a parking line;
  • the tracking processing module calculates a time for the license plate to stay on the parking space in a case where the license plate enters a parking space in the parking space region from the parking space through the parking line;
  • the tracking processing module is in the If the time when the vehicle stays on the parking space exceeds a second threshold, determining that the vehicle is an entrance vehicle, a parking event of the vehicle is an entrance parking event, and determining an admission parking event information, and
  • the admission parking event information is recorded in the presence parking information table; and the tracking processing module determines that the vehicle identified by the license plate is an entry vehicle if the time when the license plate stays on the parking space exceeds a second threshold.
  • the parking event of the vehicle identified by the license plate is an admission parking event, and the information of the admission parking event is determined, and the information of the admission parking event is recorded to the on-site parking.
  • Information table said entrance parking event information includes the license plate number, time of admission, the admission of the parking vehicle is parking event, the key entry point of the image, the key entry sequence images, video tracking admission.
  • the tracking processing module stops tracking the vehicle, and the motion detection module continues to detect motion of the vehicle Performing; if the time when the license plate stays on the parking space exceeds the second threshold, the tracking processing module stops tracking the vehicle of the license plate identification, and the motion detection module identifies the vehicle with the license plate The motion detection continues.
  • the entry key image includes: an image that is first detected by the vehicle in the entrance detection area, a license plate image that is first detected by the vehicle in the entrance detection area, the vehicle from The entrance detection area crosses the parking line to enter the image of the parking space area and the image of the vehicle parked on the parking space.
  • the admission key timing image is an image selected from the admission tracking video.
  • the entrance tracking video is a video segment in which the vehicle is first detected in the entrance detection area until the vehicle stops on the parking space.
  • the tracking processing module determines that the vehicle is an exit vehicle, and determines an exit parking event information of the vehicle, if the vehicle enters the exit detection area from a parking space in the parking space area;
  • the tracking processing module determines that the vehicle identified by the license plate is an exit vehicle and the exit parking event of the vehicle that identifies the license plate identification, in a case where the license plate enters the exit detection area from a parking space in the parking space area.
  • Information; the information of the exit parking event includes the license plate number, the playing time, the parking space where the vehicle is located when the parking event occurs, the key image of the exit, the key timing image of the exit, and the appearance tracking video.
  • the exit key image includes an image of the vehicle parked on the parking space before leaving the vehicle, an image that the vehicle is first detected in the exit detection area, and the vehicle is first detected in the exit detection area.
  • the exit key time series image is a process image selected from the exit tracking video.
  • the exit tracking video includes a first exit tracking video and a second exit tracking video, where the first exit tracking video is a video segment in which the vehicle enters the exit detection area from the parking space area until disappears from the exit detection area.
  • the second exit tracking video is a video segment in which the vehicle is detected for the first time in the exit detection area until the vehicle stops in the exit detection area.
  • the tracking processing module searches according to the license plate number. Whether the license plate number has been recorded in the presence vehicle information table, and in the case where the license plate number has been recorded, the vehicle that identifies the license plate identification is the exit vehicle, and the license plate number and appearance of the vehicle that identifies the license plate identification The time, the parking space where the vehicle identified by the license plate is located.
  • a parking management system comprising: a camera group disposed on a pole position of a roadside parking lot for acquiring a monitoring area image; parking management described in any of the above embodiments
  • the device is configured to perform image segmentation on the monitoring area image according to an image division rule, and perform detection and identification tracking processing on the monitoring area image to determine parking event information on the parking space managed by the camera group.
  • the camera group comprises a plurality of camera groups, each of the camera groups comprising at least one camera array, each of the camera arrays comprising at least one camera, one camera group mounted on one pole position, first and last The camera set on both pole positions contains one camera array, and the camera set on the remaining pole positions contains two camera arrays.
  • a computer readable storage medium having stored thereon a computer program that, when executed by a processor, implements the parking management method of any of the embodiments.
  • the camera since the camera, especially the gun type camera, has the characteristics of stable viewing angle and focal length, it can continuously acquire stable images. For example, the current 2 megapixel gun is deployed on a 6 meter monitor. When the field of view covers 2-3 parking spaces, a clear and stable license plate image can be obtained. Therefore, the method and device provided by the present invention are based on dividing the image captured by the camera, and simultaneously processing the obtained image regions, thereby realizing tracking and recognition of the vehicle entering and leaving the field, and occurring when several cameras are monitored. When a plurality of concurrent parking events are performed, a plurality of parking events can be simultaneously managed according to the divided area. Compared with the management mode of the ball machine, the scheduling response time of the camera is not required to be considered. Therefore, the parking event can be improved by using the method provided by the embodiment of the present invention. Management efficiency.
  • FIG. 1 is an exemplary flowchart of a parking management method according to an embodiment of the present invention
  • FIG. 2 is a diagram showing an installation scenario of an array camera in an embodiment of the present invention
  • FIG. 3 is a schematic diagram of dividing a first monitoring area image according to an embodiment of the present invention.
  • step S104 in FIG. 1 is an exemplary flowchart of a specific implementation of step S104 in FIG. 1;
  • FIG. 5 is a scene diagram of detecting and recognizing a vehicle according to an embodiment of the present invention.
  • FIG. 6 is a schematic block diagram showing the structure of a parking management apparatus according to an embodiment of the present invention.
  • FIG. 7 is a schematic block diagram showing another structure of a parking management apparatus according to an embodiment of the present invention.
  • FIG. 8 is a schematic block diagram showing the structure of a parking management system according to an embodiment of the present invention.
  • FIG. 1 is an exemplary flowchart of a parking management method according to an embodiment of the present invention.
  • the method specifically includes steps S102 to S104.
  • step S101 may also be included.
  • Each step can be embodied in the following manner.
  • step S101 the image division rule of the camera is preset.
  • the camera can be mounted in the manner of FIG.
  • FIG 2 shows an installation scenario diagram of an array camera, in accordance with some embodiments of the present disclosure.
  • several cameras can be used as an array to form a camera group that is fixed and fixed in the pole position in the roadside parking lot to monitor and manage the roadside parking events. After the camera is installed, set the parameters such as the focal length and shooting angle of each camera to determine the roadside parking monitoring area managed by the camera to obtain monitoring information.
  • a monitoring area image may be acquired from the monitoring area image captured by the camera as the first monitoring area image, and the coordinates of the tracking detection area are set in the first monitoring area image, and the camera is subsequently photographed according to the coordinates.
  • the area of the surveillance area is divided. For example, a frame of the latest shooting may be selected in the image of the monitoring area captured by the camera as the first monitoring area image for presetting the image dividing rule of the camera. That is, the coordinates of the tracking detection area are selected in the image of the first monitoring area as the image dividing rule of the camera and stored, so that the subsequent steps divide and identify the monitoring area image captured by the camera according to the set image dividing rule. deal with.
  • the tracking detection area is an active area of the vehicle, and may include a parking space area, an entrance detection area, and an exit detection area.
  • the coordinates of the tracking detection area include coordinates of the parking space area, the entrance detection area, and the exit detection area in the first monitoring area image.
  • the first monitoring area image can be as shown in Fig. 2.
  • FIG. 3 illustrates a schematic diagram of a first monitored area image, in accordance with some embodiments of the present disclosure.
  • the tracking detection area may include a parking space area, an entrance detection area, and an exit detection area.
  • the parking space area, the entrance detection area, and the exit detection area may be divided according to the following method.
  • the parking space area is an area for parking a vehicle, and includes a plurality of parking berths, such as an area between the boundary 2 and the boundary line 6 in FIG.
  • the entrance detection area includes adjacent areas on both sides of the boundary line of the parking space area.
  • the entrance detection area may be used as an area for detecting an entrance parking event and acquiring the parking parking event information.
  • the boundary 2 in the figure is a boundary line of a parking space area, which is also called a parking space line, and adjacent areas on both sides of the boundary line refer to an area between the boundary line 1 and the boundary line 2, and the boundary line 3
  • the distance between the boundary 1, the boundary 3 and the boundary 2 can be set according to actual needs. Since the speed of the vehicle is relatively slow when the vehicle crosses the parking line, it is a good opportunity to capture the license plate number. Therefore, the area between the boundary 1 and the boundary 2 is used as a part of the entrance detection area, and the efficiency of the license plate can be effectively captured. Accuracy.
  • the exit detection area is an area outside the parking space area, for example, an area between the boundary 2 and the boundary 4 and the boundary 5 in FIG.
  • the exit detection area may be used as an area for detecting an exit parking event and acquiring the parking parking event information.
  • irrelevant areas such as lawn areas, tree areas, etc.
  • irrelevant areas such as lawn areas, tree areas, etc.
  • the amount of calculation may be performed according to the specific monitoring scene of each camera, and used as a vehicle tracking monitoring area, thereby reducing images of subsequent tracking vehicle tracking monitoring areas. The amount of calculation.
  • step S102 a monitoring area image captured by the camera is received.
  • step S103 the monitoring area image captured by the camera is divided according to the image division rule of the camera, and the tracking detection area in the monitoring area image is obtained.
  • the roadside parking monitoring area may include a parking space area, an entrance detection area, and an exit detection area.
  • step S104 at least one of the vehicle and the license plate in the monitoring area image is monitored, and the parking event information is determined according to the tracking detection area in which at least one of the vehicle and the license plate is located.
  • Monitoring can include detection, identification, tracking, and the like.
  • all vehicles and license plates in the entire surveillance area image can be detected and identified. It is also possible to perform parallel processing using a multi-threading method or a multi-processing module or a multi-processor.
  • the tracking detection area is used as a processing unit, and vehicles and license plates are detected and identified in parallel for each area. Parallel processing can improve the detection and recognition efficiency of vehicles and license plates.
  • the parking event may include an admission parking event, an exit parking event, a violation, a violation, and an illegal parking event.
  • the parking event information may include a type of parking event, a time when the parking event occurs, a time when the vehicle is parked on the parking space, a parking space where the vehicle is located when the parking event occurs, a license plate number, a key point image, a tracking video, and a key time series image. After the parking event information is determined, the parking event information may be stored for subsequent uploading to the background for the management personnel to query and patrol.
  • the types of parking events may include an admission parking event, an exit parking event, a violation, a violation, and an illegal parking event.
  • the time when the parking event occurs may include the time when the vehicle enters the parking space (also known as the entrance time) and the time when the vehicle exits the parking space (also known as the playing time). The time the vehicle is parked on the parking space can be calculated based on the vehicle's playing time and admission time.
  • the key point image may include an entry key image of the vehicle and an exit key image.
  • the entry key image may include a vehicle image that is first detected by the vehicle in the entrance detection area, a license plate image that is first detected by the vehicle in the entrance detection area, and the vehicle is detected from the entrance.
  • the exit key image may include an image of the vehicle parked on the parking space before the vehicle exits, a vehicle image that is first detected by the vehicle in the exit detection area, and the vehicle is first detected in the exit detection area.
  • a license plate image an image of the vehicle entering the exit detection area from the parking space area over the parking space, a vehicle image that was last detected by the vehicle in the exit detection area, and the vehicle was last detected in the exit detection area. License plate image.
  • an abnormal parking event for example, the background server displays a vehicle on the parking space, but the vehicle does not display the license plate
  • the staff can inspect the key image to supplement the parking event. information.
  • the tracking video may include an entrance tracking video and an exit tracking video.
  • the entrance tracking video may be a video segment in which the vehicle is first detected in the entrance detection area until the vehicle is parked on the parking space.
  • the exit tracking video may include a first appearance tracking video and a second appearance tracking video.
  • the first exit tracking video may be a video segment in which the vehicle enters the exit detection area from the parking space area until disappears from the exit detection area.
  • the second exit tracking video may be a video segment in which the vehicle is first detected in the exit detection area until the vehicle disappears in the exit detection area. Since the processing capability of the front-end device/device for performing the method is limited, the tracking video can be stored first for later use when the background processing capability is strong.
  • the tracking video can also be uploaded to the background, and the tracking video is recognized and processed by the background, which can further improve the accuracy of identifying the vehicle and the license plate.
  • the key time series image may include an entry key time series image and an exit key time series image.
  • the entry key timing image can be selected and extracted from the admission tracking video.
  • the appearance key timing images can be selected and extracted from the exit tracking video.
  • the pictures of the corresponding moments are selected from the entrance tracking video to be combined into the key moment images according to actual requirements, and the pictures of the corresponding moments are selected from the exit tracking video segments to be combined into the exit key time series images. Since transmission tracking video requires large bandwidth and excessive traffic, in the case of small bandwidth, key timing images can be transmitted to the background, and key timing images are identified and processed by the background to further improve Identify the accuracy of vehicles and license plates.
  • step S104 can be implemented by the flow in FIG.
  • FIG. 4 illustrates an exemplary flow chart of a parking event information determination method in accordance with some embodiments of the present disclosure.
  • step S104 can be implemented by steps S1041 to S1044.
  • step S1041 vehicle detection is performed on the tracking detection area, and the vehicle in the tracking detection area is determined.
  • Vehicle tracking detection is performed on the tracking detection area, and a license plate in the tracking detection area is determined. And identifying the license plate to determine a license plate number of the license plate.
  • vehicle detection may be performed on the tracking detection area using a vehicle detection algorithm.
  • the license plate detection algorithm may also be used to perform license plate detection on the tracking detection area and the license plate recognition algorithm may be used to identify the license plate.
  • the vehicle detection algorithm and the license plate detection algorithm may adopt a Fast Rcnn algorithm based on deep learning (English name: Faster Regions with CNNs features), an SSD algorithm (English full name: single shot multibox detector), and a Yolo algorithm (English full name) :You Only Look Once) and other target detection algorithms, or other types of image target detection algorithms.
  • a Fast Rcnn algorithm based on deep learning (English name: Faster Regions with CNNs features), an SSD algorithm (English full name: single shot multibox detector), and a Yolo algorithm (English full name) :You Only Look Once) and other target detection algorithms, or other types of image target detection algorithms.
  • step S1042 a positional relationship between the license plate and the vehicle is acquired to determine whether the license plate is a license plate of the vehicle. For example, after determining the vehicle and the license plate in the tracking detection area, whether the license plate is the license plate of the vehicle may be determined according to the positional relationship between the vehicle and the license plate. In some embodiments, the determination can be made by the following method.
  • the license plate area of the license plate is located in the vehicle area of the vehicle, it is determined that both the vehicle is detected, the license plate of the vehicle is detected, and/or the license plate number of the license plate is recognized. For example, as shown in area 10411 of Figure 3, the license plate area is located within the vehicle area, thus indicating that both the vehicle is detected and the license plate number of the vehicle is detected and/or identified.
  • the license plate may be associated with the vehicle, and the feature information of the vehicle and the feature information of the license plate may be combined as the feature information of the vehicle, and the coordinates of the license plate region are preferentially selected as the coordinates of the vehicle, and the vehicle region may be As an auxiliary tracking area.
  • the license plate area of the license plate is not located in the vehicle area of any vehicle, it is determined that only the license plate of the vehicle and/or the license plate number of the license plate is detected, but the license plate identification is not detected.
  • vehicle For example, as shown in area 10412 of FIG. 3, the license plate area is not located in any of the vehicle areas, thus indicating that the vehicle in the tracking detection area image is not detected, but the license plate number of the vehicle is detected and/or identified.
  • the feature information of the license plate may be used as the feature information of the vehicle.
  • the vehicle area of the vehicle does not include a license plate area of any license plate, it is determined that only the vehicle is detected, but the license plate of the vehicle and/or the license plate number of the license plate is not recognized. For example, as shown in area 10413 of FIG. 3, the vehicle area does not contain any license plate area, thus indicating that only the vehicle in the tracking detection area image is detected, but the license plate number of the vehicle is not detected and recognized.
  • step S1043 motion detection is performed on the vehicle and the license plate, and the motion state of the vehicle and the license plate is determined.
  • the motion state of the vehicle and the license plate includes stationary and moving.
  • motion detection may be performed on the vehicle and the license plate by using a motion detection algorithm.
  • the motion detection algorithm may employ an optical flow based Lucas-knades algorithm, a Kalman filtering algorithm, or the like.
  • step S1044 parking event information is determined based on the motion state of the vehicle and the motion state of the license plate.
  • the parking event information of the vehicle is determined according to the tracking detection area in which the vehicle is located. If the motion state of the license plate is stationary, the parking event information of the vehicle identified by the license plate is determined according to the tracking detection area where the license plate is located. If the motion state of the vehicle is motion, the vehicle is tracked to determine parking event information of the vehicle. If the motion state of the license plate is motion, the license plate is tracked to determine parking event information of the vehicle identified by the license plate.
  • a vehicle tracking algorithm may be used to track the vehicles in the tracking detection area image.
  • the license plate tracking algorithm may be used to track the license plates in the tracking detection area image.
  • the license plate tracking algorithm and the vehicle tracking algorithm may adopt a KCF algorithm based on a correlation filter (English name: High-speed tracking with kernelized correlation filters), and a Staple (English name: Sum of Template And Pixel-wise LEarners) target tracking algorithm. Or other types of image target tracking algorithms.
  • FIG. 5 illustrates a scene graph of detecting recognition tracking of a vehicle, in accordance with some embodiments of the present disclosure.
  • the parking event information can be confirmed by the following procedure.
  • the parking event information of the vehicle is determined according to the tracking detection area in which the vehicle is located. If the motion state of the license plate is stationary, the parking event information of the vehicle identified by the license plate is determined according to the tracking detection area where the license plate is located.
  • the vehicle if the vehicle is located in an entrance detection area (such as vehicle A in FIG. 5), then the time at which the vehicle is stuck in the admission parking area is calculated. Or if the license plate is located in the entrance detection area, the time at which the license plate is stuck in the entrance parking area is calculated. If the vehicle or the license plate is stuck in the entrance detection area exceeding a first threshold, it is determined that the vehicle or the vehicle identified by the license plate is an illegal parking vehicle, and the type of the parking event is a parking violation event.
  • the first threshold can be set according to actual needs. For example, the first threshold can be set to 3-5 minutes.
  • the above processing means can process the situation listed in step S1042. For example, since the vehicle is in a stationary state in the case (1) of step S1042, sometimes the license plate or the vehicle is temporarily unable to be detected or recognized due to the license plate or the vehicle being blocked by a person or other objects, so that when the license plate is detected or When the vehicle is at a standstill, the confirmation process of the parking event information can be performed on the vehicle in the case (1) of step S1042.
  • the presence vehicle information table is searched for based on the license plate number. If the license plate number has not been previously recorded, the vehicle identified by the license plate is determined to be an entry vehicle, and the entry parking event information of the vehicle identified by the license plate is determined.
  • the admission parking event information may include a license plate number of the vehicle, an entry time of the vehicle, and a parking space where the vehicle is located.
  • the presence vehicle information table is searched for based on characteristic information of the vehicle (eg, color, vehicle type, etc.). If the vehicle has not been previously recorded, it is determined that the vehicle is an unlicensed vehicle and is determined to be an entry vehicle, and an entry parking event information of the vehicle identified by the license plate is determined.
  • the admission parking event information may include feature information of the vehicle, an entry time of the vehicle, and a parking space where the vehicle is located.
  • the parking event information of the vehicle cannot be recorded after the vehicle is parked in the parking space. Therefore, in some embodiments, the motion detection and recognition of the vehicle parked in the parking space area can be performed in real time or at a time, and the vehicle type, color, and entry of the vehicle can be supplementally recorded by comparing the identified license plate number with the on-site parking information table. Information on admission events such as time and license plate number.
  • the vehicle is located in an exit detection area (such as vehicle C in FIG. 5)
  • the time at which the vehicle is stuck in the exit parking area is calculated.
  • the license plate is located in the exit detection area, the time at which the license plate vehicle stays in the exit parking area is calculated. If the time when the vehicle or the license plate stays in the exit detection area exceeds a first threshold, it is determined that the vehicle or the vehicle identified by the license plate is an illegal parking vehicle, and the type of the parking event is a parking violation event.
  • the determination of illegal parking events can assist traffic management.
  • the above processing means can also handle the situation listed in step S1042.
  • the vehicle is tracked to determine parking event information of the vehicle. If the motion state of the license plate is motion, the license plate is tracked to determine parking event information of the vehicle identified by the license plate.
  • a corresponding tracking scheme can be taken for the vehicle based on the situation determined in step S1042. For example, it can be realized by the following method.
  • the vehicle tracking algorithm is used to intercept the vehicle area as an initial The target vehicle area is tracked, and a license plate tracking algorithm is used to intercept the license plate area and the area around the license plate as an initial target license plate area for tracking.
  • a vehicle such as vehicle D in FIG. 5 enters a parking space in the parking space area from the entrance detection zone across the parking line (ie, boundary 2 in FIG. 2) (ie, the target vehicle area and target)
  • the license plate area is moved into the parking space in the parking space area, and the time at which the vehicle stays on the parking space is calculated.
  • the tracking is stopped, and the motion detection of the vehicle continues (because the computing resource consumed by the tracking algorithm is large, the tracking can be stopped when it is determined that the vehicle is parked on the parking space)
  • the parking is an admission parking event (ie, the type of parking event is an entrance parking event)
  • information about the admission parking event is determined and stored.
  • the vehicle license plate number, the entrance time, the parking space where the vehicle is located at the time of the entrance parking event, and the like are written into the on-site vehicle information table. And associating and storing the entry key image of the vehicle, the entry critical time image of the vehicle, and the entrance tracking video of the vehicle with the presence vehicle information table.
  • the second threshold can be set according to actual needs. For example, the second threshold can be set to be 20-30 seconds. If the vehicle does not enter the parking space area and exits the entrance detection area, the tracking of the vehicle is stopped.
  • a vehicle such as vehicle E in FIG. 5 enters the exit detection area from a parking space in the parking space area (ie, the vehicle area and the license plate area enter from the parking space in the parking space area) Describe the field detection area), determine that the vehicle is an exit vehicle, and determine the exit parking event information.
  • the exit parking event information includes the license plate number, the exit time of the vehicle, the parking space where the vehicle is located when the exit parking event occurs, the image of the key point of the exit, the key timing image of the exit, and the appearance tracking video.
  • the presence vehicle information table is searched according to the license plate number. . If the license plate number has been previously recorded, i.e., the vehicle identified by the license plate is previously an entry vehicle, then the vehicle identified by the license plate is determined to be the exit vehicle and the exit parking event information for the vehicle.
  • the exit parking event information includes the license plate number, the playing time, and the parking space where the vehicle is identified by the license plate.
  • the vehicle tracking algorithm is used to intercept the vehicle area as the initial target vehicle area for tracking.
  • the vehicle enters the parking space in the parking space area from the entrance detection area across the parking line (ie, boundary 2 in FIG. 2) (ie, the target vehicle area is detected from the entrance).
  • the area is moved into the parking space in the parking space area, and when the time spent on the parking space exceeds a second threshold, the tracking of the vehicle is stopped and motion detection of the vehicle continues.
  • recording the vehicle parking time as an entrance parking event determining and storing the parking event information.
  • the arrival time of the vehicle and the parking space where the vehicle is located when the entrance parking event occurs are written into the on-site vehicle information table.
  • a vehicle such as vehicle E in FIG. 5 enters the exit detection area from a parking space in the parking space area (ie, the target vehicle area enters the exit from a parking space in the parking space area)
  • the detection area is determined to determine that the vehicle is an exit vehicle, and the exit parking event information is determined.
  • the exit parking event information includes the playing time, the parking space where the vehicle is located when the exit parking event occurs, the image of the key point of the exit, the key timing image of the exit, and the exit tracking video.
  • the license plate tracking algorithm is used to intercept the license plate image and the The area around the license plate image is tracked as the initial target license plate area.
  • the license plate (such as the license plate of the vehicle D in FIG. 5) enters the parking space in the parking space area from the entrance detection area beyond the parking line (ie, the boundary 2 in FIG. 2) (ie, the target license plate area) Moving into the parking space in the parking space area, and when the time spent on the parking space exceeds a second threshold, the tracking of the vehicle is stopped and motion detection of the vehicle continues.
  • the vehicle that records the license plate identifier is an entrance vehicle, and the parking is an entrance parking event (ie, the type of the parking event is an entrance parking event).
  • the information of the parking event is determined and stored, that is, the license plate number, the entrance time, and the parking space where the vehicle is located when the entrance parking event occurs are written into the presence vehicle information table. And associating and storing the entrance key image, the entry key time series image, and the entrance tracking video with the presence vehicle information table.
  • a license plate such as the license plate of vehicle E in FIG. 5 enters the exit detection area from a parking space in the parking space area (ie, the target license plate area enters from a parking space in the parking space area)
  • the exit detection area is determined, and the vehicle identified by the license plate is determined to be an exit vehicle, and the exit parking event information is determined.
  • the exit parking event information includes a license plate number, an exit time, a parking space where the vehicle is located when the exit parking event occurs, an image of a key point of the exit, a key timing image of the exit, and an exit tracking video.
  • the license plate (such as the license plate of the vehicle F in FIG. 5) is tracked in the exit detection area, and the license plate number of the license plate has been identified, the presence vehicle is searched according to the license plate number. Information Sheet. If the license plate number has been previously recorded, that is, the vehicle identified by the license plate is previously an entry vehicle, the vehicle identified by the license plate is determined to be the exit parking event information of the exit vehicle and the vehicle identified by the license plate.
  • the exit parking event information includes a license plate number, an exit time, and a parking space where the vehicle is located when the exit parking event occurs.
  • the information of the exit parking event such as the exit time and the license plate number of the vehicle is supplementally recorded. In this way, the problem that the vehicle exit parking event information is missing due to the failure to capture and recognize the vehicle during the vehicle exiting the parking space can be solved.
  • the target vehicle area in the process of tracking the target vehicle area, the latest image of the vehicle is intercepted (also in real time), and the target vehicle area is updated to prevent long-term tracking of the target vehicle area.
  • the goal is lost.
  • the latest license plate image of the vehicle and the image of the surrounding area of the license plate image are intercepted (also in real time), and the target license plate area is updated to prevent the target from being tracked for a long time.
  • the target caused by the license plate area is lost.
  • the target vehicle area may include an area around the license plate image in addition to, for example, an area including the license plate image, which can improve the accuracy of detecting the license plate.
  • the time of entering the first parking space of the vehicle is as a matter of course, the final parking space of the vehicle is subject to the last parking space.
  • the camera since the camera, especially the gun type camera, has the characteristics of stable viewing angle and focal length, it is possible to continuously and stably acquire stable images. Take the current 2 megapixel gun as an example. It is deployed on a 6-meter monitor. When the field of view covers 2-3 parking spaces, clear and stable vehicle images and license plate images can be obtained. The recognition accuracy is high and parking events can be obtained. A complete chain of evidence. Therefore, the present disclosure is based on the division of an image taken by a camera, and the obtained image regions can be processed at the same time, so that tracking and recognition of the vehicle entering and leaving the field can be realized. And when multiple concurrent parking events occur on several parking spaces monitored by a certain camera, multiple parking events can be simultaneously managed according to the divided areas.
  • the present disclosure can further improve the confidence and accuracy by jointly processing images captured by the camera.
  • the technical solution of the present disclosure can improve the management efficiency of the roadside parking event.
  • FIG. 6 is a schematic block diagram showing the structure of a parking management apparatus according to an embodiment of the present invention.
  • the apparatus is for performing the method provided in any of the above embodiments of the present disclosure.
  • the apparatus includes a processor 201 and a memory 203.
  • display 202 can also be included.
  • a watchdog (WD) 204 may also be included.
  • the various components of the device can be configured as follows.
  • the processor 201 is configured to receive a monitoring area image captured by the camera. And dividing the monitoring area image captured by the camera according to an image division rule of the camera to obtain a tracking detection area in the monitoring area image. And monitoring at least one of a vehicle and a license plate in the image of the surveillance area, and determining parking event information according to a tracking detection area in which at least one of the vehicle and the license plate is located.
  • the tracking detection area includes a parking space area, an entrance detection area, and an exit detection area.
  • the parking event information includes a license plate number of the vehicle, a type of the parking event, and a parking space where the vehicle is located when the parking event occurs.
  • the display 202 is configured to display an operation management interface, and preset an image division rule of the camera through an operation management interface.
  • the operation management interface receives a first query instruction input by a user, and the first query instruction is an operation for the user to select to query monitoring information of the camera.
  • the operation management interface acquires a first monitoring area image captured by the camera according to the first query instruction. And displaying the first monitoring area image.
  • the first monitoring area image is a captured area image of the newly captured image taken by the camera.
  • the operation management community receives the tracking detection area coordinates selected by the user in the first monitoring area image, and the coordinates of the tracking detection area are image division rules of the camera.
  • the memory 203 is used to store monitoring information for each camera.
  • the monitoring information includes: a surveillance area image captured by the camera, and an image obtained after the division (ie, a roadside parking monitoring area, a parking space area, an entrance detection area, an exit detection area, and the like, a tracking detection area image), an image division rule, and a Managed parking space information and parking event information occurring on the parking space.
  • the parking event information includes a type of parking event, a time when the parking event occurs, a parking space where the vehicle is located when the parking event occurs, a parking time of the vehicle, a license plate number of the vehicle, a key point image, a key time series image, a tracking video, and the like.
  • the watchdog 204 is used to monitor the operational status of the processor 201 and upload the operational status information of the processor 201 to the background server. When an abnormality occurs in the processor, the watchdog 204 controls the processor 201 to stop working or is stopped by the manager through the background server control processor 201.
  • the camera referred to in this embodiment may be a gun type camera.
  • the processor 201 can periodically send key point images, key time series images, and tracking videos of each parking event in the memory 203 to the server, and the user can find the key points of each parking event on the operation management interface. Images, key time series images to patrol parking events.
  • the server can identify and process the key time series image and the tracking video, and correct and supplement the parking event information determined by the device processor 201, such as the license plate number, etc., to further improve the license plate recognition rate and management efficiency of the device for the parking event.
  • the device can operate according to the following process.
  • the processor 201 stores the monitoring area image captured by the camera, the key point image, the tracking video, and the key time series image into the memory 203 via the input module 301.
  • the user can set the image segmentation rules of the camera and query the monitoring information of the camera through the operation management community on the display 202.
  • the operation management interface on the display 202 receives the first query instruction input by the user, that is, when the user selects to query the monitoring information of a certain camera, the operation management interface displays the latest captured monitoring area image of the camera (first monitoring)
  • the area image may be selected by the user on the first monitoring area image (which may be a click mode).
  • the coordinates of the vehicle tracking monitoring area (the coordinates of the parking space area, the entrance detection area, and the exit detection area) are used as the image of the camera.
  • the partitioning rules are stored to the storage module 305.
  • the processor 201 After the processor 201 subsequently receives the image of the monitoring area captured by the camera, the image of the monitoring area captured by each camera is divided according to the image dividing rule of each camera, and the tracking detection area in the image of the monitoring area is obtained, that is, the parking space area, the entrance Field detection area and exit detection area.
  • the processor 201 performs target detection, motion detection, license plate recognition, target vehicle tracking, and target license plate tracking on the vehicle tracking monitoring area image, thereby obtaining parking event information on the parking space.
  • the operation management interface can display information such as the monitoring area image, key point image, tracking video, key time series image, etc. according to the user's operation. For the user to inspect, supplement the parking event information.
  • the process of the specific detection, identification, and tracking of the vehicle in the image by the processor 201 may refer to the specific implementation process of step S104 in the method embodiment provided by the disclosure, and details are not described herein again.
  • the above method is performed by using the apparatus, that is, the image division rule of the camera is set by operating on a software (operation management interface).
  • the image captured by the camera is then divided by the processor according to the image division rule, and the obtained image regions are processed.
  • the detection, identification and tracking of vehicles entering and leaving the field can be managed simultaneously according to the divided areas.
  • the device does not have the influence of the camera scheduling response time (similar to the response time of the dome camera). Therefore, the method provided by the present disclosure can improve the management efficiency of the roadside parking event.
  • FIG. 7 is a schematic block diagram showing another structure of a parking management apparatus according to an embodiment of the present invention.
  • the apparatus is a method that can be used to perform any of the above-described embodiments of the present disclosure.
  • the apparatus includes an input module 301, an image division module 302, and a detection identification tracking module 303.
  • an operation management interface 304 can also be included.
  • the storage module 305 can also be included.
  • each module in the device can operate as follows.
  • the input module 301 is configured to receive a monitoring area image captured by the camera.
  • the image dividing module 302 is configured to divide the monitoring area image captured by the camera according to the image dividing rule of the camera, and obtain a tracking detection area in the monitoring area image.
  • the tracking detection area includes a parking space area, an entrance detection area, and an exit detection area.
  • the detection identification tracking module 303 is configured to monitor at least one of a vehicle and a license plate in the monitoring area image, and determine parking event information according to a tracking detection area where at least one of the vehicle and the license plate is located.
  • the detection identification tracking module 303 can include a vehicle detection module 3031, a license plate detection module 3032, a license plate recognition module 3033, a motion detection module 3034, a tracking processing module 3035, and an operation determination module 3036.
  • the license plate detection module 3032 is configured to perform license plate detection on the tracking detection area by using a license plate detection algorithm, and determine a license plate in the tracking detection area.
  • the license plate recognition module 3033 is configured to identify the license plate by using a license plate recognition algorithm to determine a license plate number of the license plate.
  • the operation determining module 3036 is configured to determine whether the license plate is a license plate of the vehicle.
  • the motion detection module 3034 is configured to perform motion detection on the vehicle and the license plate by using a motion detection algorithm to determine a motion state of the vehicle and the license plate; and the motion state of the target vehicle and the target license plate includes static and motion .
  • the tracking processing module 3035 is configured to determine parking event information according to the motion state of the vehicle and the license plate. For example, if the motion state of the vehicle or the license plate is stationary, the parking event information is determined based on the tracking detection area in which the vehicle or the license plate is located. If the motion state of the vehicle or license plate is motion, the vehicle is tracked and/or the license plate is tracked to determine parking event information.
  • the tracking processing module 3035 can include a vehicle tracking module 30351, a license plate tracking module 30352.
  • the vehicle tracking module 30351 is configured to track the vehicle using a vehicle tracking algorithm to determine parking event information.
  • the license plate tracking module 30352 is configured to track the license plate by using a license plate tracking algorithm to determine parking event information.
  • the operation management interface 304 is configured to display a monitoring area image captured by the camera, a tracking detection area, and parking event information. And the image of the parking space, the entrance detection area, and the exit detection area that are input when the user selects the tracking detection area are used as the image division rule of the camera, and then preset the image division rule of the camera.
  • the preset can be made as follows.
  • the operation management interface 304 receives the first query instruction input by the user.
  • the first query instruction is an operation in which the user selects to query the monitoring information of the camera.
  • the operation management interface 304 acquires the first monitoring area image captured by the camera according to the first query instruction, and displays the first monitoring area image.
  • the first monitoring area image is a captured area image of the newly captured image taken by the camera.
  • the operation management interface 304 receives the tracking detection area coordinates selected by the user in the first monitoring area image.
  • the coordinates of the tracking detection area are image division rules of the camera.
  • the storage module 305 is configured to store monitoring information of each camera.
  • the monitoring information includes: an image of the monitoring area captured by the camera, and an image obtained after the division (ie, a roadside parking monitoring area, a parking space area, an entrance detection area, an exit detection area, and the like, a tracking detection area image), an image division rule, The managed parking space information and the parking event information occurring on the parking space.
  • the parking event information includes a type of parking event, a time when the parking event occurs, a parking space where the parking event occurs, a parking time of the vehicle, a license plate number of the vehicle, a key point image, a key time series image, and a tracking video.
  • the working process of the device can be as follows.
  • the surveillance area image, the key point image, the tracking video, and the key time series image captured by the camera are stored to the storage module 305 via the input module 301.
  • the user can select an image segmentation rule of the camera and query the monitoring information of the camera through the operation management interface 304.
  • the operation management interface 304 receives the first query instruction input by the user, that is, when the user selects to query the monitoring information of a certain camera, the operation management interface 304 displays the newly detected monitoring area image of the camera (the first monitoring area image) ).
  • the user may select (in a manner of clicking) the coordinates of the vehicle tracking monitoring area (the coordinates of the parking space area, the entrance detection area, and the exit detection area) on the first monitoring area image as the image division rule of the camera is stored to Storage module 305.
  • the image division module 302 divides the image of the surveillance area captured by the camera according to the image division rule of each camera to obtain a vehicle including a parking space area, an entrance detection area, and an exit detection area.
  • the area of the monitoring area is tracked and written into the storage module 303. And sending the vehicle tracking monitoring area image to the detection identification tracking module 303 for processing.
  • the startup vehicle detection module 3031 performs vehicle detection on the tracking detection area by using a vehicle detection algorithm, and determines the vehicle in the tracking detection area.
  • the activation license plate detection module 3032 performs license plate detection on the tracking detection area by using a license plate detection algorithm, and determines a license plate in the tracking detection area.
  • the activation license plate recognition module 3033 identifies the license plate by using a license plate recognition algorithm to determine a license plate number of the license plate.
  • the operation judging module 3036 and the motion detecting module 3034 are respectively activated.
  • the operation judging module 3036 determines whether the license plate is a license plate of the vehicle based on a positional relationship between the vehicle and the license plate.
  • the start motion detection module 3034 performs motion detection on the vehicle and the license plate by using a motion detection algorithm to determine a motion state of the vehicle and the license plate.
  • the tracking processing module 3035 determines the parking event information based on the tracking detection area in which the vehicle or the license plate is located. For example, if the vehicle is located in an entrance detection area or an exit detection area, the tracking processing module 3035 calculates the time at which the vehicle is stuck in the admission parking area. Or if the license plate is located in the entrance detection area or the exit detection area, calculate the time when the vehicle of the license plate identification stays in the entrance parking area. If the vehicle or the vehicle identified by the license plate is stuck in the entrance detection area exceeding a first threshold, it is determined that the vehicle or the vehicle identified by the license plate is a parking vehicle.
  • the tracking processing module 3035 looks up the presence vehicle information table based on the license plate number. If the license plate number has not been previously recorded, it is determined that the identified vehicle is an entry vehicle, and the admission parking event information of the vehicle is determined.
  • the tracking processing module 3035 tracks the vehicle and/or tracks the license plate. For example, if the vehicle detection module 3031 detects the vehicle in the tracking detection area, and the license plate detection module 3032 detects that the license plate and/or license plate recognition module 3033 of the vehicle recognizes the license plate number of the license plate, the tracking processing module 3035 initiates vehicle tracking module 30351 to track the vehicle. And the activation license plate tracking module 30352 tracks the license plate.
  • the tracking processing module 3035 activates the vehicle tracking module 3035 to track the vehicle.
  • the processing module 3035 activates the license plate tracking module 30352 to track the license plate.
  • the tracking processing module 3035 calculates the time the vehicle is staying on the parking space. If the time the vehicle stays on the parking space exceeds a second threshold, it is determined that the parking is an admission parking event, and the admission parking event information is determined. If the time the vehicle is parked on the parking space exceeds a second threshold, the tracking processing module 3035 stops tracking the vehicle and the motion detection module 3034 continues to perform motion detection on the vehicle. If the vehicle of the license plate identification stays on the parking space exceeds a second threshold, the tracking processing module 3035 stops tracking the license plate, and the motion detection module 3034 continues to perform motion detection on the license plate.
  • the tracking processing module 3035 looks up the presence vehicle information table. If the license plate number has been previously recorded, it is determined that the vehicle identified by the license plate is an exit vehicle, and the parking stop event information of the vehicle that determines the license plate identification.
  • the operation management interface 304 can display parking event information such as key point images, key time series images, and tracking videos according to the operation of the user. For the user to inspect, supplement the parking event information.
  • the specific implementation process of the method provided in any one of the foregoing embodiments may refer to a specific process in the method embodiment, and details are not described herein again.
  • the above method is performed by using the apparatus based on the method embodiment provided by the present disclosure. That is, the image segmentation rule of the camera is set by operating on the software (operation management interface), and then the image captured by the camera is divided by the processor according to the image division rule, and the obtained image regions are processed and the vehicle is processed. Tracking and identification of entry and exit. And when multiple concurrent parking events occur on several parking spaces monitored by one camera, multiple parking events can be managed simultaneously according to the divided areas. Compared with the management mode of the dome camera, it is not necessary to consider the scheduling response time of the camera. Therefore, the method provided by the embodiment of the present disclosure can improve the management efficiency of the roadside parking event.
  • FIG. 8 is a schematic block diagram showing the structure of a parking management system according to an embodiment of the present invention.
  • the system includes a camera group 401, a parking management device 402 shown in FIG. 6 or FIG.
  • the system can also include a switch 403, a cloud server 404.
  • the camera group 401, the parking management device 402, and the cloud server 404 can all communicate through the switch.
  • the camera group 401 includes N camera groups, and the value of N is not less than 2.
  • the specific value may be determined according to actual conditions.
  • Each camera set includes at least one camera array, such as camera set 1 including camera array 11, camera set 2 including camera array 21 and camera array 22, ..., camera set N including camera array N1 and camera array N2, camera set (N +1) includes the camera array (N+1) 1.
  • Each camera array includes at least one camera, and the specific number may be determined according to actual conditions.
  • the camera is preferably a gun type camera.
  • Each camera set is mounted on the pole position of the roadside parking lot, and a camera set is mounted on one pole position.
  • the camera groups on the first and last pole positions are camera groups that contain one camera array, and the camera groups on the remaining pole positions are camera groups that contain two camera arrays. This allows the parking space between every two pole positions to be monitored and managed from both sides through two camera arrays on both pole positions.
  • Each camera array manages a certain number of parking spaces. In some embodiments, each camera in the camera array can be set to manage 2 to 5 parking spaces. After the camera group is installed on the pole position, the cameras are aligned with the managed parking spaces, the focal length and shooting angle of each camera are set, and the roadside parking monitoring area and parking space managed by each camera are determined to obtain monitoring information. .
  • the various components of the system can be configured as follows.
  • the camera group 401 is used to acquire an image of the surveillance area on the managed parking space.
  • the parking management device 402 is configured to perform image division on the monitoring area image according to an image division rule. And detecting and identifying the tracking image of the monitoring area image, and determining parking event information on the parking space managed by the camera group 401.
  • the switch 403 is configured to connect the camera group 401, the multi-camera based parking management device 402, and the cloud server 404 to perform communication.
  • the cloud server 404 is used to manage parking event information.
  • the parking event information determined by the parking management device 402 may be stored, and data such as key time series images, tracking videos, and the like may be collectively detected and identified, and the information missing from the parking event may be supplemented.
  • the working process of the system can be performed as follows.
  • the camera group 401 acquires the image of the monitoring area on the parking space and transmits it to the parking management device 402 via the switch 403.
  • the parking management device 402 performs image division on the monitoring area image according to an image division rule. And detecting and identifying the tracking image of the monitoring area image, and determining parking event information on the parking space managed by the camera group 401.
  • the parking management device 402 uploads the determined parking event information to the cloud server 404 via the switch.
  • the cloud server 404 stores the parking event information, and performs batch detection and identification tracking processing on data such as key time series images and tracking videos to supplement information missing from the parking event.
  • the parking management device 402 can perform image division and detection recognition processing on the side parking monitoring area image. Supplement the parking event information of the vehicle on the original parking space.
  • the camera that manages the parking space in the camera array 32 fails to capture the image of the vehicle when it exits the parking space.
  • the camera in the camera array 22 transmits the captured image to the parking management device 402 via the switch 403.
  • the parking management device 402 can supplement the exit parking event information when the vehicle exits the parking space 3j by the image division and detection recognition processing of the monitoring area image.
  • the system is configured to form a cluster by using a camera group composed of a camera array.
  • the camera array By installing the camera array, the parking space between the two pole positions is monitored by the first and last camera arrays in real time, thereby improving the capturing efficiency.
  • the joint detection and recognition processing of the monitoring area image captured by the camera between the pole positions by the multi-camera based parking management device 402 can improve the efficiency and completeness of determining the parking information of the road, thereby improving the roadside parking as a whole. Management efficiency.
  • a computer readable storage medium having stored thereon a computer program that, when executed by a processor, implements a multi-camera based roadside parking management method in any of the above embodiments.
  • the computer readable storage medium is a non-transitory computer readable storage medium.
  • the multi-camera based roadside parking management method the multi-camera based roadside parking management apparatus, the multi-camera based roadside parking management system, and the computer readable storage medium according to the present disclosure have been described in detail. In order to avoid obscuring the concepts of the present disclosure, some details known in the art are not described. Those skilled in the art can fully understand how to implement the technical solutions disclosed herein according to the above description.
  • the methods and systems of the present disclosure may be implemented in a number of ways.
  • the methods and systems of the present disclosure may be implemented in software, hardware, firmware, or any combination of software, hardware, or firmware.
  • the above-described sequence of steps for the method is for illustrative purposes only, and the steps of the method of the present disclosure are not limited to the order specifically described above unless otherwise specifically stated.
  • the present disclosure may also be embodied as programs recorded in a recording medium, the programs including machine readable instructions for implementing a method in accordance with the present disclosure.
  • the present disclosure also covers a recording medium storing a program for executing the method according to the present disclosure.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Signal Processing (AREA)
  • Traffic Control Systems (AREA)
  • Devices For Checking Fares Or Tickets At Control Points (AREA)

Abstract

一种基于多摄像机的路侧停车管理方法、装置和系统，该方法包括：接收摄像机拍摄的监控区域图像，对所述监控区域图像进行划分，得到所述监控区域图像中的各跟踪检测区域，对所述监控区域图像中的车辆和车牌中的至少一种进行监测，根据所述车辆和车牌中的至少一种所在的跟踪检测区域确定停车事件信息。该方法能够提高停车事件的管理效率。

Description

基于多摄像机的路侧停车管理方法、装置和系统
相关申请的交叉引用
本申请是以CN申请号为201711135187.6，申请日为2017年11月16日的申请为基础，并主张其优先权，该CN申请的公开内容在此作为整体引入本申请中。
技术领域
本公开涉及目标检测技术领域，特别涉及一种停车管理方法、停车管理装置、停车管理系统和计算机可读存储介质。
背景技术
随着城市车辆的快速增长，很多城市不得不在道路的两侧设置停车位，以对停车车辆进行统一地管理。这类停车场一般都是全开放式的停车场，车辆进出不受道闸等护栏装置的约束，因此这类停车通常称为路侧停车(也称路侧占道停车或路内停车)。而路侧停车中的平行停车场(即车位首尾相连形成一排的停车场,以下称路侧平行停车场)由人工进行管理，这种管理方式存在着停车计费不规范、计时精度低、漏收费、乱收费、停车证据取证不足且难追溯、不能全天候管理等问题，且人力管理的工作强度大、工作效率低、安全隐患大，因此目前的路侧停车一直是城市管理的难题。
近年来关于路侧停车的管理方案中，部分城市试行了基于地磁传感器的路侧停车管理方案和基于枪球相机联动的路侧停车管理方案。其中，基于枪球相机联动的管理方案主要是通过采用常用的主从式枪球联动相机，例如包括一组枪机(即多个安装视角和焦距固定的相机组，后文简称枪机)和一台球机(即能自动调整抓拍角度和焦距的相机，后文简称球机)，在使用时，将所述枪机和球机部署于路侧的L型监控杆的横臂上，通过所述L型监控杆上的枪机和球机的联动(即将枪机调成固定焦距和抓拍角度以检测停车事件，球机根据枪机的检测结果调整焦距和角度以捕捉目标车辆的图像，完成车辆的停车取证。)，在兼顾事件检测的同时和对车辆图像、车牌进行抓拍，实现大范围多车位的停车事件抓拍。
发明内容
本公开的发明人发现上述相关技术中存在如下问题：枪机和球机的视角均受安装 杆位的位置所限，对同一个目标车辆只能从单个方向进行抓拍。而且球机在被调用时，需要经过一段时间(即响应时间)，才能实施抓拍任务。以上情况限制了球机在实施抓拍任务时，只能一次一次抓拍，每次只能抓拍一个，从而影响球机抓拍效率，导致后续一连串诸如漏抓拍、识别效率低以及出现并发停车事件时，管理复杂，管理效率低的问题。针对所述问题中的至少一个问题，本公开提出了一种停车管理技术方案，能够提高停车管理效率。
根据本公开的一些实施例，提供了一种停车管理方法，包括：接收摄像机拍摄的监控区域图像；根据预设的图像划分规则，对所述监控区域图像进行划分，得到所述监控区域图像中的各跟踪检测区域；对所述监控区域图像中的车辆和车牌中的至少一种进行监测，根据所述车辆和车牌中的至少一种所在的跟踪检测区域确定停车事件信息。
可选地，所述跟踪检测区域图像包含车位区域、入场检测区域、出场检测区域；所述停车事件信息包括车辆的车牌号、停车事件的类型、所述停车事件发生时所述车辆所在的车位。
可选地，所述入场检测区域包含与所述车位区域的边界线临近的区域；所述出场检测区域为所述车位区域之外的区域。
可选地，从所述摄像机拍摄的监控区域图像中获取一张监控区域图像作为第一监控区域图像；在所述第一监控区域图像中设定跟踪检测区域的坐标；根据所述坐标对所述摄像机后续拍摄的监控区域图像进行划分。
可选地，在所述跟踪检测区域中进行车辆检测，确定所述各跟踪检测区域中的车辆；在所述跟踪检测区域中进行车牌检测，确定所述各跟踪检测区域中的车牌；对所述车牌进行识别，确定所述车牌的车牌号；对所述车辆、所述车牌进行运动检测，确定所述车辆、所述车牌的运动状态，所述运动状态包括静止和运动；根据所述车辆的运动状态，确定所述车辆的停车事件信息；以及根据所述车牌的运动状态，确定所述车牌标识的车辆的停车事件信息。
可选地，获取所述车牌与所述车辆的位置关系；在所述车牌的车牌区域位于所述车辆的车辆区域内的情况下，确定所述车牌检测和所述车辆检测的结果为既检测到所述车辆，又检测到所述车辆的车牌和/或识别到所述车牌的车牌号；在所述车牌的车牌区域没有位于任何车辆的车辆区域内的情况下，确定所述车牌检测和车辆检测的结果为仅检测到所述车牌和/或识别到所述车牌的车牌号，但未检测到所述车牌标识的车 辆；在所述车辆的车辆区域没有包含任何车牌的车牌区域的情况下，确定所述车牌检测和所述车辆检测的结果为仅检测到所述车辆，但未检测到所述车辆的车牌和所述车牌的车牌号。
可选地，在所述车辆的运动状态为静止的情况下，根据所述车辆所在的跟踪检测区域，确定所述车辆的停车事件信息；在所述车牌的运动状态为静止的情况下，根据所述车牌所在的跟踪检测区域，确定所述车牌标识的车辆的停车事件信息；在所述车辆的运动状态为运动的情况下，对所述车辆进行跟踪，确定所述车辆的停车事件信息；在所述车牌的运动状态为运动的情况下，对所述车牌进行跟踪，确定所述车牌标识的车辆的停车事件信息。
可选地，所述根据所述车牌所在的跟踪检测区域，确定所述车牌的停车事件信息，包括：在所述车牌位于车位区域，以及识别到所述车牌的车牌号的情况下，根据所述车牌号，查找在场车辆信息表中是否记录有所述车牌号；在所述在场车辆信息表没有记录所述车牌号的情况下，确定所述车牌标识的车辆为入场车辆，以及确定并记录所述车牌标识的车辆的车牌号、所述车辆的入场时间、所述车辆所在的车位；在所述车牌位于入场检测区域或者出场检测区域的情况下，计算所述车牌滞留在入场停车区域或者出场停车区域的时间；在所述时间超过第一阈值的情况下，确定所述车牌标识的车辆为违规车辆，所述车牌标识的车辆的停车事件为违规停车事件。
可选地，所述根据所述车辆所在的跟踪检测区域，确定所述车辆的停车事件信息，包括：在所述车辆位于车位区域，以及未识别到所述车辆的车牌号的情况下，根据所述车牌，查找所述在场车辆信息表在是否记录有所述车辆；在所述在场车辆信息表没有记录所述车辆的情况下，确定所述车辆为入场车辆，以及确定并记录所述车辆的入场时间、所述车辆所在的车位；在所述车辆位于入场检测区域或者出场检测区域的情况下，计算所述车辆滞留在入场停车区域或者出场停车区域的时间；在所述时间超过所述第一阈值的情况下，确定所述车辆为违规车辆，所述车辆的停车事件为违规停车事件。
可选地，在检测到所述跟踪检测区域图像中的车辆，以及检测到所述车辆的车牌或识别到所述车牌号的情况下，截取所述车辆区域作为初始的目标车辆区域进行跟踪，以及截取所述车牌区域及所述车牌周边的区域作为初始的目标车牌区域进行跟踪；在检测到所述跟踪检测区域中的车辆，但未检测到所述车辆的车牌的情况下，截取所述车辆区域作为初始的目标车辆区域进行跟踪；在未检测到所述跟踪检测区域图 像中的车辆，但检测到所述车辆的车牌或识别到所述车牌号的情况下，截取所述车牌图像及所述车牌图像周边的区域作为初始的目标车牌区域进行跟踪。
可选地，截取所述车辆的最新图像，对所述目标车辆区域进行更新；截取所述车辆的最新车牌图像及所述车牌图像周边区域的图像，对所述目标车牌区域进行更新。
可选地，在所述车辆从所述入场检测区越过车位线进入所述车位区域中的车位的情况下，计算所述车辆在所述车位上停留的时间；在所述车牌从所述入场检测区越过车位线进入所述车位区域中的车位的情况下，计算所述车牌在所述车位上停留的时间。
可选地，在所述车辆在所述车位上停留的时间超过第二阈值的情况下，确定所述车辆为入场车辆、所述车辆的停车事件为入场停车事件，以及确定入场停车事件信息，并将所述入场停车事件信息记录到在场停车信息表中。
可选地，在所述车牌在所述车位上停留的时间超过第二阈值的情况下，确定所述车牌标识的车辆为入场车辆、所述车牌标识的车辆的停车事件为入场停车事件，以及确定入场停车事件信息，并将所述入场停车事件信息记录到在场停车信息表中，所述入场停车事件信息包括车牌号、入场时间、所述入场停车事件发生时车辆所在的车位、入场关键点图像、入场关键时序图像、入场跟踪视频。
可选地，在所述车辆在所述车位上停留的时间超过所述第二阈值的情况下，停止跟踪所述车辆，对所述车辆的运动检测继续进行；在所述车牌在所述车位上停留的时间超过所述第二阈值的情况下，停止跟踪所述车牌，对所述车牌的运动检测继续进行。
可选地，所述入场关键点图像包括：所述车辆在入场检测区域中首次被检测到的图像、所述车辆在入场检测区域中首次被检测到的车牌图像、所述车辆从入场检测区域越过车位线进入车位区域的图像、所述车辆停在车位上的图像。
可选地，所述入场关键时序图像为从所述入场跟踪视频中选择提取得到的过程图像。
可选地，所述入场跟踪视频为所述车辆在入场检测区域中首次被检测到至所述车辆停在车位上的视频段。
可选地，在所述车辆从所述车位区域中的车位进入所述出场检测区域的情况下，确定所述车辆为出场车辆，以及确定所述车辆的出场停车事件信息；在所述车牌从所述车位区域中的车位进入所述出场检测区域的情况下，确定所述车牌标识的车辆为出场车辆，以及确定所述车牌标识的车辆的出场停车事件信息；所述出场停车事件信息 包括车牌号、出场时间、所述出场停车事件发生时车辆所在的车位、出场关键点图像、出场关键时序图像、出场跟踪视频。
可选地，所述出场关键点图像包括所述车辆离场前停在车位上的图像、所述车辆在出场检测区域中首次被检测到的图像、所述车辆在出场检测区域中首次被检测到的车牌图像、所述车辆从车位区域越过车位线进入出场检测区域的图像、所述车辆在出场检测区域中最后一次被检测到的图像、所述车辆在出场检测区域中最后一次被检测到的车牌图像。
可选地，所述出场关键时序图像为从所述出场跟踪视频中选择提取得到的过程图像。
可选地，所述出场跟踪视频包括第一出场跟踪视频、第二出场跟踪视频，所述第一出场跟踪视频为所述车辆从车位区域进入出场检测区域直至从出场检测区域消失的视频段，所述第二出场跟踪视频为所述车辆在出场检测区域中首次被检测到至所述车辆在出场检测区域中消失的视频段。
可选地，在所述出场检测区域中跟踪到车牌，以及识别所述车牌的车牌号的情况下，根据所述车牌号，查找在场车辆信息表中是否已记录所述车牌号；在所述车牌号之前已被记录的情况下，确定所述车牌标识的车辆为出场车辆，以及所确定述车牌标识的车辆的车牌号、出场时间、所述车牌标识的车辆所在的车位。
根据本公开的另一些实施例，提供一种停车管理装置，包括：处理器，用于接收摄像机拍摄的监控区域图像；以及对所述监控区域图像进行划分，得到所述监控区域图像中的跟踪检测区域；以及对所述跟踪检测区域进行车辆、车牌监测，确定停车事件信息；存储器，用于存储所述监控区域图像、停车事件信息。
可选地，所述跟踪检测区域图像包含车位区域、入场检测区域、出场检测区域；所述停车事件信息包括所述车辆的车牌号、停车事件的类型、所述停车事件发生时所述车辆所在的车位。
可选地，所述停车管理装置还包括显示器，用于显示操作管理界面；所述操作管理界面用于接收用户输入的第一查询指令；所述第一查询指令为用户选择查询所述摄像机的监控信息的操作；所述操作管理界面根据所述第一查询指令，获取所述摄像机拍摄的监控区域图像，以及将所述监控区域图像显示；所述操作管理界面接收用户在所述监控区域图像中选择的跟踪检测区域坐标；所述跟踪检测区域的坐标为划分所述监控区域图像的依据。
可选地，所述处理器在所述各跟踪检测区域中进行车辆检测，确定所述各跟踪检测区域中的车辆，所述处理器在所述各跟踪检测区域中进行车牌检测，确定所述各跟踪检测区域中的车牌，所述处理器对所述车牌进行识别，确定所述车牌的车牌号，所述处理器对所述车辆、所述车牌进行运动检测，确定所述车辆、所述车牌的运动状态，所述运动状态包括静止和运动，所述处理器根据所述车辆、所述车牌的运动状态确定所述车辆、所述车牌的所述停车事件信息。
可选地，所述处理器获取所述车牌与所述车辆的位置关系，所述处理器在所述车牌的车牌区域位于所述车辆的车辆区域内的情况下，确定所述车牌检测和所述车辆检测的结果为既检测到所述车辆，又检测到所述车牌和/或所述车牌的车牌号，所述处理器在所述车牌的车牌区域没有位于任何车辆的车辆区域内的情况下，确定所述车牌检测和车辆检测的结果为仅检测到所述车牌和/或所述车牌的车牌号，但未检测到所述车牌标识的车辆，所述处理器在所述车辆的车辆区域没有包含任何车牌的车牌区域的情况下，确定所述车牌检测和所述车辆检测的结果为仅检测到所述车辆，但未检测到所述车辆的车牌和所述车牌的车牌号。
可选地，所述处理器在所述车辆的运动状态为静止的情况下，根据所述车辆所在的跟踪检测区域，确定所述车辆的停车事件信息，所述处理器在所述车牌的运动状态为静止的情况下，根据所述车牌所在的跟踪检测区域，确定所述车牌标识的车辆的停车事件信息，所述处理器在所述车辆的运动状态为运动的情况下，对所述车辆进行跟踪，确定所述车辆的停车事件信息，所述处理器在所述车牌的运动状态为运动的情况下，对所述车牌进行跟踪，确定所述车牌标识的车辆的停车事件信息。
可选地，所述处理器在所述车牌位于车位区域，以及识别到所述车牌的车牌号的情况下，根据所述车牌号，查找在场车辆信息表中是否记录有所述车牌号,所述处理器在所述在场车辆信息表没有记录所述车牌号的情况下，确定所述车牌标识的车辆为入场车辆，以及确定并记录所述车牌标识的车辆的车牌号、所述车辆的入场时间、所述车辆所在的车位，所述处理器在所述车辆位于车位区域，以及未识别到所述车辆的车牌号的情况下，根据所述车牌，查找所述在场车辆信息表在是否记录有所述车辆，所述处理器在所述在场车辆信息表没有记录所述车辆的情况下，确定所述车辆为入场车辆，以及确定并记录所述车辆的入场时间、所述车辆所在的车位。
可选地，所述处理器在所述车辆位于车位区域，以及未识别到所述车辆的车牌号的情况下，根据所述车辆，查找所述在场车辆信息表是否记录有所述车辆；在所述在 场车辆信息表没有记录所述车辆的情况下，确定所述车辆为入场车辆，以及确定并记录所述车辆的入场停车事件信息；所述车辆的入场停车事件信息包括车辆的入场时间、车辆所在的车位；所述处理器在所述车辆位于入场检测区域或者出场检测区域的情况下，计算所述车辆滞留在入场停车区域或者出场停车区域的时间，在所述时间超过所述第一阈值的情况下，确定所述车辆为违规车辆，所述车辆的停车事件为违规停车事件。
可选地，所述处理器在检测到所述跟踪检测区域图像中的车辆，以及检测到所述车辆的车牌或车牌号中的情况下，截取所述车辆区域作为初始的目标车辆区域进行跟踪，以及截取所述车牌区域及所述车牌周边的区域作为初始的目标车牌区域进行跟踪，所述处理器在检测到所述跟踪检测区域中的车辆，但未检测到所述车辆的车牌的情况下，截取所述车辆区域作为初始的目标车辆区域进行跟踪，所述处理器在未检测到所述跟踪检测区域图像中的车辆，但检测到所述车辆的车牌或车牌号中的情况下，截取所述车牌图像及所述车牌图像周边的区域作为所述初始的目标车牌区域进行跟踪。
可选地，所述处理器截取所述车辆最新的图像，对所述目标车辆区域进行更新；所述处理器截取所述车辆最新的车牌图像及车牌图像周边区域的图像，对所述目标车牌区域进行更新。
可选地，所述处理器在所述车辆从所述入场检测区越过车位线进入所述车位区域中的车位的情况下，计算所述车辆在所述车位上停留的时间；所述处理器在所述车牌从所述入场检测区越过车位线进入所述车位区域中的车位的情况下，计算所述车牌在所述车位上停留的时间；所述处理器在所述车辆在所述车位上停留的时间超过第二阈值的情况下，确定所述车辆为入场车辆、所述车辆的停车事件为入场停车事件，以及确定入场停车事件信息，并将所述入场停车事件信息记录到在场停车信息表中；所述处理器在所述车牌在所述车位上停留的时间超过第二阈值的情况下，确定所述车牌标识的车辆为入场车辆、所述车牌标识的车辆的停车事件为入场停车事件，以及确定入场停车事件信息，并将所述入场停车事件信息记录到在场停车信息表中；所述入场停车事件信息包括车牌号、入场时间、所述入场停车事件发生时车辆所在的车位、入场关键点图像、入场关键时序图像、入场跟踪视频。
可选地，所述处理器在所述车辆在所述车位上停留的时间超过所述第二阈值的情况下，停止跟踪所述车辆，对所述车辆的运动检测继续进行；所述处理器在所述车牌 在所述车位上停留的时间超过所述第二阈值的情况下，停止跟踪所述车牌，对所述车牌的运动检测继续进行。
可选地，所述入场关键点图像包括：所述车辆在入场检测区域中首次被检测到的图像、所述车辆在入场检测区域中首次被检测到的车牌图像、所述车辆从入场检测区域越过车位线进入车位区域的图像、车辆停在车位上的过程图像。
可选地，所述入场关键时序图像为从所述入场跟踪视频中选择提取得到的图像。
可选地，所述入场跟踪视频为所述车辆在入场检测区域中首次被检测到至所述车辆停在车位上的视频段。
可选地，所述处理器在所述车辆从所述车位区域中的车位进入所述出场检测区域的情况下，确定所述车辆为出场车辆，以及确定所述车辆的出场停车事件信息；所述处理器在所述车牌从所述车位区域中的车位进入所述出场检测区域的情况下，确定所述车牌标识的车辆为出场车辆，以及确定所述车牌标识的车辆的出场停车事件信息；所述出场停车事件信息包括车牌号、出场时间、出场停车事件发生时车辆所在的车位、出场关键点图像、出场关键时序图像、出场跟踪视频。
可选地，所述出场关键点图像包括所述车辆离场前停在车位上的图像、所述车辆在出场检测区域中首次被检测到的图像、所述车辆在出场检测区域中首次被检测到的车牌图像、所述车辆从车位区域越过车位线进入出场检测区域的图像、所述车辆在出场检测区域中最后一次被检测到的图像、所述车辆在出场检测区域中最后一次被检测到的车牌图像。
可选地，所述出场关键时序图像为从所述出场跟踪视频中选择提取得到的过程图像。
可选地，所述出场跟踪视频包括第一出场跟踪视频、第二出场跟踪视频；所述第一出场跟踪视频为所述车辆从车位区域进入出场检测区域直至从出场检测区域消失的视频段；所述第二出场跟踪视频为所述车辆在出场检测区域中首次被检测到至所述车辆停在出场检测区域中消失的视频段。
可选地，所述处理器在所述出场检测区域中跟踪到车牌，以及识别所述车牌的车牌号的情况下，根据所述车牌号，查找在场车辆信息表中是否已记录所述车牌号，所述处理器在所述车牌号之前已被记录的情况下，确定所述车牌标识的车辆为出场车辆，以及所确定述车牌标识的车辆的车牌号、出场时间、所述车牌标识的车辆所在的车位。
根据本公开的又一些实施例，提供一种停车管理装置，包括：输入模块，用于接收摄像机拍摄的监控区域图像；图像划分模块，用于根据预设的图像划分规则，对所述监控区域图像进行划分，得到所述监控区域图像中的跟踪检测区域；检测识别跟踪模块，用于对所述跟踪检测区域进行车辆、车牌监测，确定停车事件信息。
可选地，所述跟踪检测区域图像包含车位区域、入场检测区域、出场检测区域；所述停车事件信息包括所述车辆的车牌号、停车事件的类型、所述停车事件发生时所述车辆所在的车位。
可选地，所述入场检测区域包含所述车位区域边界线两侧的临近区域；所述出场检测区域为车位区域之外的区域。
可选地，所述停车管理装置还包括操作管理界面，用于接收用户输入的第一查询指令，所述第一查询指令为用户选择查询所述摄像机的监控信息的操作，根据所述第一查询指令，获取所述摄像机拍摄的第一监控区域图像，以及将所述第一监控区域图像显示，接收用户在所述第一监控区域图像中选择的跟踪检测区域坐标；所述跟踪检测区域的坐标为所述摄像机的图像划分规则。
可选地，所述跟踪检测区域的坐标包含车位区域的坐标、入场检测区域的坐标、出场检测区域的坐标。
可选地，所述检测识别跟踪模块包括：车辆检测模块，用于采用车辆检测算法对跟踪检测区域进行车辆检测，确定所述跟踪检测区域中的车辆；车牌检测模块，用于采用车牌检测算法对跟踪检测区域进行车牌检测，确定所述跟踪检测区域中的车牌；车牌识别模块，用于采用车牌识别算法对所述车牌进行识别，确定所述车牌的车牌号；运动检测模块，用于采用运动检测算法对所述车辆、所述车牌进行运动检测，确定所述车辆、所述车牌的运动状态，所述目标车辆、所述目标车牌的运动状态包括静止和运动；跟踪处理模块，用于根据所述车辆、所述车牌的运动状态确定所述车辆、所述车牌的所述停车事件信息。
可选地，所述停车管理装置还包括：运算判断模块，用于获取所述车牌与所述车辆的位置关系，运算判断模块在所述车牌的车牌区域位于所述车辆的车辆区域内的情况下，确定所述车牌检测和所述车辆检测的结果为既检测到所述车辆，又检测到所述车辆的车牌和/或所述车牌的车牌号；运算判断模块在所述车牌的车牌区域没有位于任何车辆的车辆区域内的情况下，确定所述车牌检测和车辆检测的结果为仅检测到所述车牌和/或所述车牌的车牌号，但未检测到所述车牌标识的车辆；运算判断模块在所述 车辆的车辆区域没有包含任何车牌的车牌区域的情况下，确定所述车牌检测和所述车辆检测的结果为仅检测到所述车辆，但未检测到所述车辆的车牌和所述车牌的车牌号。
可选地，所述跟踪处理模块在所述车辆的运动状态为静止的情况下，根据所述车辆所在的跟踪检测区域，确定所述车辆的停车事件信息，所述跟踪处理模块在所述车牌的运动状态为静止的情况下，根据所述车牌所在的跟踪检测区域，确定所述车牌标识的车辆的停车事件信息，所述跟踪处理模块在所述车辆的运动状态为运动的情况下，对所述车辆进行跟踪，确定所述车辆的停车事件信息，所述跟踪处理模块在所述车牌的运动状态为运动的情况下，对所述车牌进行跟踪，确定所述车牌标识的车辆的停车事件信息。
可选地，在所述车牌位于车位区域，以及所述车牌识别模块识别到所述车牌的车牌号的情况下，所述跟踪处理模块根据所述车牌号，查找在场车辆信息表中是否记录有所述车牌号，在所述在场车辆信息表没有记录所述车牌号的情况下，确定所述车牌标识的车辆为入场车辆，以及确定并记录所述车牌标识的车辆的车牌号、所述车辆的入场时间、所述车辆所在的车位。
可选地，在所述车牌位于入场检测区域或者出场检测区域的情况下，所述跟踪处理模块计算所述车牌滞留在入场停车区域或者出场停车区域的时间，在所述时间超过第一阈值的情况下，确定所述车牌标识的车辆为违规车辆，所述车牌标识的车辆的停车事件为违规停车事件。
可选地，在所述车辆位于车位区域，以及所述车牌识别模块未识别到所述车辆的车牌号的情况下，所述跟踪处理模块根据所述车牌，查找所述在场车辆信息表是否记录有所述车辆，在所述在场车辆信息表没有记录所述车辆的情况下，确定所述车辆为入场车辆，以及确定并记录所述车辆的入场停车事件信息，所述车辆的入场停车事件信息包括车辆的入场时间、车辆所在的车位。
可选地，在所述车辆位于入场检测区域或者出场检测区域的情况下，所述跟踪处理模块计算所述车辆滞留在入场停车区域或者出场停车区域的时间，在所述时间超过所述第一阈值的情况下，确定所述车辆为违规车辆，所述车辆的停车事件为违规停车事件。
可选地，在所述车辆检测模块检测到所述跟踪检测区域图像中的车辆，以及所述车牌检测模块检测到所述车辆的车牌或所述车牌识别模块识别到车牌号的情况下，所 述跟踪处理模块启动车辆跟踪模块截取所述车辆区域作为初始的目标车辆区域进行跟踪，以及启动车牌跟踪模块截取所述车牌区域及所述车牌周边的区域作为初始的目标车牌区域进行跟踪；在所述车辆检测模块检测到所述跟踪检测区域中的车辆，但所述车牌检测模块未检测到所述车辆的车牌的情况下，所述跟踪处理模块启动车辆跟踪模块截取所述车辆区域作为初始的目标车辆区域进行跟踪；在所述车辆检测模块未检测到所述跟踪检测区域图像中的车辆，但所述车牌检测模块检测到所述车辆的车牌或所述车牌识别模块识别到车牌号的情况下，所述跟踪处理模块启动车牌跟踪模块截取所述车牌图像及所述车牌图像周边的区域作为所述初始的目标车牌区域进行跟踪。
可选地，所述车辆跟踪模块截取所述车辆的最新图像，对所述目标车辆区域进行更新；所述车牌跟踪模块截取所述车辆的最新车牌图像及所述车牌图像周边区域的图像，对所述目标车牌区域进行更新。
可选地，所述跟踪处理模块在所述车辆从所述入场检测区越过车位线进入所述车位区域中的车位的情况下，计算所述车辆在所述车位上停留的时间；所述跟踪处理模块在所述车牌从所述入场检测区越过车位线进入所述车位区域中的车位的情况下，计算所述车牌在所述车位上停留的时间；所述跟踪处理模块在所述车辆在所述车位上停留的时间超过第二阈值的情况下，确定所述车辆为入场车辆、所述车辆的停车事件为入场停车事件，以及确定入场停车事件信息，并将所述入场停车事件信息记录到在场停车信息表中；所述跟踪处理模块在所述车牌在所述车位上停留的时间超过第二阈值的情况下，确定所述车牌标识的车辆为入场车辆、所述车牌标识的车辆的停车事件为入场停车事件，以及确定入场停车事件信息，并将所述入场停车事件信息记录到在场停车信息表中；所述入场停车事件信息包括车牌号、入场时间、所述入场停车事件发生时车辆所在的车位、入场关键点图像、入场关键时序图像、入场跟踪视频。
可选地，在所述车辆在所述车位上停留的时间超过所述第二阈值的情况下，所述跟踪处理模块停止跟踪所述车辆，所述运动检测模块对所述车辆的运动检测继续进行；在所述车牌在所述车位上停留的时间超过所述第二阈值的情况下，所述跟踪处理模块停止跟踪所述车牌标识的车辆，所述运动检测模块对所述车牌标识的车辆的运动检测继续进行。
可选地，所述入场关键点图像包括：所述车辆在入场检测区域中首次被检测到的图像、所述车辆在入场检测区域中首次被检测到的车牌图像、所述车辆从入场检测区域越过车位线进入车位区域的图像、车辆停在车位上的图像。
可选地，所述入场关键时序图像为从所述入场跟踪视频中选择提取得到的图像。
可选地，所述入场跟踪视频为所述车辆在入场检测区域中首次被检测到至所述车辆停在车位上的视频段。
可选地，所述跟踪处理模块在所述车辆从所述车位区域中的车位进入所述出场检测区域的情况下，确定所述车辆为出场车辆，以及确定所述车辆的出场停车事件信息；所述跟踪处理模块在所述车牌从所述车位区域中的车位进入所述出场检测区域的情况下，确定所述车牌标识的车辆为出场车辆，以及确定所述车牌标识的车辆的出场停车事件信息；所述出场停车事件信息包括车牌号、出场时间、出场停车事件发生时车辆所在的车位、出场关键点图像、出场关键时序图像、出场跟踪视频。
可选地，所述出场关键点图像包括所述车辆离场前停在车位上的图像、所述车辆在出场检测区域中首次被检测到的图像、所述车辆在出场检测区域中首次被检测到的车牌图像、所述车辆从车位区域越过车位线进入出场检测区域的图像、所述车辆在出场检测区域中最后一次被检测到的图像、所述车辆在出场检测区域中最后一次被检测到的车牌图像。
可选地，所述出场关键时序图像为从所述出场跟踪视频中选择提取得到的过程图像。
可选地，所述出场跟踪视频包括第一出场跟踪视频、第二出场跟踪视频，所述第一出场跟踪视频为所述车辆从车位区域进入出场检测区域直至从出场检测区域消失的视频段，所述第二出场跟踪视频为所述车辆在出场检测区域中首次被检测到至所述车辆停在出场检测区域中消失的视频段。
可选地，在所述车牌跟踪模块在所述出场检测区域中跟踪到车牌，以及所述车牌识别模块识别所述车牌的车牌号的情况下，根据所述车牌号，所述跟踪处理模块查找在场车辆信息表中是否已记录所述车牌号，在所述车牌号之前已被记录的情况下，确定所述车牌标识的车辆为出场车辆，以及所确定述车牌标识的车辆的车牌号、出场时间、所述车牌标识的车辆所在的车位。
根据本公开的再一些实施例，提供一种停车管理系统，包括：布置在路侧停车场的杆位上的相机群，用于获取监控区域图像；上述任一个实施例中所述的停车管理装置，用于按图像划分规则对监控区域图像进行图像划分，以及对监控区域图像进行检测识别跟踪处理，确定所述相机群管理的车位上的停车事件信息。
可选地，所述相机群包括多个相机组，每个所述相机组包括至少一个相机阵列， 每个所述相机阵列包括至少一个摄像机，一个杆位上安装一个所述相机组，首末两个杆位上的相机组均包含一个相机阵列，其余杆位上的相机组均包含两个相机阵列。
根据本公开的再一些实施例，提供一种计算机可读存储介质，其上存储有计算机程序，该程序被处理器执行时实现所述任一个实施例中的停车管理方法。
在上述实施例中，由于摄像机，尤其是枪型摄像机，具有视角和焦距稳定的特点，能持续地采集稳定的图像，以目前200万像素的枪机为例，部署于6米监控杆，当视野覆盖2-3个车位时，能得到清晰、稳定的车牌图像。因此本发明提供的方法及装置基于对摄像机拍摄的图像进行划分，可同时对得到的各图像区域进行处理，进而实现对车辆进出场的跟踪和识别，以及当一个摄像机监控的若干个车位上发生多个并发停车事件时，能根据划分的区域同时管理多个停车事件，相较于球机的管理方式，无需考虑相机的调度响应时间，因此采用本发明实施例提供的方法，能提高停车事件的管理效率。
通过以下参照附图对本公开的示例性实施例的详细描述，本公开的其它特征及其优点将会变得清楚。
附图说明
此处所说明的附图用来提供对本公开的进一步理解，构成本申请的一部分，本公开的示意性实施例及其说明用于解释本公开，并不构成对本公开的不当限定。在附图中：
图1是本发明实施例提供的一种停车管理方法的示例性流程图；
图2是本发明实施例中阵列摄像机的安装场景图；
图3是本发明实施例提供的第一监控区域图像的划分示意图；
图4是图1中步骤S104的具体实施示例性流程图；
图5是本发明实施例对车辆进行检测识别跟踪的场景图；
图6是本发明实施例提供的一种停车管理装置的结构示意性框图；
图7是本发明实施例提供的另一种停车管理装置的结构示意性框图；
图8是本发明实施例提供的一种停车管理系统的结构示意性框图。
具体实施方式
现在将参照附图来详细描述本公开的各种示例性实施例。应注意到：除非另外具 体说明，否则在这些实施例中阐述的部件和步骤的相对布置、数字表达式和数值不限制本公开的范围。
同时，应当明白，为了便于描述，附图中所示出的各个部分的尺寸并不是按照实际的比例关系绘制的。
以下对至少一个示例性实施例的描述实际上仅仅是说明性的，决不作为对本公开及其应用或使用的任何限制。
对于相关领域普通技术人员已知的技术、方法和设备可能不作详细讨论，但在适当情况下，所述技术、方法和设备应当被视为授权说明书的一部分。
在这里示出和讨论的所有示例中，任何具体值应被解释为仅仅是示例性的，而不是作为限制。因此，示例性实施例的其它示例可以具有不同的值。
应注意到：相似的标号和字母在下面的附图中表示类似项，因此，一旦某一项在一个附图中被定义，则在随后的附图中不需要对其进行进一步讨论。
图1是本发明实施例提供的一种停车管理方法的示例性流程图。
如图1所示，所述方法具体包括步骤S102～S104。在一些实施例中，还可以包括步骤S101。各步骤可以按照如下的方式具体实施。
在步骤S101中，预设摄像机的图像划分规则。在一些实施例中，可以根据图2中的方式安装摄像机。
图2示出根据本公开的一些实施例的阵列摄像机的安装场景图。
如图2所示，可将若干个摄像机作为一个阵列，组成一个相机组安装固定在路侧停车场中的杆位上，对路侧停车事件进行监控管理。摄像机安装好后，设定各个摄像机的焦距、拍摄角度等参数，确定摄像机管理的路侧停车监控区域，以获取监控信息。
在一些实施例中，可以从摄像机拍摄的监控区域图像中获取一张监控区域图像作为第一监控区域图像，在第一监控区域图像中设定跟踪检测区域的坐标，根据坐标对摄像机后续拍摄的监控区域图像进行划分。例如，可在所述摄像机拍摄的监控区域图像中选取最新拍摄的一帧，作为第一监控区域图像，用于预设摄像机的图像划分规则。即通过在所述第一监控区域图像中选择跟踪检测区域的坐标，作为摄像机的图像划分规则并存储，以便后续步骤根据设定好的图像划分规则对摄像机后续拍摄的监控区域图像进行划分和识别处理。所述跟踪检测区域为车辆的活动区域，可包含车位区域、入场检测区域、出场检测区域。所述跟踪检测区域的坐标包括车位区域、入场检测区域、出场检测区域在所述第一监控区域图像中的坐标。例如，所述第一监控区域图像 可以如图2所示。
图3示出根据本公开的一些实施例的第一监控区域图像的示意图。
如图3所示，所述跟踪检测区域可以包含车位区域、入场检测区域、出场检测区域。在一些实施例中，所述车位区域、入场检测区域、出场检测区域可以按照下面的方法进行划分。
所述车位区域为停放车辆的区域，包含多个停车泊位，可如图3中界线2和界线6之间的区域。
所述入场检测区域包含车位区域边界线两侧的临近区域。所述入场检测区域可作为检测入场停车事件，以及获取所述入场停车事件信息的区域。例如，如图3所示，图中的界线2为车位区域的边界线，也称为车位线，所述边界线两侧的临近区域指的是界线1和界线2之间的区域、界线3和界线2之间的区域(亦界线1和界线3之间的区域)。界线1、界线3分别和界线2之间的距离可根据实际需求而设定。由于车辆在越过车位线时，车辆的速度比较慢，是抓拍识别车牌号的良好时机，因此将界线1和界线2之间的区域作为入场检测区域的一部分，可以有效地抓拍车牌的效率和准确率。
所述出场检测区域为车位区域之外的区域，例如，如图3中界线2和界线4、界线5之间的区域。所述出场检测区域可作为检测出场停车事件，以及获取所述出场停车事件信息的区域。
在实际应用中，可根据每个摄像机的具体监控场景，将图像中的不相关区域(例如草坪区域、树木区域等)屏蔽后，作为车辆跟踪监测区域，从而可以减少后续处理车辆跟踪监测区域图像时的计算量。
在步骤S102中，接收摄像机拍摄的监控区域图像。
在步骤S103中，根据所述摄像机的图像划分规则，对所述摄像机拍摄的监控区域图像进行划分，得到所述监控区域图像中的跟踪检测区域。
例如，所述路侧停车监控区域可以包括车位区域、入场检测区域、出场检测区域。
在步骤S104中，对所述监控区域图像中的车辆和车牌中的至少一种进行监测，根据所述车辆和车牌中的至少一种所在的跟踪检测区域确定停车事件信息。监测可以包括检测、识别、跟踪等处理方式。
在一些实施例中，可以对整张监控区域图像中的所有车辆、车牌进行检测和识别。也可以采用多线程方式或者多处理模块或者多处理器进行并行处理，例如，以跟踪检 测区域为处理单位，并行地对各区域进行车辆、车牌的检测和识别。并行处理方式可以提高车辆、车牌的检测识别效率。
例如，所述停车事件可以包括入场停车事件、出场停车事件、违章、违规、违法停车事件。所述停车事件信息可以包括停车事件的类型、停车事件发生的时间、车辆在车位上停放的时间、停车事件发生时车辆所在的车位、车牌号、关键点图像、跟踪视频、关键时序图像。停车事件信息被确定后，可以将所述停车事件信息存储，以便后续上传后台，供管理人员进行查询、巡检。
所述停车事件的类型可以包括入场停车事件、出场停车事件、违章、违规、违法停车事件。所述停车事件发生的时间可以包括车辆驶入车位的时间(亦称入场时间)、车辆驶出车位的时间(亦称出场时间)。车辆在车位上停放的时间可根据车辆的出场时间和入场时间计算。
所述关键点图像可以包括车辆的入场关键点图像、出场关键点图像。所述入场关键点图像可以包括所述车辆在入场检测区域中首次被检测到的车辆图像、所述车辆在入场检测区域中首次被检测到的车牌图像、所述车辆从入场检测区域越过车位线进入车位区域的过程图像、车辆停在车位上的图像。所述出场关键点图像可以包括所述车辆离场前停在车位上的图像、所述车辆在出场检测区域中首次被检测到的车辆图像、所述车辆在出场检测区域中首次被检测到的车牌图像、所述车辆从车位区域越过车位线进入出场检测区域的图像、所述车辆在出场检测区域中最后一次被检测到的车辆图像、所述车辆在出场检测区域中最后一次被检测到的车牌图像。在实际的停车管理中，当出现异常停车事件时(例如，后台服务器显示车位上有车辆，但信息中没有显示车牌)，工作人员可对所述关键点图像进行巡检，以补充停车事件的信息。
所述跟踪视频可以包括入场跟踪视频和出场跟踪视频。所述入场跟踪视频可以为所述车辆在入场检测区域中首次被检测到至所述车辆停在车位上的视频段。所述出场跟踪视频可以包括第一出场跟踪视频、第二出场跟踪视频。所述第一出场跟踪视频可以为所述车辆从车位区域进入出场检测区域直至从出场检测区域消失的视频段。所述第二出场跟踪视频可以为所述车辆在出场检测区域中首次被检测到至所述车辆在出场检测区域中消失的视频段。由于通常情况下，执行本方法的前端设备/装置的处理能力有限，因此可基于后台处理能力强大的特点，可将所述跟踪视频先进行存储，以备后续需要时。还可以将所述跟踪视频上传至后台，由后台对跟踪视频进行识别处理，能进一步提高识别车辆、车牌的准确率。
所述关键时序图像可以包括入场关键时序图像、出场关键时序图像。入场关键时序图像可从所述入场跟踪视频中选择提取得到。出场关键时序图像可从所述出场跟踪视频中选择提取得到。
可选的，可根据实际需求从入场跟踪视频中选择相应时刻的图片组合成所述入场关键时序图像，从出场跟踪视频段中选择相应时刻的图片组合成所述出场关键时序图像。由于传输跟踪视频需要较大的带宽、流量过大，因此对于带宽小的情况，可采用传输关键时序图像的方式，将关键时序图像传输至后台，由后台对关键时序图像进行识别处理，进一步提高识别车辆、车牌的准确率。
在一些实施例中，可以通过图4中的流程来实施步骤S104。
图4示出根据本公开的一些实施例的停车事件信息确定方法的示例性流程图。
如图4所示，步骤S104可以通过步骤S1041～步骤S1044来实施。
在步骤S1041中，对跟踪检测区域进行车辆检测，确定所述跟踪检测区域中的车辆。对所述跟踪检测区域进行车牌检测，确定所述跟踪检测区域中的车牌。以及对所述车牌进行识别，确定所述车牌的车牌号。
例如，可采用车辆检测算法对跟踪检测区域进行车辆检测。还可采用车牌检测算法对所述跟踪检测区域进行车牌检测以及可采用车牌识别算法对所述车牌进行识别。
可选的，所述车辆检测算法、车牌检测算法可以采用基于深度学习的Fast Rcnn算法(英文全称：Faster Regions with CNNs features)、SSD算法(英文全称：single shot multibox detector)、Yolo算法(英文全称：You Only Look Once)等目标检测算法，或者其他类型的图像目标检测算法。
在步骤S1042中，获取所述车牌与所述车辆的位置关系，以判断所述车牌是否为所述车辆的车牌。例如，可以在确定所述跟踪检测区域中的车辆、车牌后，根据所述车辆和所述车牌的位置关系，确定所述车牌是否为所述车辆的车牌。在一些实施例中，可以通过如下方法来进行判断。
(1)如果所述车牌的车牌区域位于所述车辆的车辆区域内，则确定既检测到所述车辆，又检测到所述车辆的车牌和/或识别到所述车牌的车牌号。例如，如图3中区域10411所示，车牌区域位于车辆区域内，因此表示既检测所述车辆，又检测和/或识别到所述车辆的车牌号。可以将所述车牌和车辆关联，以及将所述车辆的特征信息和所述车牌的特征信息合并，作为所述车辆的特征信息，且优先选择车牌区域的坐标作为该车辆的坐标，车辆区域可作为辅助的跟踪区域。
(2)如果所述车牌的车牌区域没有位于任何车辆的车辆区域内，则确定仅检测到所述车辆的车牌和/或识别到所述车牌的车牌号，但未检测到所述车牌标识的车辆。例如，如图3中区域10412所示，车牌区域没有位于任何车辆区域内，因此表示未检测到所述跟踪检测区域图像中的车辆，但检测和/或识别到所述车辆的车牌号。可以将所述车牌的特征信息作为所述车辆的特征信息。
(3)如果所述车辆的车辆区域没有包含任何车牌的车牌区域，则确定仅检测到所述车辆，但未检测到所述车辆的车牌和/或未识别到所述车牌的车牌号。例如，如图3中区域10413所示，车辆区域没有包含任何车牌区域，因此表示仅检测到所述跟踪检测区域图像中的车辆，但未能检测和识别到所述车辆的车牌号。
在步骤S1043中，对所述车辆、所述车牌进行运动检测，确定所述车辆、所述车牌的运动状态。
例如，所述车辆、所述车牌的运动状态包括静止和运动。
可选的，可采用运动检测算法对所述车辆、所述车牌进行运动检测。所述运动检测算法可以采用基于光流的Lucas-knades算法、Kalman滤波算法等。
在步骤S1044中，根据所述车辆的运动状态、所述车牌的运动状态，确定停车事件信息。
例如，如果所述车辆的运动状态为静止，则根据所述车辆所在的跟踪检测区域，确定所述车辆的停车事件信息。如果所述车牌的运动状态为静止，则根据所述车牌所在的跟踪检测区域，确定所述车牌标识的车辆的停车事件信息。如果所述车辆的运动状态为运动，则对所述车辆进行跟踪，确定所述车辆的停车事件信息。如果所述车牌的运动状态为运动，则对所述车牌进行跟踪，确定所述车牌标识的车辆的停车事件信息。
可选的，可采用车辆跟踪算法对所述跟踪检测区域图像中的车辆进行跟踪。可采用车牌跟踪算法对所述跟踪检测区域图像中的车牌进行跟踪。所述车牌跟踪算法、车辆跟踪算法可以采用基于相关滤波器的KCF算法(英文全称：High-speed tracking with kernelized correlation filters)，Staple(英文全称：Sum of Template And Pixel-wise LEarners)目标跟踪算法，或其他类型的图像目标跟踪算法。
图5示出根据本公开的一些实施例的对车辆进行检测识别跟踪的场景图。
如图5所示，停车事件信息可以有如下流程来确认。
(1)如果所述车辆的运动状态为静止，则根据所述车辆所在的跟踪检测区域， 确定所述车辆的停车事件信息。如果所述车牌的运动状态为静止，则根据所述车牌所在的跟踪检测区域，确定所述车牌标识的车辆的停车事件信息。
在一些实施例中，如果所述车辆位于入场检测区域(如图5中的车辆A)，则计算所述车辆滞留在入场停车区域的时间。或者如果所述车牌位于入场检测区域，则计算所述车牌滞留在入场停车区域的时间。如果所述车辆或者所述车牌滞留在入场检测区域超过第一阈值，则确定所述车辆或者所述车牌标识的车辆为违规停车车辆，所述停车事件的类型为违规停车事件。所述第一阈值可根据实际需求进行设定。例如，可以设定所述第一阈值为3-5分钟。
以上处理手段可处理步骤S1042所列的情况。例如，由于在步骤S1042的情况(1)中，车辆处于静止状态时，有时由于车牌或者车辆被人、其他物体遮挡，而导致车牌或车辆暂时无法被检测或识别，因此只要当检测到车牌或者车辆处于静止状态时，即可对步骤S1042的情况(1)中的车辆执行停车事件信息的确认过程。
在一些实施例中，如果所述车牌位于车位区域(如图5中的车辆B)，以及所述车牌的车牌号已被识别，则根据所述车牌号，查找在场车辆信息表。如果所述车牌号之前未被记录，则确定所述车牌标识的车辆为入场车辆，以及确定所述车牌标识的车辆的入场停车事件信息。所述入场停车事件信息可以包括车辆的车牌号、车辆的入场时间、车辆所在的车位。
如果所述车辆位于车位区域，以及所述车辆的车牌号未被识别，则根据所述车辆的特征信息(如颜色、车型等)，查找在场车辆信息表。如果所述车辆之前未被记录，则确定所述车辆为无牌车，并确定为入场车辆，以及确定所述车牌标识的车辆的入场停车事件信息。所述入场停车事件信息可以包括车辆的特征信息、车辆的入场时间、车辆所在的车位。
由于在车辆驶入车位的过程中，有时没能抓拍、识别到车辆，因而导致车辆停在车位上后，没能记录该车辆的停车事件信息。因此在一些实施例中可以通过实时或者定时对停在车位区域中的车辆进行运动检测和识别，通过将识别到的车牌号与在场停车信息表进行对比，补充记录该车辆的车型、颜色、入场时间、车牌号等入场停车事件信息。
在一些实施例中，如果所述车辆位于出场检测区域(如图5中的车辆C)，则计算所述车辆滞留在出场停车区域的时间。或者如果所述车牌位于出场检测区域，则计算所述车牌车辆滞留在出场停车区域的时间。如果所述车辆或者所述车牌滞留在出场 检测区域的时间超过第一阈值，则确定所述车辆或者所述车牌标识的车辆为违规停车车辆，所述停车事件的类型为违规停车事件。违规停车事件的判定，可以辅助交通管理。以上处理手段亦可以处理步骤S1042所列的情况。
(2)如果所述车辆的运动状态为运动，则对所述车辆进行跟踪，确定所述车辆的停车事件信息。如果所述车牌的运动状态为运动，则对所述车牌进行跟踪，确定所述车牌标识的车辆的停车事件信息。
可以基于步骤S1042所确定的情况，对车辆采取相应的跟踪方案。例如可以通过如下方法实现。
1)如果检测到所述跟踪检测区域图像中的车辆，以及检测到所述车辆的车牌和/或识别到所述车牌上的车牌号，则采用车辆跟踪算法，截取所述车辆区域作为初始的目标车辆区域进行跟踪，以及采用车牌跟踪算法，截取所述车牌区域及所述车牌周边的区域作为初始的目标车牌区域进行跟踪。
在一些实施例中，如果车辆(如图5中的车辆D)从入场检测区越过车位线(即如图2中的界线2)进入车位区域中的车位(即所述目标车辆区域和目标车牌区域移入车位区域中的车位)，则计算所述车辆在所述车位上停留的时间。如果车辆在车位上停留的时间超过第二阈值，则停止跟踪，对所述车辆的运动检测继续进行(由于跟踪算法消耗的运算资源较大，因此当确定车辆停在车位上后，可停止跟踪，以节省运算资源)，并记录此次停车为入场停车事件(即所述停车事件的类型为入场停车事件)，以及确定并存储所述入场停车事件的信息。即将车辆的车牌号、入场时间、入场停车事件发生时车辆所在的车位等写入所述在场车辆信息表。以及将车辆的入场关键点图像、车辆的入场关键时序图像、车辆的入场跟踪视频与所述在场车辆信息表进行关联并存储。所述第二阈值可根据实际需求进行设定。例如，可以设定所述第二阈值为20-30秒。如果车辆没有进入车位区域，而驶出入场检测区域，则停止对车辆的跟踪。
在一些实施例中，如果车辆(如图5中的车辆E)从所述车位区域中的车位进入所述出场检测区域(即所述车辆区域和车牌区域从所述车位区域中的车位进入所述出场检测区域)，则确定所述车辆为出场车辆，确定所述出场停车事件信息。出场停车事件信息包括车牌号、车辆的出场时间、所述出场停车事件发生时车辆所在的车位、出场关键点图像、出场关键时序图像、出场跟踪视频。
在一些实施例中，如果在出场检测区域中跟踪到车辆(如图5中的车辆F)的车牌，以及所述车牌的车牌号已被识别，则根据所述车牌号，查找在场车辆信息表。如 果所述车牌号之前已被记录，即所述车牌标识的车辆之前是入场车辆，则确定所述车牌标识的车辆为出场车辆和该车辆的出场停车事件信息。出场停车事件信息包括车牌号、出场时间、所述车牌标识的车辆所在的车位。通过对出场检测区域中的车辆进行跟踪识别，以及将识别到的车牌号与在场停车信息表进行对比，补充记录该车辆的出场时间、车牌号等出场停车事件信息，可解决由于在车辆驶出车位的过程中，没能抓拍、识别到车辆，而导致的车辆出场停车事件信息缺失的问题。
2)如果检测到所述跟踪检测区域中的车辆，但未检测到所述车辆的车牌，则采用车辆跟踪算法，截取所述车辆区域作为初始的目标车辆区域进行跟踪。
在一些实施例中，如果车辆(如图5中的车辆D)从入场检测区域越过车位线(即如图2中的界线2)进入车位区域中的车位(即目标车辆区域从入场检测区域移入车位区域中的车位)，以及在所述车位上停留的时间超过第二阈值，则停止跟踪所述车辆，继续对所述车辆进行运动检测。以及记录所述车辆此次停车为入场停车事件，确定并存储所述停车事件的信息。即将所述车辆的入场时间、入场停车事件发生时车辆所在的车位写入所述在场车辆信息表。以及将所述车辆的入场关键点图像、入场关键时序图像、入场跟踪视频与所述在场车辆信息表进行关联并存储。如果所述车辆没有进入车位区域，而驶出入场检测区域，则停止对所述车辆的跟踪。
在一些实施例中，如果车辆(如图5中的车辆E)从所述车位区域中的车位进入所述出场检测区域(即所述目标车辆区域从所述车位区域中的车位进入所述出场检测区域)，则确定所述车辆为出场车辆，确定所述出场停车事件信息。所述出场停车事件信息包括所述出场时间、所述出场停车事件发生时车辆所在的车位、出场关键点图像、出场关键时序图像、出场跟踪视频。
3)如果未检测到所述跟踪检测区域图像中的车辆，但检测到所述车辆的车牌和/或识别到所述车牌的车牌号，则采用车牌跟踪算法，截取所述车牌图像及所述车牌图像周边的区域作为初始的目标车牌区域进行跟踪。
在一些实施例中，如果车牌(如图5中的车辆D的车牌)从入场检测区越过车位线(即如图2中的界线2)进入车位区域中的车位(即所述目标车牌区域移入车位区域中的车位)，以及在所述车位上停留的时间超过第二阈值，则停止跟踪所述车辆，对所述车辆的运动检测继续进行。以及记录所述车牌标识的车辆为入场车辆，此次停车为入场停车事件(即所述停车事件的类型为入场停车事件)。确定并存储所述停车事件的信息，即将车牌号、入场时间、入场停车事件发生时车辆所在的车位写入所述 在场车辆信息表。以及将入场关键点图像、入场关键时序图像、入场跟踪视频与所述在场车辆信息表进行关联并存储。
在一些实施例中，如果车牌(如图5中的车辆E的车牌)从所述车位区域中的车位进入所述出场检测区域(即所述目标车牌区域从所述车位区域中的车位进入所述出场检测区域)，则确定所述车牌标识的车辆为出场车辆，确定出场停车事件信息。所述出场停车事件信息包括车牌号、出场时间、所述出场停车事件发生时车辆所在的车位、出场关键点图像、出场关键时序图像、出场跟踪视频。
在一些实施例中，如果在出场检测区域中跟踪到所述车牌(如图5中的车辆F的车牌)，以及所述车牌的车牌号已被识别，则根据所述车牌号，查找在场车辆信息表。如果所述车牌号之前已被记录，即所述车牌标识的车辆之前是入场车辆，则确定所述车牌标识的车辆为出场车辆和所述车牌标识的车辆的出场停车事件信息。所述出场停车事件信息包括车牌号、出场时间、所述出场停车事件发生时车辆所在的车位。通过对出场检测区域中的车辆的车牌进行跟踪识别，以及将识别到的车牌号与在场停车信息表进行对比，补充记录该车辆的出场时间、车牌号等出场停车事件信息。这样可解决由于在车辆驶出车位的过程中，没能抓拍、识别到车辆，而导致的车辆出场停车事件信息缺失的问题。
在一些实施例中，在对目标车辆区域跟踪的过程中，定时(也可实时)截取所述车辆最新的图像，对所述目标车辆区域进行更新，防止长时间跟踪所述目标车辆区域而造成的目标丢失。在对所述目标车牌区域跟踪的过程中，定时(也可实时)截取所述车辆最新的车牌图像及车牌图像周边区域的图像，对所述目标车牌区域进行更新，防止长时间跟踪所述目标车牌区域而造成的目标丢失。此外，目标车辆区域除例如可以包括所述车牌图像的区域外，还可以包括车牌图像周边的区域，能提高检测车牌的准确率。
可选的，如果车辆进入一个车位，停留时间超过第一阈值后，所述车辆又选择同一车位区域中的其他车位进行停泊，则所述车辆的入场时间以进入第一个车位的时间为准，所述车辆最终的停车车位以最后停泊的车位为准。
上述实施例中，由于摄像机，尤其是枪型摄像机，具有视角和焦距稳定的特点，能高速持续地采集稳定的图像。以目前200万像素的枪机为例，部署于6米监控杆，当视野覆盖2-3个车位时，能得到清晰、稳定的车辆图像、车牌图像，识别准确率高，且能获得停车事件全程完整的证据链。因此本公开基于对摄像机拍摄的图像进行划 分，可同时对得到的各图像区域进行处理，可实现对车辆进出场的跟踪和识别。以及当某一摄像机监控的若干个车位上发生多个并发停车事件时，能根据划分的区域同时管理多个停车事件。相较于球机的管理方式，无需考虑相机的调度响应时间。此外，本公开可通过对相机各拍摄的图像进行联合处理，进一步提高置信度、准确率。综上，采用本公开的技术方案，能提高路侧停车事件的管理效率。
图6是本发明实施例提供的一种停车管理装置的结构示意性框图。
所述装置用于执行本公开上述任一个实施例中提供的方法。如图6所示，所述装置包括处理器201和存储器203。在一些实施例中，还可以包括显示器202。在另一些实施例中，还可以包括看门狗(WD，Watch Dog)204。例如，所述装置的各组成部分的可以按照如下方式配置。
处理器201用于接收摄像机拍摄的监控区域图像。以及根据所述摄像机的图像划分规则，对所述摄像机拍摄的监控区域图像进行划分，得到所述监控区域图像中的跟踪检测区域。以及对所述监控区域图像中的车辆和车牌中的至少一种进行监测，根据所述车辆和车牌中的至少一种所在的跟踪检测区域确定停车事件信息。所述跟踪检测区域包括车位区域、入场检测区域、出场检测区域。所述停车事件信息包括车辆的车牌号、停车事件的类型、停车事件发生时车辆所在的车位。
显示器202用于显示操作管理界面，通过操作管理界面预设所述摄像机的图像划分规则。例如，所述操作管理界面接收用户输入的第一查询指令，所述第一查询指令为用户选择查询所述摄像机的监控信息的操作。所述操作管理界面根据所述第一查询指令，获取所述摄像机拍摄的第一监控区域图像。以及将所述第一监控区域图像显示。所述第一监控区域图像为所述摄像机拍摄最新拍摄的监控区域图像。所述操作管理界接收用户在第一监控区域图像中选择的跟踪检测区域坐标，所述跟踪检测区域的坐标为所述摄像机的图像划分规则。
存储器203用于存储每个摄像机的监控信息。所述监控信息包括：摄像机拍摄的监控区域图像、经划分后得到的图像(即路侧停车监控区域、车位区域、入场检测区域、出场检测区域等跟踪检测区域图像)、图像划分规则、所管理的车位信息以及所述车位上发生的停车事件信息。所述停车事件信息包括停车事件的类型、停车事件发生的时间、停车事件发生时车辆所在的车位、车辆的停车时间、车辆的车牌号、关键点图像、关键时序图像、跟踪视频等。
看门狗204用于监控处理器201的工作状态，以及将处理器201的工作状态信息 上传到后台服务器。当处理器出现异常时，看门狗204控制处理器201停止工作或者由管理人员通过后台服务器控制处理器201停止工作。
可选的，本实施例中所指的摄像机可以是枪型摄像机。
在一些实施例中，处理器201可以定时将存储器203中每个停车事件的关键点图像、关键时序图像、跟踪视频发送给服务器，用户可通过在操作管理界面上查找每个停车事件的关键点图像、关键时序图像，以对停车事件进行巡检。服务器可对关键时序图像、跟踪视频进行识别处理，修正、补充本装置处理器201所确定的停车事件信息，如车牌号等，进一步提高本装置对停车事件的车牌识别率和管理效率。
在一些实施例中，所述装置可以根据如下流程工作。
处理器201将摄像机拍摄的监控区域图像、关键点图像、跟踪视频、关键时序图像经输入模块301存储至存储器203中。
用户可通过显示器202上的操作管理界设定摄像机的图像划分规则和查询摄像机的监控信息。当显示器202上的操作管理界面接收到用户输入的第一查询指令时，即用户选择要查询某个摄像机的监控信息时，操作管理界面将该摄像机最新拍摄的监控区域图像进行显示(第一监控区域图像)，用户可在所述第一监控区域图像上选择(可以是点击的方式)车辆跟踪监测区域的坐标(车位区域、入场检测区域、出场检测区域的坐标)作为所述摄像机的图像划分规则存储至存储模块305。
处理器201后续接收到摄像机拍摄的监控区域图像后，根据各个摄像机的图像划分规则，对各个摄像机拍摄的监控区域图像进行划分，得到所述监控区域图像中的跟踪检测区域，即车位区域、入场检测区域、出场检测区域。
处理器201对车辆跟踪监测区域图像进行目标检测、运动检测、车牌识别、目标车辆跟踪、目标车牌跟踪，进而得到车位上的停车事件信息。
当用户通过操作管理界面进行查询、巡检等操作，以补充停车事件信息时，操作管理界面根据用户的操作，可将监控区域图像、关键点图像、跟踪视频、关键时序图像等信息进行显示，以供用户巡检，补充停车事件信息。
例如，处理器201对图像中车辆的具体检测、识别、跟踪过程，可参考本公开提供的方法实施例中步骤S104的具体实施过程，此处不再赘述。
上述实施例中，基于本公开提供的方法实施例，采用本装置执行上述方法，即通过在软件(操作管理界面)上操作的方式设定摄像机的图像划分规则。然后由处理器按图像划分规则对摄像机拍摄的图像进行划分，同时对得到的各图像区域进行处理。 并对车辆进出场的检测、识别、跟踪。以及当摄像机监控的若干个车位上发生多个并发停车事件时，能根据划分的区域同时管理多个停车事件。相较于球机的管理方式，本装置不存在摄像机调度响应时间(类似球机的响应时间)的影响，因此采用本公开提供的方法，能提高路侧停车事件的管理效率。
图7是本发明实施例提供的另一种停车管理装置的结构示意性框图。
所述装置为可用于执行本公开的上述任一个实施例中方法。如图7所示，所述装置包括：输入模块301、图像划分模块302和检测识别跟踪模块303。在一些实施例中，还可以包括：操作管理界面304。在另一些实施例中，还可以包括：和存储模块305。例如，所述装置中各模块可以按照如下流程工作。
输入模块301用于接收摄像机拍摄的监控区域图像。
图像划分模块302用于根据所述摄像机的图像划分规则，对所述摄像机拍摄的监控区域图像进行划分，得到所述监控区域图像中的跟踪检测区域。所述跟踪检测区域包括车位区域、入场检测区域、出场检测区域。
检测识别跟踪模块303用于对所述监控区域图像中的车辆和车牌中的至少一种进行监测，根据所述车辆和车牌中的至少一种所在的跟踪检测区域确定停车事件信息。
在一些实施例中，检测识别跟踪模块303可以包括车辆检测模块3031、车牌检测模块3032、车牌识别模块3033、运动检测模块3034、跟踪处理模块3035、运算判断模块3036。
车牌检测模块3032用于采用车牌检测算法对跟踪检测区域进行车牌检测，确定所述跟踪检测区域中的车牌。车牌识别模块3033用于采用车牌识别算法对所述车牌进行识别，确定所述车牌的车牌号。运算判断模块3036用于判断所述车牌是否为所述车辆的车牌。运动检测模块3034用于采用运动检测算法对所述车辆、所述车牌进行运动检测，确定所述车辆、所述车牌的运动状态；所述目标车辆、所述目标车牌的运动状态包括静止和运动。
跟踪处理模块3035用于根据所述车辆、车牌的运动状态，确定停车事件信息。例如，如果所述车辆或者车牌的运动状态为静止，则根据所述车辆或者车牌所在的跟踪检测区域，确定停车事件信息。如果所述车辆或者车牌的运动状态为运动，则对所述车辆进行跟踪和/或对所述车牌进行跟踪，确定停车事件信息。
在一些实施例中，跟踪处理模块3035可以包括车辆跟踪模块30351、车牌跟踪模块30352。车辆跟踪模块30351用于采用车辆跟踪算法对所述车辆进行跟踪，确定停 车事件信息。车牌跟踪模块30352用于采用车牌跟踪算法对所述车牌进行跟踪，确定停车事件信息。
操作管理界面304用于显示摄像机拍摄的监控区域图像、跟踪检测区域、停车事件信息。以及用于接收用户选择跟踪检测区域时输入的车位区域、入场检测区域、出场检测区域坐标作为所述摄像机的图像划分规则，进而对所述摄像机的图像划分规则进行预设。例如，可以根据如下方法进行预设。
操作管理界面304接收用户输入的第一查询指令。所述第一查询指令为用户选择查询所述摄像机的监控信息的操作。
操作管理界面304根据所述第一查询指令，获取所述摄像机拍摄的第一监控区域图像，以及将所述第一监控区域图像显示。所述第一监控区域图像为所述摄像机拍摄最新拍摄的监控区域图像。
操作管理界面304接收用户在第一监控区域图像中选择的跟踪检测区域坐标。所述跟踪检测区域的坐标为所述摄像机的图像划分规则。
存储模块305用于存储每个摄像机的监控信息。所述监控信息包括：摄像机拍摄的监控区域图像、经划分后得到的图像(即路侧停车监控区域、车位区域、入场检测区域、出场检测区域等跟踪检测区域图像。)、图像划分规则、所管理的车位信息以及所述车位上发生的停车事件信息。所述停车事件信息包括停车事件的类型、停车事件发生的时间、停车事件发生时所在的车位、车辆的停车时间、车辆的车牌号、关键点图像、关键时序图像、跟踪视频。
在一些实施例中，所述装置的工作过程可以如下方流程。
摄像机拍摄的监控区域图像、关键点图像、跟踪视频、关键时序图像经输入模块301存储至存储模块305。
用户可通过操作管理界面304选择设定摄像机的图像划分规则和查询摄像机的监控信息。当操作管理界面304接收到用户输入的第一查询指令时，即用户选择要查询某个摄像机的监控信息时，操作管理界面304将该摄像机最新拍摄的监控区域图像进行显示(第一监控区域图像)。用户可在所述第一监控区域图像上选择(可以是点击的方式)车辆跟踪监测区域的坐标(车位区域、入场检测区域、出场检测区域的坐标)作为所述摄像机的图像划分规则存储至存储模块305。
当输入模块301后续接收摄像机拍摄的监控区域图像时，图像划分模块302根据各摄像机的图像分规则将其拍摄的监控区域图像进行划分，得到包含车位区域、入场 检测区域、出场检测区域的车辆跟踪监测区域图像，并写入存储模块303中。以及将车辆跟踪监测区域图像发送给检测识别跟踪模块303进行处理。
检测识别跟踪模块303接收到所述跟踪检测区域后，启动车辆检测模块3031采用车辆检测算法对跟踪检测区域进行车辆检测，确定所述跟踪检测区域中的车辆。以及启动车牌检测模块3032采用车牌检测算法对所述跟踪检测区域进行车牌检测，确定所述跟踪检测区域中的车牌。以及启动车牌识别模块3033采用车牌识别算法对所述车牌进行识别，确定所述车牌的车牌号。
检测到跟踪检测区域中的车辆、车牌后，分别启动运算判断模块3036和运动检测模块3034。运算判断模块3036根据所述车辆和所述车牌的位置关系，判断所述车牌是否为所述车辆的车牌。启动运动检测模块3034采用运动检测算法对所述车辆、所述车牌进行运动检测，确定所述车辆、所述车牌的运动状态。
如果所述车辆或者车牌的运动状态为静止，则跟踪处理模块3035根据所述车辆或者车牌所在的跟踪检测区域，确定停车事件信息。例如，如果所述车辆位于入场检测区域或者出场检测区域，则跟踪处理模块3035计算所述车辆滞留在入场停车区域的时间。或者如果所述车牌位于入场检测区域或者出场检测区域，则计算所述车牌标识的车辆滞留在入场停车区域的时间。如果所述车辆或者所述车牌标识的车辆滞留在入场检测区域超过第一阈值，则确定所述车辆或者所述车牌标识的车辆为违规停车车辆。如果所述车牌位于车位区域，以及所述车牌已被车牌识别模块3033识别，则跟踪处理模块3035根据所述车牌号，查找在场车辆信息表。如果所述车牌号之前未被记录，则确定所述标识的车辆为入场车辆，确定所述车辆的入场停车事件信息。
如果所述车辆和/或者车牌的运动状态为运动，则所述跟踪处理模块3035对所述车辆进行跟踪和/或对所述车牌进行跟踪。例如，如果车辆检测模块3031检测到所述跟踪检测区域中的车辆，以及车牌检测模块3032检测到所述车辆的车牌和/或车牌识别模块3033识别到所述车牌的车牌号，则跟踪处理模块3035启动车辆跟踪模块30351所述车辆进行跟踪。以及启动车牌跟踪模块30352对所述车牌进行跟踪。
如果车辆检测模块3031检测到所述跟踪检测区域中的车辆，但车牌检测模块3032未检测到所述车辆的车牌，则跟踪处理模块3035启动车辆跟踪模块3035对所述车辆进行跟踪。
如果车辆检测模块3031未检测到所述跟踪检测区域图像中的车辆，但车牌检测模块3032检测到所述车辆的车牌和/或所述车牌识别模块3033识别到所述车牌的车牌 号，则跟踪处理模块3035启动车牌跟踪模块30352对所述车牌进行跟踪。
如果跟踪到车辆从入场检测区越过车位线进入车位区域中的车位，则跟踪处理模块3035计算所述车辆在所述车位上停留的时间。如果所述车辆在所述车位上停留的时间超过第二阈值，则确定此次停车为入场停车事件，以及确定入场停车事件信息。如果所述车辆在车位上停留的时间超过第二阈值，则跟踪处理模块3035停止跟踪所述车辆，运动检测模块3034继续对所述车辆进行运动检测。如果所述车牌标识的车辆在所述车位上停留的时间超过第二阈值，则跟踪处理模块3035停止跟踪所述车牌，运动检测模块3034继续对所述车牌进行运动检测。
如果跟踪到车辆从所述车位区域中的车位进入所述出场检测区域，则确定所述车辆为出场车辆，以及确定所述出场停车事件信息。
如果车牌跟踪模块30352在出场检测区域中跟踪到车辆的车牌，以及车牌识别模块3033识别所述车牌的车牌号，则跟踪处理模块3035查找在场车辆信息表。如果所述车牌号之前已被记录，则确定所述车牌标识的车辆为出场车辆，以及所确定述车牌标识的车辆的出场停车事件信息。
当用户通过操作管理界面304进行查询、巡检等操作，以补充停车事件信息时，操作管理界面304根据用户的操作，可将关键点图像、关键时序图像、跟踪视频等停车事件信息进行显示，以供用户巡检，补充停车事件信息。
在一些实施例中，所述装置对上述任一个实施例中提供方法的具体执行过程，可参照方法实施例中的具体过程，此处不再赘述。
上述实施例中，基于本公开提供的方法实施例，采用本装置执行上述方法。即通过在软件(操作管理界面)上操作的方式设定摄像机的图像划分规则，然后由处理器按图像划分规则对摄像机拍摄的图像进行划分，同时对得到的各图像区域进行处理，并对车辆进出场的跟踪和识别。以及当一个摄像机监控的若干个车位上发生多个并发停车事件时，能根据划分的区域同时管理多个停车事件。相较于球机的管理方式，无需考虑相机的调度响应时间，因此采用本公开实施例提供的方法，能提高路侧停车事件的管理效率。
图8是本发明实施例提供的一种停车管理系统的结构示意性框图。
如图8所示，所述系统包括相机群401、图6或者图7所示的停车管理装置402。在一些实施例中，所述系统还可以包括交换机403、云端服务器404。相机群401、停车管理装置402、云端服务器404均可以通过交换机通信。
相机群401包括N个相机组，N的值不小于2，具体取值可根据实际情况而定。每个相机组至少包括一个相机阵列，如相机组1包括相机阵列11、相机组2包括相机阵列21和相机阵列22、……、相机组N包括相机阵列N1和相机阵列N2、相机组(N+1)包括相机阵列(N+1)1。每个相机阵列包括至少一个摄像机，具体数量可根据实际情况而定。所述摄像机优选枪型摄像机。
将各个相机组安装在路侧停车场的杆位上，一个杆位上安装一个相机组。首末两个杆位上的相机组均为包含一个相机阵列的相机组，其余杆位上的相机组均为包含两个相机阵列的相机组。这样可以使得每两个杆位间的车位都可通过两个杆位上的两个相机阵列从两侧进行监控和管理。每个相机阵列管理一定数量的车位。在一些实施例中，相机阵列中的每个摄像机可设定为管理2至5个车位。相机组安装在杆位上后，将各个摄像机对准所管理的车位，设定每个摄像机的焦距、拍摄角度等参数，确定每个摄像机管理的路侧停车监控区域和车位，以获取监控信息。
在一些实施例中，所述系统的各组成部分可以按照如下方式配置。
相机群401用于获取所管理的车位上的监控区域图像。
停车管理装置402用于按图像划分规则对监控区域图像进行图像划分。以及对监控区域图像进行检测识别跟踪处理，确定相机群401管理的车位上的停车事件信息。
交换机403用于连接相机群401、基于多摄像机的停车管理装置402、云端服务器404，进行通信。
云端服务器404用于管理停车事件信息。例如可包括将停车管理装置402确定的停车事件信息存储，以及对关键时序图像、跟踪视频等数据进行批量地检测识别跟踪处理，补充停车事件所缺失的信息。
在一些实施例中，所述系统的工作过程可以按照如下流程进行。
相机群401获取个车位上的监控区域图像后，经交换机403传输给所述停车管理装置402。所述停车管理装置402按图像划分规则对监控区域图像进行图像划分。以及对监控区域图像进行检测识别跟踪处理，确定相机群401管理的车位上的停车事件信息。所述停车管理装置402将所确定的停车事件信息经交换机上传至云端服务器404。云端服务器404将所述停车事件信息进行存储，以及对关键时序图像、跟踪视频等数据进行批量地检测识别跟踪处理，补充停车事件所缺失的信息。
需要强调的是，相机群401在获取监控区域图像的过程中，当某一摄像机无法抓拍(例如错过抓拍时机造成的漏拍等)车辆进出车位的监控区域图像时，如果所述车 辆进入另一摄像机(包括本相机组中的其他摄像机和其他相机组中的摄像机)的管理区域，被所述另一摄像机抓拍到时(例如所述车辆出现在所述另一摄像机抓拍的侧停车监控区域图像中)，则所述另一摄像机将该侧停车监控区域图像上传给所述路车管理装置402后，所述停车管理装置402即可通过对该侧停车监控区域图像的图像划分和检测识别处理，补充该车辆在原车位上的停车事件信息。
例如，如图8所示，原来停在车位3j上车辆在驶出车位时，相机阵列32中管理该车位的摄像机未能抓拍到该车辆驶出车位时的图像。当该车辆出现在相机阵列22中某一摄像机管理区域中的出厂检测区域时，所述相机阵列22中的该摄像机将抓拍到的图像经交换机403发送给停车管理装置402。停车管理装置402通过对该监控区域图像的图像划分和检测识别处理，即可补充该车辆驶出车位3j时的出场停车事件信息。
上述实施例中，所述系统通过采用由相机阵列组成的相机组构成向集群，通过对相机阵列的安装，使两个杆位之间的车位由首尾两个相机阵列实时监控抓拍，提高抓拍效率。以及通过所述基于多摄像机的停车管理装置402对各个杆位间的摄像机抓拍的监控区域图像进行联合检测识别处理，可提高确定路车停车信息的效率和完成整性，从而整体提高路侧停车的管理效率。
在一些实施例中，提供一种计算机可读存储介质，其上存储有计算机程序，该程序被处理器执行时实现上述任一个实施例中的基于多摄像机的路侧停车管理方法。例如，该计算机可读存储介质为非瞬时性计算机可读存储介质。
至此，已经详细描述了根据本公开的基于多摄像机的路侧停车管理方法、基于多摄像机的路侧停车管理装置、基于多摄像机的路侧停车管理系统和计算机可读存储介质。为了避免遮蔽本公开的构思，没有描述本领域所公知的一些细节。本领域技术人员根据上面的描述，完全可以明白如何实施这里公开的技术方案。
可能以许多方式来实现本公开的方法和系统。例如，可通过软件、硬件、固件或者软件、硬件、固件的任何组合来实现本公开的方法和系统。用于所述方法的步骤的上述顺序仅是为了进行说明，本公开的方法的步骤不限于以上具体描述的顺序，除非以其它方式特别说明。此外，在一些实施例中，还可将本公开实施为记录在记录介质中的程序，这些程序包括用于实现根据本公开的方法的机器可读指令。因而，本公开还覆盖存储用于执行根据本公开的方法的程序的记录介质。
虽然已经通过示例对本公开的一些特定实施例进行了详细说明，但是本领域的技 术人员应该理解，以上示例仅是为了进行说明，而不是为了限制本公开的范围。本领域的技术人员应该理解，可在不脱离本公开的范围和精神的情况下，对以上实施例进行修改。本公开的范围由所附权利要求来限定。

Claims (65)

  1. 一种停车管理方法，所述方法包括：
    接收摄像机拍摄的监控区域图像；
    对所述监控区域图像进行划分，得到所述监控区域图像中的跟踪检测区域；
    对所述跟踪检测区域进行车辆、车牌监测，确定停车事件信息。
  2. 根据权利要求1所述的停车管理方法，其中，
    所述跟踪检测区域包括车位区域、入场检测区域、出场检测区域；
    所述停车事件信息包括车辆的车牌号、停车事件的类型、车辆所在的车位。
  3. 根据权利要求2所述的停车管理方法，其中，
    所述入场检测区域包括与所述车位区域的边界线临近的区域；
    所述出场检测区域包括所述车位区域之外的区域。
  4. 根据权利要求1-3任一项所述的停车管理方法，其中，对所述监控区域图像进行划分包括：
    从所述摄像机拍摄的监控区域图像中获取一张监控区域图像作为第一监控区域图像；
    在所述第一监控区域图像中设定跟踪检测区域的坐标；
    根据所述坐标对所述摄像机后续拍摄的监控区域图像进行划分。
  5. 根据权利要求1-4任一项所述的停车管理方法，其中，所述对所述跟踪检测区域进行车辆、车牌监测，确定停车事件信息，包括：
    对所述跟踪检测区域进行车辆检测，确定各跟踪检测区域中的车辆；
    对所述跟踪检测区域进行车牌检测，确定各跟踪检测区域中的车牌；
    对所述车牌进行识别，确定所述车牌的车牌号；
    对所述车辆、所述车牌进行运动检测，确定所述车辆、所述车牌的运动状态，所述运动状态包括静止和运动；
    根据所述车辆的运动状态，确定所述车辆的停车事件信息，以及根据所述车牌的运动状态，确定所述车牌标识的车辆的停车事件信息。
  6. 根据权利要求5所述的停车管理方法，还包括：
    获取所述车牌与所述车辆的位置关系；
    在所述车牌的车牌区域位于所述车辆的车辆区域内的情况下，确定所述车牌检测 和所述车辆检测的结果为既检测到所述车辆，又检测到所述车辆的车牌和/或识别到所述车牌的车牌号；
    在所述车牌的车牌区域没有位于任何车辆的车辆区域内的情况下，确定所述车牌检测和所述车辆检测的结果为仅检测到所述车牌和/或识别到所述车牌的车牌号，但未检测到所述车牌标识的车辆；
    在所述车辆的车辆区域没有包含任何车牌的车牌区域的情况下，确定所述车牌检测和所述车辆检测的结果为仅检测到所述车辆，但未检测到所述车辆的车牌。
  7. 根据权利要求6所述的停车管理方法，其中，根据所述车辆的运动状态，确定所述车辆的停车事件信息，以及根据所述车牌的运动状态，确定所述车牌标识的车辆的停车事件信息，包括：
    在所述车辆的运动状态为静止的情况下，根据所述车辆所在的跟踪检测区域，确定所述车辆的停车事件信息；
    在所述车牌的运动状态为静止的情况下，根据所述车牌所在的跟踪检测区域，确定所述车牌标识的车辆的停车事件信息；
    在所述车辆的运动状态为运动的情况下，对所述车辆进行跟踪，确定所述车辆的停车事件信息；
    在所述车牌的运动状态为运动的情况下，对所述车牌进行跟踪，确定所述车牌标识的车辆的停车事件信息。
  8. 根据权利要求7所述的停车管理方法，其中，所述根据所述车牌所在的跟踪检测区域，确定所述车牌标识的车辆的停车事件信息，包括：
    在所述车牌位于车位区域，以及识别到所述车牌的车牌号的情况下，根据所述车牌号，查找在场车辆信息表中是否记录有所述车牌号；
    在所述在场车辆信息表没有记录所述车牌号的情况下，确定所述车牌标识的车辆为入场车辆，以及确定并记录所述车牌标识的车辆的入场停车事件信息，所述车牌标识的车辆的入场停车事件信息包括车辆的车牌号、车辆的入场时间、车辆所在的车位；
    在所述车牌位于入场检测区域或者出场检测区域的情况下，计算所述车牌滞留在入场停车区域或者出场停车区域的时间；
    在所述时间超过第一阈值的情况下，确定所述车牌标识的车辆为违规车辆，所述车牌标识的车辆的停车事件为违规停车事件。
  9. 根据权利要求7所述的停车管理方法，其中，根据所述车辆所在的跟踪检测 区域，确定所述车辆的停车事件信息，包括：
    在所述车辆位于车位区域，以及未识别到所述车辆的车牌号的情况下，根据所述车辆，查找所述在场车辆信息表是否记录有所述车辆；
    在所述在场车辆信息表没有记录所述车辆的情况下，确定所述车辆为入场车辆，以及确定并记录所述车辆的入场停车事件信息，所述车辆的入场停车事件信息包括车辆的入场时间、车辆所在的车位；
    在所述车辆位于入场检测区域或者出场检测区域的情况下，计算所述车辆滞留在入场停车区域或者出场停车区域的时间；
    在所述时间超过所述第一阈值的情况下，确定所述车辆为违规车辆，所述车辆的停车事件为违规停车事件。
  10. 根据权利要求6-9任一项所述的停车管理方法，其中，对所述车辆或所述车牌进行跟踪，包括：
    在检测到所述跟踪检测区域图像中的车辆，以及检测到所述车辆的车牌或识别到所述车牌号的情况下，截取所述车辆区域作为初始的目标车辆区域进行跟踪，以及截取所述车牌区域及所述车牌周边的区域作为初始的目标车牌区域进行跟踪；
    在检测到所述跟踪检测区域中的车辆，但未检测到所述车辆的车牌的情况下，截取所述车辆区域作为初始的目标车辆区域进行跟踪；
    在未检测到所述跟踪检测区域图像中的车辆，但检测到所述车辆的车牌或识别到所述车牌号中的情况下，截取所述车牌图像及所述车牌图像周边的区域作为初始的目标车牌区域进行跟踪。
  11. 根据权利要求10所述的停车管理方法，其中，对所述车辆或所述车牌进行跟踪，包括：
    截取所述车辆的最新图像，对所述目标车辆区域进行更新；
    截取所述车辆的最新车牌图像及所述车牌图像周边区域的图像，对所述目标车牌区域进行更新。
  12. 根据权利要求10所述的停车管理方法，其中，对所述车辆、车牌进行跟踪，确定停车事件信息，包括：
    在所述车辆从所述入场检测区越过车位线进入所述车位区域中的车位的情况下，计算所述车辆在所述车位上停留的时间；
    在所述车牌从所述入场检测区越过车位线进入所述车位区域中的车位的情况下， 计算所述车牌在所述车位上停留的时间；
    在所述车辆在所述车位上停留的时间超过第二阈值的情况下，确定所述车辆为入场车辆、所述车辆的停车事件为入场停车事件，以及确定入场停车事件信息，并将所述入场停车事件信息记录到在场停车信息表中；
    在所述车牌在所述车位上停留的时间超过第二阈值的情况下，确定所述车牌标识的车辆为入场车辆、所述车牌标识的车辆的停车事件为入场停车事件，以及确定入场停车事件信息，并将所述入场停车事件信息记录到在场停车信息表中；
    所述入场停车事件信息包括车牌号、入场时间、所述入场停车事件发生时车辆所在的车位、入场关键点图像、入场关键时序图像、入场跟踪视频。
  13. 根据权利要求12所述的停车管理方法，还包括：
    在所述车辆在所述车位上停留的时间超过所述第二阈值的情况下，停止跟踪所述车辆，对所述车辆的运动检测继续进行；
    在所述车牌在所述车位上停留的时间超过所述第二阈值的情况下，停止跟踪所述车牌，对所述车牌的运动检测继续进行。
  14. 根据权利要求12所述的停车管理方法，其中，
    所述入场关键点图像包括：所述车辆在入场检测区域中首次被检测到的图像、所述车辆在入场检测区域中首次被检测到的车牌图像、所述车辆从入场检测区域越过车位线进入车位区域的图像、所述车辆停在车位上的图像。
  15. 根据权利要求12所述的停车管理方法，其中，
    所述入场关键时序图像为从所述入场跟踪视频中选择提取得到的过程图像。
  16. 根据权利要求12所述的停车管理方法，其中，
    所述入场跟踪视频为所述车辆在入场检测区域中首次被检测到至所述车辆停在车位上的视频段。
  17. 根据权利要求10所述的停车管理方法，其中，对所述车辆、车牌进行跟踪，确定停车事件信息，包括：
    在所述车辆从所述车位区域中的车位进入所述出场检测区域的情况下，确定所述车辆为出场车辆，以及确定所述车辆的出场停车事件信息；
    在所述车牌从所述车位区域中的车位进入所述出场检测区域的情况下，确定所述车牌标识的车辆为出场车辆，以及确定所述车牌标识的车辆的出场停车事件信息；
    所述出场停车事件信息包括车牌号、出场时间、所述出场停车事件发生时车辆所 在的车位、出场关键点图像、出场关键时序图像、出场跟踪视频。
  18. 根据权利要求17所述的停车管理方法，其中，
    所述出场关键点图像包括：所述车辆离场前停在车位上的图像、所述车辆在出场检测区域中首次被检测到的图像、所述车辆在出场检测区域中首次被检测到的车牌图像、所述车辆从车位区域越过车位线进入出场检测区域的图像、所述车辆在出场检测区域中最后一次被检测到的图像、所述车辆在出场检测区域中最后一次被检测到的车牌图像。
  19. 根据权利要求17所述的停车管理方法，其中，
    所述出场关键时序图像为从所述出场跟踪视频中选择提取得到的过程图像。
  20. 根据权利要求17所述的停车管理方法，其中，
    所述出场跟踪视频包括第一出场跟踪视频、第二出场跟踪视频，所述第一出场跟踪视频为所述车辆从车位区域进入出场检测区域直至从出场检测区域消失的视频段，所述第二出场跟踪视频为所述车辆在出场检测区域中首次被检测到至所述车辆在出场检测区域中消失的视频段。
  21. 根据权利要求10所述的停车管理方法，其中，对所述车辆或车牌进行跟踪，确定停车事件信息包括：
    在所述出场检测区域中跟踪到车牌，以及识别所述车牌的车牌号的情况下，根据所述车牌号，查找在场车辆信息表中是否已记录所述车牌号；
    在所述车牌号之前已被记录的情况下，确定所述车牌标识的车辆为出场车辆，以及确定所述车牌标识的车辆的出场停车事件信息，所述出场停车事件信息包括车牌号、出场时间、所述车牌标识的车辆所在的车位。
  22. 一种停车管理装置，所述装置包括：
    处理器(201)，用于接收摄像机拍摄的监控区域图像，以及对所述监控区域图像进行划分，得到所述监控区域图像中的跟踪检测区域，以及对所述跟踪检测区域进行车辆、车牌监测，确定停车事件信息；
    存储器(203)，用于存储所述监控区域图像、停车事件信息。
  23. 根据权利要求22所述的停车管理装置，其中，
    所述跟踪检测区域包括车位区域、入场检测区域、出场检测区域；
    所述停车事件信息包括车辆的车牌号、停车事件的类型、所述停车事件发生时所述车辆所在的车位。
  24. 根据权利要求22或23中所述的停车管理装置，还包括：
    显示器(202)，用于显示操作管理界面；
    所述操作管理界面用于接收用户输入的第一查询指令，所述第一查询指令为用户选择查询所述摄像机的监控信息的操作，所述操作管理界面根据所述第一查询指令，获取所述摄像机拍摄的监控区域图像，以及将所述监控区域图像显示，所述操作管理界面接收用户在所述监控区域图像中选择的跟踪检测区域坐标，所述跟踪检测区域的坐标为划分所述监控区域图像的依据。
  25. 根据权利要求22-24任一项所述的停车管理装置，其中，所述处理器(201)进一步被配置为：
    在所述各跟踪检测区域中进行车辆检测，确定所述各跟踪检测区域中的车辆，
    在所述各跟踪检测区域中进行车牌检测，确定所述各跟踪检测区域中的车牌，
    对所述车牌进行识别，确定所述车牌的车牌号，
    对所述车辆、所述车牌进行运动检测，确定所述车辆、所述车牌的运动状态，所述运动状态包括静止和运动，
    根据所述车辆、所述车牌的运动状态确定所述车辆、所述车牌的所述停车事件信息。
  26. 根据权利要求25所述的停车管理装置，其中，所述处理器(201)进一步被配置为：
    获取所述车牌与所述车辆的位置关系，
    在所述车牌的车牌区域位于所述车辆的车辆区域内的情况下，确定所述车牌检测和所述车辆检测的结果为既检测到所述车辆，又检测到所述车辆的车牌和/或所述车牌的车牌号，
    在所述车牌的车牌区域没有位于任何车辆的车辆区域内的情况下，确定所述车牌检测和车辆检测的结果为仅检测到所述车牌和/或所述车牌的车牌号，但未检测到所述车牌标识的车辆，
    在所述车辆的车辆区域没有包含任何车牌的车牌区域的情况下，确定所述车牌检测和所述车辆检测的结果为仅检测到所述车辆，但未检测到所述车辆的车牌。
  27. 根据权利要求26所述的停车管理装置，其中，所述处理器(201)进一步被配置为：
    在所述车辆的运动状态为静止的情况下，根据所述车辆所在的跟踪检测区域，确 定所述车辆的停车事件信息，
    在所述车牌的运动状态为静止的情况下，根据所述车牌所在的跟踪检测区域，确定所述车牌标识的车辆的停车事件信息，
    在所述车辆的运动状态为运动的情况下，对所述车辆进行跟踪，确定所述车辆的停车事件信息，
    在所述车牌的运动状态为运动的情况下，对所述车牌进行跟踪，确定所述车牌标识的车辆的停车事件信息。
  28. 根据权利要求27所述的停车管理装置，其中，所述处理器(201)进一步被配置为：
    在所述车牌位于车位区域，以及识别到所述车牌的车牌号的情况下，根据所述车牌号，查找在场车辆信息表中是否记录有所述车牌号，
    在所述在场车辆信息表没有记录所述车牌号的情况下，确定所述车牌标识的车辆为入场车辆，以及确定并记录所述车牌标识的车辆的入场停车事件信息，所述车牌标识的车辆的入场停车事件信息包括车辆的车牌号、车辆的入场时间、车辆所在的车位，
    在所述车牌位于入场检测区域或者出场检测区域的情况下，计算所述车牌滞留在入场停车区域或者出场停车区域的时间，
    在所述时间超过第一阈值的情况下，确定所述车牌标识的车辆为违规车辆，所述车牌标识的车辆的停车事件为违规停车事件。
  29. 根据权利要求27所述的停车管理装置，其中，所述处理器(201)进一步被配置为：
    在所述车辆位于车位区域，以及未识别到所述车辆的车牌号的情况下，根据所述车辆，查找所述在场车辆信息表是否记录有所述车辆，
    在所述在场车辆信息表没有记录所述车辆的情况下，确定所述车辆为入场车辆，以及确定并记录所述车辆的入场停车事件信息；所述车辆的入场停车事件信息包括车辆的入场时间、车辆所在的车位，
    在所述车辆位于入场检测区域或者出场检测区域的情况下，计算所述车辆滞留在入场停车区域或者出场停车区域的时间，
    在所述时间超过所述第一阈值的情况下，确定所述车辆为违规车辆，所述车辆的停车事件为违规停车事件。
  30. 根据权利要求26至29任一项所述的停车管理装置，其中，所述处理器(201) 进一步被配置为：
    在检测到所述跟踪检测区域图像中的车辆，以及检测到所述车辆的车牌或车牌号中的情况下，截取所述车辆区域作为初始的目标车辆区域进行跟踪，以及截取所述车牌区域及所述车牌周边的区域作为初始的目标车牌区域进行跟踪，
    在检测到所述跟踪检测区域中的车辆，但未检测到所述车辆的车牌的情况下，截取所述车辆区域作为初始的目标车辆区域进行跟踪，
    在未检测到所述跟踪检测区域图像中的车辆，但检测到所述车辆的车牌或车牌号中的情况下，截取所述车牌图像及所述车牌图像周边的区域作为所述初始的目标车牌区域进行跟踪。
  31. 根据权利要求30所述的停车管理装置，其中，所述处理器(201)进一步被配置为：
    截取所述车辆最新的图像，对所述目标车辆区域进行更新，
    截取所述车辆最新的车牌图像及车牌图像周边区域的图像，对所述目标车牌区域进行更新。
  32. 根据权利要求30所述的停车管理装置，其中，所述处理器(201)进一步被配置为：
    在所述车辆从所述入场检测区越过车位线进入所述车位区域中的车位的情况下，计算所述车辆在所述车位上停留的时间，
    在所述车牌从所述入场检测区越过车位线进入所述车位区域中的车位的情况下，计算所述车牌在所述车位上停留的时间，
    在所述车辆在所述车位上停留的时间超过第二阈值的情况下，确定所述车辆为入场车辆、所述车辆的停车事件为入场停车事件，以及确定入场停车事件信息，并将所述入场停车事件信息记录到在场停车信息表中，
    在所述车牌在所述车位上停留的时间超过第二阈值的情况下，确定所述车牌标识的车辆为入场车辆、所述车牌标识的车辆的停车事件为入场停车事件，以及确定入场停车事件信息，并将所述入场停车事件信息记录到在场停车信息表中，
    所述入场停车事件信息包括车牌号、入场时间、所述入场停车事件发生时车辆所在的车位、入场关键点图像、入场关键时序图像、入场跟踪视频。
  33. 根据权利要求32所述的停车管理装置，其中，所述处理器(201)进一步被配置为：
    在所述车辆在所述车位上停留的时间超过所述第二阈值的情况下，停止跟踪所述车辆，对所述车辆的运动检测继续进行，
    在所述车牌在所述车位上停留的时间超过所述第二阈值的情况下，停止跟踪所述车牌，对所述车牌的运动检测继续进行。
  34. 根据权利要求32所述的停车管理装置，其中，
    所述入场关键点图像包括：所述车辆在入场检测区域中首次被检测到的图像、所述车辆在入场检测区域中首次被检测到的车牌图像、所述车辆从入场检测区域越过车位线进入车位区域的图像、车辆停在车位上的过程图像。
  35. 根据权利要求32所述的停车管理装置，其中，
    所述入场关键时序图像为从所述入场跟踪视频中选择提取得到的图像。
  36. 根据权利要求32所述的停车管理装置，其中，
    所述入场跟踪视频为所述车辆在入场检测区域中首次被检测到至所述车辆停在车位上的视频段。
  37. 根据权利要求30所述的停车管理装置，其中，
    所述处理器(201)在所述车辆从所述车位区域中的车位进入所述出场检测区域的情况下，确定所述车辆为出场车辆，以及确定所述车辆的出场停车事件信息；所述处理器(201)在所述车牌从所述车位区域中的车位进入所述出场检测区域的情况下，确定所述车牌标识的车辆为出场车辆，以及确定所述车牌标识的车辆的出场停车事件信息；所述出场停车事件信息包括车牌号、出场时间、出场停车事件发生时车辆所在的车位、出场关键点图像、出场关键时序图像、出场跟踪视频。
  38. 根据权利要求37所述的停车管理装置，其中，
    所述出场关键点图像包括所述车辆离场前停在车位上的图像、所述车辆在出场检测区域中首次被检测到的图像、所述车辆在出场检测区域中首次被检测到的车牌图像、所述车辆从车位区域越过车位线进入出场检测区域的图像、所述车辆在出场检测区域中最后一次被检测到的图像、所述车辆在出场检测区域中最后一次被检测到的车牌图像。
  39. 根据权利要求37所述的停车管理装置，其中，
    所述出场关键时序图像为从所述出场跟踪视频中选择提取得到的过程图像。
  40. 根据权利要求37所述的停车管理装置，其中，
    所述出场跟踪视频包括第一出场跟踪视频、第二出场跟踪视频；所述第一出场跟 踪视频为所述车辆从车位区域进入出场检测区域直至从出场检测区域消失的视频段；所述第二出场跟踪视频为所述车辆在出场检测区域中首次被检测到至所述车辆停在出场检测区域中消失的视频段。
  41. 根据权利要求30所述的停车管理装置，其中，
    所述处理器(201)在所述出场检测区域中跟踪到车牌，以及识别所述车牌的车牌号的情况下，根据所述车牌号，查找在场车辆信息表中是否已记录所述车牌号，在所述车牌号之前已被记录的情况下，确定所述车牌标识的车辆为出场车辆，以及所确定述车牌标识的车辆的出场停车事件信息，所述出场停车事件信息包括车牌号、出场时间、所述车牌标识的车辆所在的车位。
  42. 一种停车管理装置，所述装置包括：
    输入模块(301)，用于接收摄像机拍摄的监控区域图像；
    图像划分模块(302)，用于对所述监控区域图像进行划分，得到所述监控区域图像中的跟踪检测区域；
    检测识别跟踪模块(303)，用于对所述跟踪检测区域进行车辆、车牌监测，确定停车事件信息。
  43. 根据权利要求42所述的停车管理装置，其中，
    所述跟踪检测区域包括车位区域、入场检测区域、出场检测区域；
    所述停车事件信息包括所述车辆的车牌号、停车事件的类型、所述停车事件发生时所述车辆所在的车位。
  44. 根据权利要求43所述的停车管理装置，其中，
    所述入场检测区域包括与所述车位区域的边界线临近的区域；
    所述出场检测区域包括所述车位区域之外的区域。
  45. 根据权利要求42-44任一项所述的停车管理装置，还包括：
    操作管理界面(304)，用于接收用户输入的第一查询指令，所述第一查询指令为用户选择查询所述摄像机的监控信息的操作，根据所述第一查询指令，获取所述摄像机拍摄的第一监控区域图像，以及将所述第一监控区域图像显示，接收用户在所述第一监控区域图像中选择的跟踪检测区域坐标，所述跟踪检测区域的坐标为所述摄像机的图像划分依据。
  46. 根据权利要求42-45任一项所述的停车管理装置，其中，所述检测识别跟踪模块(303)包括：
    车辆检测模块(3031)，用于采用车辆检测算法对跟踪检测区域进行车辆检测，确定所述跟踪检测区域中的车辆；
    车牌检测模块(3032)，用于采用车牌检测算法对跟踪检测区域进行车牌检测，确定所述跟踪检测区域中的车牌；
    车牌识别模块(3033)，用于采用车牌识别算法对所述车牌进行识别，确定所述车牌的车牌号；
    运动检测模块(3034)，用于采用运动检测算法对所述车辆、所述车牌进行运动检测，确定所述车辆、所述车牌的运动状态，所述目标车辆、所述目标车牌的运动状态包括静止和运动；
    跟踪处理模块(3035)，用于根据所述车辆、所述车牌的运动状态确定所述车辆、所述车牌的所述停车事件信息。
  47. 根据权利要求46所述的停车管理装置，还包括运算判断模块(3036)，所述运算判断模块(3036)被配置为：
    获取所述车牌与所述车辆的位置关系，
    在所述车牌的车牌区域位于所述车辆的车辆区域内的情况下，确定所述车牌检测和所述车辆检测的结果为既检测到所述车辆，又检测到所述车辆的车牌和/或所述车牌的车牌号，
    在所述车牌的车牌区域没有位于任何车辆的车辆区域内的情况下，确定所述车牌检测和车辆检测的结果为仅检测到所述车牌和/或所述车牌的车牌号，但未检测到所述车牌标识的车辆，
    在所述车辆的车辆区域没有包含任何车牌的车牌区域的情况下，确定所述车牌检测和所述车辆检测的结果为仅检测到所述车辆，但未检测到所述车辆的车牌。
  48. 根据权利要求47所述的停车管理装置，其中，
    所述跟踪处理模块(3035)在所述车辆的运动状态为静止的情况下，根据所述车辆所在的跟踪检测区域，确定所述车辆的停车事件信息，
    所述跟踪处理模块(3035)在所述车牌的运动状态为静止的情况下，根据所述车牌所在的跟踪检测区域，确定所述车牌标识的车辆的停车事件信息，
    所述跟踪处理模块(3035)在所述车辆的运动状态为运动的情况下，对所述车辆进行跟踪，确定所述车辆的停车事件信息，
    所述跟踪处理模块(3035)在所述车牌的运动状态为运动的情况下，对所述车牌 进行跟踪，确定所述车牌标识的车辆的停车事件信息。
  49. 根据权利要求48所述的停车管理装置，其中，
    在所述车牌位于车位区域，以及所述车牌识别模块(3033)识别到所述车牌的车牌号的情况下，所述跟踪处理模块(3035)根据所述车牌号，查找在场车辆信息表中是否记录有所述车牌号，在所述在场车辆信息表没有记录所述车牌号的情况下，确定所述车牌标识的车辆为入场车辆，以及确定并记录所述车牌标识的车辆的入场停车事件信息，所述车牌标识的车辆的入场停车事件信息包括车辆的车牌号、车辆的入场时间、车辆所在的车位；
    在所述车牌位于入场检测区域或者出场检测区域的情况下，所述跟踪处理模块(3035)计算所述车牌滞留在入场停车区域或者出场停车区域的时间，在所述时间超过第一阈值的情况下，确定所述车牌标识的车辆为违规车辆，所述车牌标识的车辆的停车事件为违规停车事件。
  50. 根据权利要求48所述的停车管理装置，其中，
    在所述车辆位于车位区域，以及所述车牌识别模块(3033)未识别到所述车辆的车牌号的情况下，所述跟踪处理模块(3035)根据所述车牌，查找所述在场车辆信息表是否记录有所述车辆，在所述在场车辆信息表没有记录所述车辆的情况下，确定所述车辆为入场车辆，以及确定并记录所述车辆的入场停车事件信息，所述车辆的入场停车事件信息包括车辆的入场时间、车辆所在的车位；
    在所述车辆位于入场检测区域或者出场检测区域的情况下，所述跟踪处理模块(3035)计算所述车辆滞留在入场停车区域或者出场停车区域的时间，在所述时间超过所述第一阈值的情况下，确定所述车辆为违规车辆，所述车辆的停车事件为违规停车事件。
  51. 根据权利要求47-50任一项所述的停车管理装置，其中，
    在所述车辆检测模块(3031)检测到所述跟踪检测区域图像中的车辆，以及所述车牌检测模块(3032)检测到所述车辆的车牌或所述车牌识别模块(3033)识别到车牌号的情况下，所述跟踪处理模块(3035)启动车辆跟踪模块(30351)截取所述车辆区域作为初始的目标车辆区域进行跟踪，以及启动车牌跟踪模块(30352)截取所述车牌区域及所述车牌周边的区域作为初始的目标车牌区域进行跟踪；
    在所述车辆检测模块(3031)检测到所述跟踪检测区域中的车辆，但所述车牌检测模块(3032)未检测到所述车辆的车牌的情况下，所述跟踪处理模块(3035)启动 车辆跟踪模块(30351)截取所述车辆区域作为初始的目标车辆区域进行跟踪；
    在所述车辆检测模块(3031)未检测到所述跟踪检测区域图像中的车辆，但所述车牌检测模块(3032)检测到所述车辆的车牌或所述车牌识别模块(3033)识别到车牌号的情况下，所述跟踪处理模块(3035)启动车牌跟踪模块(30352)截取所述车牌图像及所述车牌图像周边的区域作为所述初始的目标车牌区域进行跟踪。
  52. 根据权利要求51所述的停车管理装置，其中，
    所述车辆跟踪模块(30351)截取所述车辆的最新图像，对所述目标车辆区域进行更新；
    所述车牌跟踪模块(30352)截取所述车辆的最新车牌图像及所述车牌图像周边区域的图像，对所述目标车牌区域进行更新。
  53. 根据权利要求51所述的停车管理装置，其中，
    所述跟踪处理模块(3035)在所述车辆从所述入场检测区越过车位线进入所述车位区域中的车位的情况下，计算所述车辆在所述车位上停留的时间；
    所述跟踪处理模块(3035)在所述车牌从所述入场检测区越过车位线进入所述车位区域中的车位的情况下，计算所述车牌在所述车位上停留的时间；
    所述跟踪处理模块(3035)在所述车辆在所述车位上停留的时间超过第二阈值的情况下，确定所述车辆为入场车辆、所述车辆的停车事件为入场停车事件，以及确定入场停车事件信息，并将所述入场停车事件信息记录到在场停车信息表中；
    所述跟踪处理模块(3035)在所述车牌在所述车位上停留的时间超过第二阈值的情况下，确定所述车牌标识的车辆为入场车辆、所述车牌标识的车辆的停车事件为入场停车事件，以及确定入场停车事件信息，并将所述入场停车事件信息记录到在场停车信息表中，所述入场停车事件信息包括车牌号、入场时间、所述入场停车事件发生时车辆所在的车位、入场关键点图像、入场关键时序图像、入场跟踪视频。
  54. 根据权利要求53所述的停车管理装置，其中：
    在所述车辆在所述车位上停留的时间超过所述第二阈值的情况下，所述跟踪处理模块(3035)停止跟踪所述车辆，所述运动检测模块(3034)对所述车辆的运动检测继续进行；
    在所述车牌标识的车辆在所述车位上停留的时间超过所述第二阈值的情况下，所述跟踪处理模块(3035)停止跟踪所述车牌标识的车辆，所述运动检测模块(3034)对所述车牌标识的车辆的运动检测继续进行。
  55. 根据权利要求53所述的停车管理装置，其中，
    所述入场关键点图像包括：所述车辆在入场检测区域中首次被检测到的图像、所述车辆在入场检测区域中首次被检测到的车牌图像、所述车辆从入场检测区域越过车位线进入车位区域的图像、车辆停在车位上的图像。
  56. 根据权利要求53所述的停车管理装置，其中，
    所述入场关键时序图像为从所述入场跟踪视频中选择提取得到的图像。
  57. 根据权利要求53所述的停车管理装置，其中，
    所述入场跟踪视频为所述车辆在入场检测区域中首次被检测到至所述车辆停在车位上的视频段。
  58. 根据权利要求51所述的停车管理装置，其中，
    所述跟踪处理模块(3035)在所述车辆从所述车位区域中的车位进入所述出场检测区域的情况下，确定所述车辆为出场车辆，以及确定所述车辆的出场停车事件信息；
    所述跟踪处理模块(3035)在所述车牌从所述车位区域中的车位进入所述出场检测区域的情况下，确定所述车牌标识的车辆为出场车辆，以及确定所述车牌标识的车辆的出场停车事件信息，所述出场停车事件信息包括车牌号、出场时间、出场停车事件发生时车辆所在的车位、出场关键点图像、出场关键时序图像、出场跟踪视频。
  59. 根据权利要求58所述的停车管理装置，其中，
    所述出场关键点图像包括所述车辆离场前停在车位上的图像、所述车辆在出场检测区域中首次被检测到的图像、所述车辆在出场检测区域中首次被检测到的车牌图像、所述车辆从车位区域越过车位线进入出场检测区域的图像、所述车辆在出场检测区域中最后一次被检测到的图像、所述车辆在出场检测区域中最后一次被检测到的车牌图像。
  60. 根据权利要求58所述的停车管理装置，其中，
    所述出场关键时序图像为从所述出场跟踪视频中选择提取得到的过程图像。
  61. 根据权利要求58所述的停车管理装置，其中，
    所述出场跟踪视频包括第一出场跟踪视频、第二出场跟踪视频，所述第一出场跟踪视频为所述车辆从车位区域进入出场检测区域直至从出场检测区域消失的视频段，所述第二出场跟踪视频为所述车辆在出场检测区域中首次被检测到至所述车辆停在出场检测区域中消失的视频段。
  62. 根据权利要求51所述的停车管理装置，其中，
    在所述车牌跟踪模块(30352)在所述出场检测区域中跟踪到车牌，以及所述车牌识别模块(3033)识别所述车牌的车牌号的情况下，根据所述车牌号，所述跟踪处理模块(3035)查找在场车辆信息表中是否已记录所述车牌号，在所述车牌号之前已被记录的情况下，确定所述车牌标识的车辆为出场车辆，以及所确定述车牌标识的车辆的出场停车事件信息，所述出场停车事件信息包括车牌号、出场时间、所述车牌标识的车辆所在的车位。
  63. 一种停车管理系统，包括：
    布置在路侧停车场的杆位上的相机群(401)，用于获取监控区域图像；
    如权利要求22-41中或如权利要求42-62中任一项所述的停车管理装置(402)，用于按图像划分规则对监控区域图像进行图像划分，以及对监控区域图像进行检测识别跟踪处理，确定所述相机群(401)管理的车位上的停车事件信息。
  64. 根据权利要求63所述的停车管理系统，其中，
    所述相机群(401)包括多个相机组，每个所述相机组包括至少一个相机阵列，每个所述相机阵列包括至少一个摄像机，一个杆位上安装一个所述相机组，首末两个杆位上的相机组均包含一个相机阵列，其余杆位上的相机组均包含两个相机阵列。
  65. 一种计算机可读存储介质，其上存储有计算机程序，该程序被处理器执行时实现如权利要求1-21中任一项所述的停车管理方法。
PCT/CN2018/078300 2017-11-16 2018-03-07 基于多摄像机的路侧停车管理方法、装置和系统 WO2019095588A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US16/762,508 US10930151B2 (en) 2017-11-16 2018-03-07 Roadside parking management method, device, and system based on multiple cameras

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201711135187.6 2017-11-16
CN201711135187.6A CN107767673B (zh) 2017-11-16 2017-11-16 一种基于多摄像机的路侧停车管理方法、装置及系统

Publications (1)

Publication Number Publication Date
WO2019095588A1 true WO2019095588A1 (zh) 2019-05-23

Family

ID=61279361

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/078300 WO2019095588A1 (zh) 2017-11-16 2018-03-07 基于多摄像机的路侧停车管理方法、装置和系统

Country Status (3)

Country Link
US (1) US10930151B2 (zh)
CN (1) CN107767673B (zh)
WO (1) WO2019095588A1 (zh)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111652143A (zh) * 2020-06-03 2020-09-11 浙江大华技术股份有限公司 一种车辆检测方法、装置以及计算机存储介质
CN113556506A (zh) * 2020-04-24 2021-10-26 英研智能移动股份有限公司 基于图像的物体追踪方法
CN113611154A (zh) * 2021-08-13 2021-11-05 武汉虹信技术服务有限责任公司 停车位管理方法、装置和系统以及计算机可读存储介质
CN114973055A (zh) * 2022-03-25 2022-08-30 成都臻识科技发展有限公司 车辆运动状态检测方法、装置、设备及存储介质
WO2022198897A1 (zh) * 2021-03-23 2022-09-29 超级视线科技有限公司 一种路侧停车的管理方法及装置
EP3876213B1 (en) * 2020-10-26 2023-11-01 Beijing Baidu Netcom Science Technology Co., Ltd. Traffic monitoring method, apparatus, device and storage medium

Families Citing this family (44)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107767673B (zh) 2017-11-16 2019-09-27 智慧互通科技有限公司 一种基于多摄像机的路侧停车管理方法、装置及系统
CN108510744B (zh) * 2018-03-14 2020-12-29 智慧互通科技有限公司 基于长短焦距相机实现半封闭车场管理的系统及方法
CN108520236A (zh) * 2018-04-09 2018-09-11 迪蒙智慧交通（深圳）有限公司 驶入及驶出车辆信息采集方法、装置及车辆信息处理系统
CN108364480A (zh) * 2018-04-19 2018-08-03 智慧互通科技有限公司 基于多球机联合的路侧停车管理系统
CN108810616B (zh) * 2018-05-31 2019-06-14 广州虎牙信息科技有限公司 目标定位方法、视频显示方法、装置、设备和存储介质
CN108765976B (zh) * 2018-06-21 2020-09-25 智慧互通科技有限公司 路侧平行停车信息管理系统及方法
CN108921955A (zh) * 2018-06-28 2018-11-30 西安艾润物联网技术服务有限责任公司 停车处理方法、装置及计算机可读存储介质
CN110827571A (zh) * 2018-08-14 2020-02-21 华为技术有限公司 一种信息处理方法及其相关设备
CN109035852A (zh) * 2018-08-16 2018-12-18 智慧互通科技有限公司 一种基于杆间监控图像与传感器信号的路侧停车管理系统
CN109377694B (zh) * 2018-10-10 2023-09-05 北京奇虎科技有限公司 社区车辆的监控方法及系统
CN109493634A (zh) * 2018-12-21 2019-03-19 深圳信路通智能技术有限公司 一种基于多设备协同工作的停车管理系统及方法
CN109727484A (zh) * 2019-02-27 2019-05-07 北京猎户智芯科技有限公司 一种用于停车场出入管理的虚拟道闸系统
CN109785664A (zh) * 2019-03-05 2019-05-21 北京悦畅科技有限公司 一种停车场的剩余停车位数量的统计方法和装置
CN111696360A (zh) * 2019-03-15 2020-09-22 杭州海康威视数字技术股份有限公司 一种停车监控方法、系统及摄像机
CN109993138A (zh) * 2019-04-08 2019-07-09 北京易华录信息技术股份有限公司 一种车牌检测与识别方法及装置
CN111798666B (zh) * 2019-04-09 2021-12-24 杭州海康威视数字技术股份有限公司 一种车辆抓拍方法及装置
CN111932901B (zh) * 2019-05-13 2022-08-09 斑马智行网络(香港)有限公司 道路车辆跟踪检测设备、方法及存储介质
CN111951598B (zh) * 2019-05-17 2022-04-26 杭州海康威视数字技术股份有限公司 一种车辆跟踪监测方法、装置及系统
CN110111581A (zh) * 2019-05-21 2019-08-09 哈工大机器人（山东）智能装备研究院 目标识别方法、装置、计算机设备及存储介质
CN110310486A (zh) * 2019-07-08 2019-10-08 深圳市道尔智控科技股份有限公司 一种停车场向服务端传送图片数据的方法
CN110491165A (zh) * 2019-08-21 2019-11-22 杭州逐一科技有限公司 一种停车场运营管理系统
CN112446268B (zh) * 2019-09-04 2024-04-05 浙江宇视科技有限公司 抬杆控制方法、装置、监控设备及监控系统
CN110930729A (zh) * 2019-11-13 2020-03-27 北京智芯原动科技有限公司 一种基于视频的低杆路侧停车检测方法及装置
CN112908019B (zh) * 2019-12-04 2023-07-18 博泰车联网科技(上海)股份有限公司 用于管理车辆泊车的方法、电子设备和计算机存储介质
CN111145558B (zh) * 2019-12-27 2021-04-06 北京容联易通信息技术有限公司 一种基于高点视频监控的违法行为识别方法
CN111405196B (zh) * 2019-12-31 2022-08-02 智慧互通科技股份有限公司 一种基于视频拼接的车辆管理的方法及系统
CN111652106B (zh) * 2020-05-28 2024-02-02 韶关学院 一种目标监控方法、装置、电子设备及其储存介质
CN111739175B (zh) * 2020-08-11 2022-07-05 上海高德威智能交通系统有限公司 一种车辆停车监控方法和装置
CN112217994A (zh) * 2020-09-29 2021-01-12 国网浙江省电力有限公司嘉兴供电公司 一种电力高压线周边安全作业预警方法
CN112382104B (zh) * 2020-11-13 2021-11-02 重庆盘古美天物联网科技有限公司 基于车辆轨迹分析的路侧停车管理方法
CN112836699A (zh) * 2020-11-30 2021-05-25 爱泊车美好科技有限公司 一种基于长时多目标跟踪的泊位出入场事件分析方法
CN113205689B (zh) * 2021-04-09 2022-09-20 超级视线科技有限公司 一种基于多维度的路侧停车入场事件判断方法及系统
CN113191221B (zh) * 2021-04-15 2022-04-19 浙江大华技术股份有限公司 基于全景相机的车辆检测方法、装置以及计算机存储介质
CN113380047A (zh) * 2021-06-22 2021-09-10 重庆盛海科技发展有限公司 一种基于传统摄像机的违规停车检测方法
CN113569648A (zh) * 2021-06-29 2021-10-29 深圳市捷顺科技实业股份有限公司 一种车辆挪位记录的方法及装置
CN113870551B (zh) * 2021-08-16 2023-07-28 清华大学 一种能够鉴别危险和非危险驾驶行为的路侧监控系统
FR3127062A1 (fr) * 2021-09-10 2023-03-17 Cocoparks Dispositif autonome et procédé pour détecter si des emplacements sont occupés ou non
CN114038232B (zh) * 2021-10-28 2022-09-20 超级视线科技有限公司 一种基于边缘端算存结合的路侧停车管理方法及系统
CN114155619A (zh) * 2021-12-09 2022-03-08 济南博观智能科技有限公司 一种对停车位自动监管的方法、装置、介质及系统
CN114973636B (zh) * 2022-05-23 2023-12-29 安徽文康科技有限公司 基于物联网的停车识别匹配系统
CN115240466B (zh) * 2022-09-23 2023-01-10 杭州立方控股股份有限公司 一种地磁与摄像机协同的智能停车管理方法
CN116189137B (zh) * 2022-12-07 2023-08-04 深圳市速腾聚创科技有限公司 车位检测方法、电子设备及计算机可读存储介质
CN115683237B (zh) * 2023-01-04 2023-03-10 中国市政工程西南设计研究总院有限公司 一种智慧交通停车管理系统及方法
CN116863712B (zh) * 2023-09-01 2023-11-28 成都宜泊信息科技有限公司 路侧巡检车精确判断车辆停放位置的方法及系统

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB2441382A (en) * 2006-08-29 2008-03-05 Ranger Services Ltd Automated parking payment system
US7369940B2 (en) * 2004-03-12 2008-05-06 Magna Donnelly Gmbh & Co. Kg Method for operating a display system in a vehicle for driving into a parking space
CN101470967A (zh) * 2007-12-28 2009-07-01 爱信艾达株式会社 导航装置及计算机程序
CN103824474A (zh) * 2014-03-25 2014-05-28 宁波市江东元典知识产权服务有限公司 基于图像识别技术的车位提示系统
CN105957395A (zh) * 2016-05-26 2016-09-21 智慧互通科技有限公司 一种基于摄像机阵列的路侧停车管理系统及其方法
CN107134145A (zh) * 2017-06-10 2017-09-05 智慧互通科技有限公司 基于多类型图像采集的路侧停车管理装置、系统及方法
CN107767673A (zh) * 2017-11-16 2018-03-06 智慧互通科技有限公司 一种基于多摄像机的路侧停车管理方法、装置及系统

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB616976A (en) 1945-12-18 1949-01-31 Otto William Heinz Improvements in or relating to goggles, particularly adapted for use as water goggles
US8698895B2 (en) * 2012-08-06 2014-04-15 Cloudparc, Inc. Controlling use of parking spaces using multiple cameras
JP6618766B2 (ja) * 2015-10-27 2019-12-11 株式会社デンソーテン 画像処理装置および画像処理方法
CN206149402U (zh) * 2016-07-29 2017-05-03 国网河南省电力公司郑州供电公司 基于视觉的目标空间运动轨迹特征的安全管理系统
CN106952477B (zh) * 2017-04-26 2020-01-14 智慧互通科技有限公司 基于多相机图像联合处理的路侧停车管理方法
CN111052198A (zh) * 2017-06-05 2020-04-21 西提弗耶德公司 停放物体检测系统

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7369940B2 (en) * 2004-03-12 2008-05-06 Magna Donnelly Gmbh & Co. Kg Method for operating a display system in a vehicle for driving into a parking space
GB2441382A (en) * 2006-08-29 2008-03-05 Ranger Services Ltd Automated parking payment system
CN101470967A (zh) * 2007-12-28 2009-07-01 爱信艾达株式会社 导航装置及计算机程序
CN103824474A (zh) * 2014-03-25 2014-05-28 宁波市江东元典知识产权服务有限公司 基于图像识别技术的车位提示系统
CN105957395A (zh) * 2016-05-26 2016-09-21 智慧互通科技有限公司 一种基于摄像机阵列的路侧停车管理系统及其方法
CN107134145A (zh) * 2017-06-10 2017-09-05 智慧互通科技有限公司 基于多类型图像采集的路侧停车管理装置、系统及方法
CN107767673A (zh) * 2017-11-16 2018-03-06 智慧互通科技有限公司 一种基于多摄像机的路侧停车管理方法、装置及系统

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113556506A (zh) * 2020-04-24 2021-10-26 英研智能移动股份有限公司 基于图像的物体追踪方法
CN111652143A (zh) * 2020-06-03 2020-09-11 浙江大华技术股份有限公司 一种车辆检测方法、装置以及计算机存储介质
CN111652143B (zh) * 2020-06-03 2023-09-29 浙江大华技术股份有限公司 一种车辆检测方法、装置以及计算机存储介质
EP3876213B1 (en) * 2020-10-26 2023-11-01 Beijing Baidu Netcom Science Technology Co., Ltd. Traffic monitoring method, apparatus, device and storage medium
WO2022198897A1 (zh) * 2021-03-23 2022-09-29 超级视线科技有限公司 一种路侧停车的管理方法及装置
CN113611154A (zh) * 2021-08-13 2021-11-05 武汉虹信技术服务有限责任公司 停车位管理方法、装置和系统以及计算机可读存储介质
CN114973055A (zh) * 2022-03-25 2022-08-30 成都臻识科技发展有限公司 车辆运动状态检测方法、装置、设备及存储介质

Also Published As

Publication number Publication date
US10930151B2 (en) 2021-02-23
CN107767673A (zh) 2018-03-06
CN107767673B (zh) 2019-09-27
US20200380864A1 (en) 2020-12-03

Similar Documents

Publication Publication Date Title
WO2019095588A1 (zh) 基于多摄像机的路侧停车管理方法、装置和系统
KR101404153B1 (ko) 지능형 cctv 통합관제시스템
JP6013923B2 (ja) ビデオエピソードの閲覧及び検索のためのシステム及び方法
CN104754302A (zh) 一种基于枪球联动系统的目标检测跟踪方法
KR101496390B1 (ko) 차량번호인식 시스템
JP2013218679A (ja) 短時間の駐車違反を取り締まるための、ビデオベースの検知装置及び通知装置
CN103726879A (zh) 利用摄像头自动捕捉矿井矿震坍塌并及时记录报警的方法
WO2021012953A1 (zh) 一种巡检方法及装置
KR101832274B1 (ko) 영상 촬영에 의한 지능형 방범 시스템 및 그 수행 방법
CN103929592A (zh) 全方位智能监控设备及方法
CN105185108A (zh) 一种黄网格线违法停车自动抓拍系统
KR101697060B1 (ko) 이벤트 감지 방법 및 이를 실행하는 장치
KR20100103109A (ko) 전방위 카메라 및 팬/틸트/줌 카메라를 이용한 무인 감시 장치 및 방법
CN204833705U (zh) 一种黄网格线违法停车自动抓拍系统
US10586130B2 (en) Method, system and apparatus for providing access to videos
CN102244769B (zh) 物件及其关键人监控系统与方法
CN112601049B (zh) 视频的监控方法、装置、计算机设备及存储介质
CN103824454A (zh) 基于视频事件检测的多区域违法自动抓拍方法
KR101599614B1 (ko) 영상을 활용한 통합형 차량감시 시스템
KR20190026625A (ko) 영상 표시 방법 및 컴퓨터 프로그램, 그 기록매체
CN115601738B (zh) 停车信息获取方法、装置、设备、存储介质及程序产品
KR101859329B1 (ko) 주정차 단속 시스템
TWI514890B (zh) Monitoring method of panoramic surveillance video
JP7293174B2 (ja) 道路周辺物監視装置、道路周辺物監視プログラム
KR102434154B1 (ko) 영상감시시스템에서의 고속 이동물체의 위치 및 모션 캡쳐 방법

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: 18879932

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18879932

Country of ref document: EP

Kind code of ref document: A1