CN108122414A - The detection method and device of car on-board and off-board on highway - Google Patents
The detection method and device of car on-board and off-board on highway Download PDFInfo
- Publication number
- CN108122414A CN108122414A CN201611083131.6A CN201611083131A CN108122414A CN 108122414 A CN108122414 A CN 108122414A CN 201611083131 A CN201611083131 A CN 201611083131A CN 108122414 A CN108122414 A CN 108122414A
- Authority
- CN
- China
- Prior art keywords
- car
- pedestrian
- video
- field video
- board
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- 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
Abstract
The present invention relates to video and technical field of image processing, the detection method and device of car on-board and off-board more particularly on highway.The detection method of car on-board and off-board on the highway of the present invention, applied to the video monitoring equipment of monitoring highway, this method includes：Whether car appearance is had according to far visual field video detection；When detecting that car occurs, according to the speed of far visual field video detection car；When the speed for detecting car is less than threshold speed, car is tracked in near-sighted field video；Whether pedestrian's appearance is had according to near-sighted field video and/or far visual field video detection；When having detected that pedestrian occurs, judge that car is separated or merged with the relation of pedestrian according to near-sighted field video and/or far visual field video, determine whether that there are upper visitor or lower objective behaviors.It looks far present invention utilizes far visual field and is conducive to reaction in time and the advantages of near-sighted field resolution capability is conducive to by force accurate judgement, can both reduce on-board and off-board and break rules and regulations the detection missing rate of situation, judgment accuracy can also be improved.
Description
Technical field
The present invention relates to video and technical field of image processing, more particularly on highway the detection method of car on-board and off-board and
Device.
Background technology
In order to avoid contingency, according to the regulation of China's traffic law, car does not allow on-board and off-board on a highway.
However, some operation cars, in order to attract traveller, disregard of law regulation is accommodated to Emergency Vehicle Lane when driving on a highway
Upper objective and lower visitor.For this act of violating regulations, on duty due to having no traffic-police on highway, therefore, it is difficult to artificially collect evidence.
Simultaneously because the uncertainty of scene, it is also difficult to realize that automatic decision is collected evidence by camera.Such as it is detected using wide-angle lens
When, since vehicle, pedestrian image are smaller, thus it is difficult to accurate judgement illegal activities and extraction vehicle characteristics；Use narrow visual field mirror
During head detection, then since field range is relatively narrow, thus it is big to monitor dead angle.
The content of the invention
In view of this, an embodiment of the present invention provides the detection method and device of car on-board and off-board on highway, realize to public affairs
The automatic detection of road car on-board and off-board behavior improves judging nicety rate, reduces detection missing rate.The technical solution is as follows：
In a first aspect, on a kind of highway car on-board and off-board detection method, applied to video monitoring equipment, the method bag
It includes：
Whether car appearance is had according to far visual field video detection；
When detecting that car occurs, according to the speed of car described in far visual field video and/or far visual field video detection；
When the speed for detecting the car is less than threshold speed, the car is tracked in near-sighted field video；
Whether pedestrian's appearance is had according to near-sighted field video and/or far visual field video detection；
When having detected that pedestrian occurs, according to near-sighted field video and/or far visual field video judge the car with it is described
The relation of pedestrian is separation or fusion, determines whether that there are upper visitor or lower objective behaviors.
Selectively, this method further includes：
The speed of the car is being detected less than after threshold speed, crawl photo forms photographic evidence and/or continuous
Video recording forms video recording evidence.
Selectively, this method further includes：
According to the characteristic of car described in near-sighted field video extraction, the characteristic includes the license plate number of the car
Code.
Selectively, the crawl photo, which forms photographic evidence, includes：
In the speed for detecting the car less than after threshold speed, first photo is captured；
When the pedestrian is close to the car or when the pedestrian leaves the car, second photo and the 3rd is captured
Open photo；
When the car leaves, the 4th photo is captured.
Selectively, it is described that the car and the pass of the pedestrian are judged according to near-sighted field video and/or far visual field video
System is separation or fusion, determines whether that there are upper visitors or lower objective behavior to include：
When detecting that the pedestrian and the relative distance of the car becomes smaller, and the pedestrian disappears at the car,
Then judge that the car with the relation of the pedestrian is to merge, determine that there are upper objective behaviors；
Occur when detecting the pedestrian at the car, and become larger with the relative distance of the car, then judge institute
It is to merge that car, which is stated, with the relation of the pedestrian, objective behavior in the presence of determining.
Selectively, the far visual field video is obtained by the first video camera, and the myopia field video passes through the second camera shooting
Machine obtains；Alternatively, the far visual field and the near-sighted field are obtained respectively by two camera lenses of same video camera；It is alternatively, described remote
Visual field video and the near-sighted field video are obtained by same video camera by focusing.
Selectively, it is described to have detected whether that car occurs and/or detected whether pedestrian and occurs being based on Fast R-
CNN algorithm of target detection.
On the other hand, on a kind of highway car on-board and off-board detection device, which includes：
Car detection module, for whether having car appearance according to far visual field video detection；
Velocity measuring module, for when having detected that car occurs, according to the speed of car described in far visual field video detection
Degree；
Car tracking module, for when detect the car speed be less than threshold speed when, in near-sighted field video
Track the car；
Pedestrian detection module, for whether having pedestrian's appearance according to near-sighted field video and/or far visual field video detection；
On-board and off-board judgment module, for when having detected that pedestrian occurs, according to near-sighted field video and/or far visual field video
Judge that the car is separated or merged with the relation of the pedestrian, determine whether that there are upper visitor or lower objective behaviors.
Selectively, which further includes：
On-board and off-board evidence obtaining module is shone in the speed for detecting the car less than after threshold speed, crawl photo is formed
Piece evidence and/or continuously recording form video recording evidence.
Selectively, described device further includes：
Characteristic extracting module, for the characteristic of the car according to near-sighted field video extraction, the characteristic bag
Include the number-plate number of the car.
Selectively, the on-board and off-board judgment module includes：
Upper visitor's judging submodule, detects that the pedestrian and the relative distance of the car become smaller for working as, and the row
People disappears at the car, then judges that the car with the relation of the pedestrian is to merge, and determines that there are upper objective behaviors；
Lower visitor's judging submodule, for when detecting that the pedestrian occurs at the car, and with the phase of the car
It adjusts the distance and becomes larger, then judge that the car with the relation of the pedestrian is to merge, objective behavior in the presence of determining.
Selectively, the on-board and off-board evidence obtaining module includes：
Jogging is taken pictures submodule, for, less than after threshold speed, capturing first in the speed for detecting the car
Photo；
People's vehicle photo module, for when the pedestrian is close to the car or when the pedestrian leaves the car, grabbing
Take second photo and the 3rd photo；
Submodule of taking pictures is sailed out of, when the car leaves, captures the 4th photo.
Selectively, the far visual field video is obtained by the first video camera, and the myopia field video passes through the second camera shooting
Machine obtains；Alternatively, the far visual field video and the near-sighted field video are obtained respectively by two camera lenses of same video camera；Or
Person, the far visual field video and the near-sighted field video are obtained by same video camera by focusing.
Selectively, the car detection module has detected whether that car occurs and/or pedestrian detection module detects whether
Pedestrian occurs being based on Fast R-CNN algorithm of target detection.
The advantageous effect that technical solution provided in an embodiment of the present invention is brought is：
In a manner that far visual field and near-sighted field combine, appearance and its speed of car are detected in far visual field video,
When car speed is less than limit value, near-sighted field video tracking car is called, when detecting that pedestrian occurs, it is to separate also to judge vehicle people
It is fusion, determines whether, there are upper objective or lower visitor, so to be prejudged according to far visual field video, after anticipation has possibility violating the regulations
It recalls the cooperation of near-sighted field video and carries out on-board and off-board judgement, take full advantage of far visual field and look far and be conducive to reaction in time and myopia
The advantages of resolution capability is conducive to by force accurate judgement is broken rules and regulations the detection missing rate of situation so as to can both reduce on-board and off-board,
On-board and off-board can be improved to break rules and regulations the judgment accuracy of situation.The person's law enforcement means that extend road management simultaneously, continuation law enforcement radius,
The technological means automatically processed by video improves law enforcement efficiency, solves act of violating regulations on-site law-enforcing evidence obtaining hardly possible, seized rate
The problem of low.
Description of the drawings
To describe the technical solutions in the embodiments of the present invention more clearly, make required in being described below to embodiment
Attached drawing is briefly described, it should be apparent that, the accompanying drawings in the following description is only some embodiments of the present invention, for
For those of ordinary skill in the art, without creative efforts, other are can also be obtained according to these attached drawings
Attached drawing.
Fig. 1 is a kind of schematic diagram of illustrative implementation environment of the present invention；
Fig. 2 is the flow chart of the detection method of car on-board and off-board on the highway of one embodiment of the invention offer；
Fig. 3 is the structure diagram of the detection device of car on-board and off-board on the highway of another embodiment of the present invention offer；
Fig. 4 is the structure diagram of the illustrative detection device of car on-board and off-board on highway of the present invention.
Specific embodiment
To make the object, technical solutions and advantages of the present invention clearer, below in conjunction with attached drawing to embodiment party of the present invention
Formula is described in further detail.
Fig. 1 is the schematic diagram for implementing a kind of illustrative implementation environment of the present invention.Referring to Fig. 1, which includes：
Monitor terminal 101, vehicle 102, pedestrian 103 and monitoring center 104.
Monitor terminal 101 generally may include image collection device, processor, storage device and communication interface.Such as Fig. 1 institutes
Show, image collection device is, for example, video camera, for recording a video and shooting photo under the control of a processor；Processor runs journey
Sequence controls each device to work, and the video and photo that collect to image collection device analyze and process, and judge whether to disobey
Chapter and extraction evidence；Storage device stores the video and photo of pre-set programs and the acquisition of image collection device；Communication interface is used for
It communicates with monitoring center 104, result violating the regulations and evidence is uploaded by wired or wireless network.Processor, storage device and communication
In a housing, image collection device from housing can stretch out or be separated by one by communication line and housing interface encapsulation
Set a distance is connected.
Vehicle 102 travels on highway, under normal circumstances will not pulling over observing, pedestrian 103 under normal circumstances will not be into
Enter highway.Once but 102 pulling over observing of vehicle and there is pedestrian to appear on highway, just illustrate vehicle 102 be accused of it is violating the regulations up and down
Visitor.Therefore monitoring device 101 needs video monitoring vehicle 102 and pedestrian 103, judges whether to break rules and regulations, and it is violating the regulations to capture vehicle 102
The evidence of on-board and off-board uploads result violating the regulations and evidence to monitoring center 104, and monitoring center 104 achieves result violating the regulations and evidence,
And can automatic or manual generation penalty note, law enforcement agency can also carry out other processing accordingly.Note that the Highway Main of the present invention will
Highway is referred to, this is because on-board and off-board are breakneck behaviors on a highway, it should be appreciated that, the present invention
Equally it is equally applicable to highway or the section similar with highway that other forbid on-board and off-board.
Fig. 2 is the flow chart of the detection method of car on-board and off-board on the highway of one embodiment of the invention offer.Referring to Fig. 2,
This method includes：
S21：Whether car appearance is had according to far visual field video detection；
Far visual field video can be obtained by the first video camera, and the first video camera for example sets wide-angle lens, for example, is propped up
The structural support of strut etc is directed at direction to the car above highway or above side, thus the video recorded is far visual field video.
First video camera recorded video, real-time delivery to video data processing device, be provided in video processing equipment accordingly from regarding
The program of automatic identification car in frequency, operation program identify the car in video, once recognizing car, then enter step
S22, while continue to have detected whether other cars.
It should be understood that those skilled in the art may be employed the video processing technique based on various algorithms come in video from
Dynamic identification car.Fast R-CNN algorithm of target detection may be employed to realize in the present embodiment：Key frame is extracted from video, profit
It is identified with Fast R-CNN algorithm of target detection, whether confirmation wherein has car.Fast R-CNN algorithm of target detection is
A kind of algorithm of target detection based on deep learning, in the present invention it is possible to which first containing various vehicles and pedestrian on various roads
Picture in the scene of face is sent in the video data processing device for including Fast R-CNN algorithm of target detection relative programs
It is trained, obtains that car and the target detection model of pedestrian can be recognized, after test reaches the accuracy of setting, by it in reality
To realize the presence of detection video microbus or pedestrian, recognition speed and accuracy are all higher for middle use.
S22：When having detected that car occurs, according to the speed of far visual field video detection car；
After detecting car, video processing equipment continues to analyze the video that the first video camera transmits in real time, really
Set the goal the travel speed of car.If bus driving speed is normal always, more than preset threshold speed, then need not
Call the second video camera (equipment for obtaining near-sighted field video) that need not run following step into line trace.
It should be understood that the video processing technique based on various algorithms may be employed to determine target visitor in those skilled in the art
The travel speed of vehicle.
S23：When the speed for detecting car is less than threshold speed, car is tracked in near-sighted field video.
The speed of car be less than threshold speed when, illustrate car may reduction of speed parking, finally there is the behavior of on-board and off-board.
It needs to transfer the second video camera therefore, video processing equipment is controlled according to the position of current target car in near-sighted field video
Second Camera location target car, the mode of tracking can be focusing, camera lens movement or both combination etc., so as to obtain myopia
Field video, and be transmitted to video processing equipment and handled.
Step S24：Whether pedestrian's appearance is had according to near-sighted field video and/or far visual field video detection；
In generation during objective behavior, car needs stop, and pedestrian, which goes to, enters car on highway；And when lower objective behavior occurs,
Car is also required to stop, and then pedestrian gets off, and car is left in finally walking.Obviously, objective or lower visitor no matter is gone up, all can
There is pedestrian's appearance, thus detect whether that the required step for being to detect whether on-board and off-board behavior occurs in pedestrian.
In this step, since near-sighted field video resolution is high, near-sighted field video can be fully utilized and complete detection；Certainly
Far visual field video can also be utilized to complete detection.Alternatively, in a manner that near-sighted field video and far visual field video detect simultaneously, only
One of them is wanted to detect pedestrian it can be assumed that there is pedestrian's appearance.
In this step, the video of the first video camera and/or the recording of the second video camera, real-time delivery to video processing equipment,
The program of the automatic identification pedestrian from video accordingly is provided in video processing equipment, runs the program to identify in video
Pedestrian determines no there is pedestrian's appearance.
It should be understood that those skilled in the art may be employed the video processing technique based on various algorithms come in video from
Dynamic identification pedestrian.Fast R-CNN algorithm of target detection equally may be employed to realize in the present embodiment：It is extracted from video crucial
Frame is identified using Fast R-CNN algorithm of target detection, and whether confirmation wherein has pedestrian.
If there is pedestrian occurs, then step S25 is gone to, is not turned then.
Step S25：When having detected that pedestrian occurs, according to near-sighted field video and/or far visual field video judge car with
The relation of pedestrian is separation or fusion, determines whether that there are upper visitor or lower objective behaviors.
When having detected that pedestrian occurs, car and the pass of pedestrian are judged according to near-sighted field video and/or far visual field video
System is separation or fusion, determines whether that there are upper visitor or lower objective behaviors.
If getting on the bus, then the relation of car and pedestrian be from as far as near, it is final to merge, if getting off, then car with
The relation of pedestrian is from closely to remote, final separation.It therefore, can be according to myopia when having detected that pedestrian occurs in this step
Field video judges that car is separated or merged with the relation of pedestrian, the former is upper visitor, and the latter is lower visitor.
In this step, since near-sighted field video resolution is high, complete to detect preferably by near-sighted field video；Certainly also may be used
To complete detection using far visual field video.Alternatively, in a manner that near-sighted field video and far visual field video detect simultaneously, wherein one
It is a detect, both cooperation detect or both and all to detect separation or fusion it is determined that upper objective or lower visitor.
A kind of preferred judgment mode is：When the relative distance for detecting pedestrian and car becomes smaller, and the pedestrian is in visitor
It disappears at vehicle, then judges that car with the relation of pedestrian is merge, determine that there are upper visitor's behaviors；Go out when detecting pedestrian at car
It is existing, and become larger with the relative distance of car, then judge that car with the relation of the pedestrian is to separate, visitor's behavior in the presence of determining.
This can be realized based on the movement locus of both detections in video or each moment position.
It should be appreciated that possible existing upper visitor also has lower visitor.One distance threshold can be set, when pedestrian's first time quilt
With the distance of target car more than the distance threshold when identifying, and the last pedestrian then says by disappearing at close-target car
Bright is upper visitor；It is less than the distance threshold with the distance of target car when pedestrian is identified for the first time, and the last pedestrian exists
Away from disappearing at target car, then explanation is lower visitor.
Method provided in an embodiment of the present invention in a manner that far visual field and near-sighted field combine, is examined in far visual field video
Appearance and its speed of car are surveyed, when car speed is less than limit value, goes in near-sighted field video and tracks car, occur in pedestrian
When, judge that vehicle people is separation or fusion, determine whether, there are upper objective or lower visitor, so to be carried out according to far visual field video pre-
Sentence, anticipation have it is violating the regulations may after recall near-sighted field video cooperation and carry out on-board and off-board judgement, take full advantage of far visual field and look far
Be conducive to the advantages of reaction in time is conducive to by force accurate judgement with near-sighted field resolution capability, break rules and regulations so as to which both on-board and off-board can be reduced
The detection missing rate of situation can also improve on-board and off-board and break rules and regulations the judgment accuracy of situation.Extending road management simultaneously, person holds
Method means, continuation law enforcement radius, the technological means automatically processed by video improve law enforcement efficiency, solve the act of violating regulations and show
The problem of field law enforcement evidence obtaining is difficult, seized rate is low.
In the present embodiment, after on-board and off-board behavior is detected, video data processing device can alarm to monitoring center, by
Monitoring center manually transfers video data confirmation.But more preferably processing mode is automatic evidence-collecting, and reduction is manually paid, therefore this reality
The means of automatic evidence-collecting can also be provided on the basis of automatic detection on-board and off-board behavior is realized by applying example, that is, detect car
Speed is less than after threshold speed, and crawl photo forms photographic evidence and/or continuously recording forms video recording evidence.In this way, true
Recognize there are after on-board and off-board behavior, photographic evidence and/or video recording evidence can be uploaded in monitoring by video data processing device
The heart is punished so as to the driver to vehicles peccancy and/or place company, is improved and is broken rules and regulations to arrest rate and cost violating the regulations, from
And on-board and off-board act of violating regulations is reduced, it ensures traffic safety.
Crawl photo can be carried out when video is shot or carried out after video analysis is complete, i.e. root
Photo is captured in video according to the time shaft of video scene.Since near-sighted field video intermediate-resolution is high, crawl photo preferably exists
It carries out in near-sighted field video, but can also be carried out as needed in far visual field video.
In the present embodiment, the identity of vehicles peccancy can be determined from photo and/or video recording by manual type, but it is more excellent
The mode of choosing is characteristic of the video data processing device from near-sighted field video extraction car, and characteristic can include car plate
Number, body color, vehicle brand etc., the most importantly number-plate number certainly, so that personnel violating the regulations have no way of denying.It can be
The extraction of data characteristics is carried out while car is tracked in near-sighted field video, after can also capturing photo in near-sighted field video
It is extracted in photo, this can realize that details are not described herein by image recognition technology.
When capturing photo evidence obtaining, it is necessary to capture several key photos, such as can so select：
In the speed for detecting car less than after threshold speed, first photo is captured, this shows that car will stop,
It is beginning violating the regulations, thus records the time of beginning violating the regulations.
When pedestrian is close to car or when pedestrian leaves car, crawl second and the 3rd photo, this shows pedestrian and vehicle
It is contacted, the former shows upper visitor, and the latter shows lower visitor.Two distance thresholds can be set therefore：First threshold and small
In the second threshold of first threshold, when distance is in first threshold shooting one, place between video microbus and the position of pedestrian
One is also shot when second threshold.Upper objective or lower visitor is assured that according to the time sequencing of photo, in this way, the first threshold
It is worth the photo of shooting preceding, explanation is upper visitor, and rear, explanation is lower visitor for the photo of first threshold shooting.Second and third photo
It can so capture, one is captured when vehicle and pedestrian begin to appear in same video pictures, will have car in pedestrian
Video pictures capture one again when disappearing, the distance of pedestrian and car is more than the latter in the former, then explanation is upper visitor, otherwise
Explanation is lower visitor, notices that disappearance here can be the size of pedestrian image less than a threshold value or can not identify again.In car
When leaving, the 4th photo is captured, records the time of end violating the regulations.
The mode for capturing photo can be extracted from the key frame of video.In the present invention, same video data can divide
For two-way, all the way as video recording caching, another way is used for being analyzed as successive image frame, while can be grabbed from the another way
Key frame is taken as photo.
This series of photos can form complete chain of evidence, improve the accuracy of law enforcement.And continuously recording can be to violating the regulations
Whole process is recorded, the two cooperation, and evidence is more conclusive.
The present embodiment as described above can obtain far visual field and near-sighted field by twin-lens, but in other embodiments,
Far visual field and near-sighted field can also be obtained by automatic focusing, alignment using single-lens, also belong to the model that the present invention protects
It encloses.The form of video camera can be gunlock, ball machine etc..Twin-lens can come from two video cameras or set there are two camera lens
Same video camera, i.e. binocular camera.
Fig. 3 is the flow chart of the detection device of car on-board and off-board on the highway of another embodiment of the present invention offer.Referring to figure
3, which includes：
Car detection module 31, for whether having car appearance according to far visual field video detection；
Velocity measuring module 32, for when having detected that car occurs, according to the speed of far visual field video detection car；
Car tracking module 33, for when detect car speed be less than threshold speed when, in near-sighted field video with
Track car；
Pedestrian detection module 34, for whether having pedestrian's appearance according to near-sighted field video and/or far visual field video detection；
On-board and off-board judgment module 35, for when having detected that pedestrian occurs, being regarded according to near-sighted field video and/or far visual field
Frequency judges that car is separated or merged with the relation of pedestrian, determines whether that there are upper visitor or lower objective behaviors.
Device provided in an embodiment of the present invention, in a manner that far visual field and near-sighted field combine, in far visual field video detection
The appearance of car and its speed when car speed is less than limit value, call near-sighted field video tracking car, when pedestrian occurs,
Judge that vehicle people is separation or fusion, determine whether, there are upper objective or lower visitor, so to be prejudged according to far visual field video,
Anticipation have it is violating the regulations may after recall near-sighted field video cooperation and carry out on-board and off-board judgement, take full advantage of far visual field and look far favorably
In timely reaction and near-sighted field resolution capability are conducive to accurate judgement by force the advantages of, break rules and regulations situation so as to can both reduce on-board and off-board
Detection missing rate, on-board and off-board can also be improved and broken rules and regulations the judgment accuracy of situation.The person that extends road management enforces the law hand simultaneously
Section, continuation law enforcement radius, the technological means automatically processed by video improve law enforcement efficiency, solve the act of violating regulations scene and hold
The problem of method evidence obtaining is difficult, seized rate is low.
In a kind of possible realization method, which further includes：
On-board and off-board evidence obtaining module forms photo card in the speed for detecting car less than photo after threshold speed, is captured
According to and/or continuously recording formed video recording evidence.
In a kind of possible realization method, which further includes：
Characteristic extracting module, extracts the characteristic of car in near-sighted field video, and characteristic includes the car plate of car
Number.
In a kind of possible realization method, on-board and off-board evidence obtaining module includes：
Jogging is taken pictures submodule, in the speed for detecting car less than after threshold speed, captures first photo；
People's vehicle photo module when pedestrian is close to car or when pedestrian leaves car, captures second photo and the 3rd
Photo；
Submodule of taking pictures is sailed out of, when car leaves, captures the 4th photo.
In a kind of possible realization method, judge that car is separated or melted with the relation of pedestrian according to near-sighted field video
It closes, determines whether that there are upper visitor or lower objective behaviors：
Become smaller when according to the relative distance of near-sighted field video detection to pedestrian and car, and pedestrian disappears at car, then
Judgement car is to merge with the relation of pedestrian, determines that there are upper objective behaviors；
Occur when according to near-sighted field video detection to pedestrian at car, and become larger with the relative distance of car, then judge
Car is to merge with the relation of pedestrian, objective behavior in the presence of determining.
In a kind of possible realization method, far visual field video is obtained by the first video camera, and near-sighted field video passes through second
Video camera obtains；Alternatively, far visual field video and near-sighted field video are obtained respectively by two camera lenses of same video camera；It is alternatively, remote
Visual field video and near-sighted field video are obtained by same video camera by focusing.
In a kind of possible realization method, car detection module according to far visual field video detection whether have car occur and/
Or pedestrian detection module is based on Fast R-CNN algorithm of target detection according to whether near-sighted field video detection has pedestrian to occur.
The alternative embodiment that any combination forms the disclosure may be employed, herein no longer in above-mentioned all optional technical solutions
It repeats one by one.
On the highway provided on the highway provided in the present embodiment in the detection device of car on-board and off-board and a upper embodiment
The method of car on-board and off-board belongs to same inventive concept, and specific implementation process refers to an embodiment, and details are not described herein.
Fig. 4 is the block diagram of the detection device 400 of car on-board and off-board on a kind of highway provided according to embodiments of the present invention.Dress
It for example can be a video monitoring equipment to put 400.With reference to Fig. 4, device 400 includes processing component 422, further comprises one
Or multiple processors and as the memory resource representated by memory 432, for storing what can be performed by processing component 422
Instruction, such as application program.The application program stored in memory 432 can include each one or more correspondence
In the module of one group of instruction.In addition, processing component 522 is configured as executing instruction, to perform the inspection of car on-board and off-board on highway
Survey method.
Device 400 can also include the power management that a power supply module 426 is configured as executive device 400, and one has
Line or radio network interface 450 are configured as device 400 being connected to network and input and output (I/O) interface 458, defeated
Entering to be connected to one or two video camera 460 at output interface 458, video camera 460 provides video data source for processor, and
Receive the control instruction of processor.Device 400 can be operated based on the operating system for being stored in memory 432, such as Windows
ServerTM, Mac OS XTM, UnixTM,LinuxTM, FreeBSDTMOr similar other systems.
One of ordinary skill in the art will appreciate that hardware can be passed through by realizing all or part of step of above-described embodiment
It completes, relevant hardware can also be instructed to complete by program, the program can be stored in a kind of computer-readable
In storage medium, storage medium mentioned above can be read-only memory, disk or CD etc..
The foregoing is merely presently preferred embodiments of the present invention, is not intended to limit the invention, it is all the present invention spirit and
Within principle, any modifications, equivalent replacements and improvements are made should all be included in the protection scope of the present invention.
Claims (14)
1. the detection method of car on-board and off-board on a kind of highway, which is characterized in that applied to video monitoring equipment, the method bag
It includes：
Whether car appearance is had according to far visual field video detection；
When detecting that car occurs, according to the speed of car described in far visual field video detection；
When the speed for detecting the car is less than threshold speed, the car is tracked in near-sighted field video；
Whether pedestrian's appearance is had according to near-sighted field video and/or far visual field video detection；
When having detected that pedestrian occurs, the car and the pedestrian are judged according to near-sighted field video and/or far visual field video
Relation be separation or fusion, determine whether that there are upper visitor or lower objective behaviors.
2. according to the method described in claim 1, it is characterized in that, this method further includes：
After the speed for detecting the car is less than threshold speed, crawl photo forms photographic evidence and/or continuously recording
Form video recording evidence.
3. according to the method described in claim 2, it is characterized in that, this method further includes：
According to the characteristic of car described in near-sighted field video extraction, the characteristic includes the number-plate number of the car.
4. according to the method described in claim 2, it is characterized in that, the crawl photo forms photographic evidence and includes：
In the speed for detecting the car less than after threshold speed, first photo is captured；
When the pedestrian is close to the car or when the pedestrian leaves the car, second photo and the 3rd photograph are captured
Piece；
When the car leaves, the 4th photo is captured.
5. according to the method described in claim 1, it is characterized in that, described sentence according to near-sighted field video and/or far visual field video
The fixed car is separated or merged with the relation of the pedestrian, determines whether that there are upper visitors or lower objective behavior to include：
When detecting that the pedestrian and the relative distance of the car becomes smaller, and the pedestrian disappears at the car, then sentences
The fixed car is to merge with the relation of the pedestrian, determines that there are upper objective behaviors；
Occur when detecting the pedestrian at the car, and become larger with the relative distance of the car, then judge the visitor
Vehicle is to merge with the relation of the pedestrian, objective behavior in the presence of determining.
6. according to claim 1-5 any one of them methods, which is characterized in that the far visual field video passes through the first video camera
It obtains, the myopia field video is obtained by the second video camera；Alternatively, the far visual field and the near-sighted field are by same video camera
Two camera lenses obtain respectively；Alternatively, the far visual field video and the near-sighted field video are burnt by adjusting by same video camera
Away from acquisition.
7. according to claim 1-5 any one of them methods, which is characterized in that it is described detected whether car occur and/or
It is based on Fast R-CNN algorithm of target detection to have detected whether that pedestrian occurs.
8. the detection device of car on-board and off-board on a kind of highway, which is characterized in that the device includes：
Car detection module, for whether having car appearance according to far visual field video detection；
Velocity measuring module, for when having detected that car occurs, according to the speed of car described in far visual field video detection；
Car tracking module, for when the speed for detecting the car is less than threshold speed, being tracked in near-sighted field video
The car；
Pedestrian detection module, for whether having pedestrian's appearance according to near-sighted field video and/or far visual field video detection；
On-board and off-board judgment module, for when having detected that pedestrian occurs, being judged according to near-sighted field video and/or far visual field video
The car is separated or merged with the relation of the pedestrian, determines whether that there are upper visitor or lower objective behaviors.
9. device according to claim 8, which is characterized in that the device further includes：
On-board and off-board evidence obtaining module forms photo card in the speed for detecting the car less than photo after threshold speed, is captured
According to and/or continuously recording formed video recording evidence.
10. device according to claim 9, which is characterized in that described device further includes：
Characteristic extracting module, for the characteristic of the car according to near-sighted field video extraction, the characteristic includes institute
State the number-plate number of car.
11. device according to claim 8, which is characterized in that the on-board and off-board judgment module includes：
Upper visitor's judging submodule, detects that the pedestrian and the relative distance of the car become smaller, and the pedestrian exists for working as
It disappears at the car, then judges that the car with the relation of the pedestrian is to merge, determine that there are upper visitor's behaviors；
Lower visitor's judging submodule, for when detecting that the pedestrian occurs at the car, and with the car with respect to away from
From becoming larger, then judge that the car with the relation of the pedestrian is to merge, objective behavior in the presence of determining.
12. device according to claim 9, which is characterized in that the on-board and off-board evidence obtaining module includes：
Jogging is taken pictures submodule, for, less than after threshold speed, capturing first photo in the speed for detecting the car；
People's vehicle photo module, for when the pedestrian is close to the car or when the pedestrian leaves the car, crawl the
Two photos and the 3rd photo；
Submodule of taking pictures is sailed out of, when the car leaves, captures the 4th photo.
13. according to the device described in claim 8-12 any of which items, which is characterized in that far visual field video passes through the first camera lens
It obtains, near-sighted field video is obtained by the second camera lens；Alternatively, the far visual field video and the near-sighted field video are by same camera shooting
Two camera lenses of machine obtain respectively；Alternatively, far visual field video and near-sighted field video are obtained by same camera lens by focusing.
14. according to the device described in claim 8-12 any of which items, which is characterized in that the car detection module, which detects, is
It is no have car occur and/or pedestrian detection module to have detected whether that pedestrian occurs be based on FastR-CNN algorithm of target detection.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611083131.6A CN108122414A (en) | 2016-11-30 | 2016-11-30 | The detection method and device of car on-board and off-board on highway |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611083131.6A CN108122414A (en) | 2016-11-30 | 2016-11-30 | The detection method and device of car on-board and off-board on highway |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108122414A true CN108122414A (en) | 2018-06-05 |
Family
ID=62226324
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611083131.6A Pending CN108122414A (en) | 2016-11-30 | 2016-11-30 | The detection method and device of car on-board and off-board on highway |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108122414A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109447032A (en) * | 2018-11-14 | 2019-03-08 | 中设设计集团股份有限公司 | The detection method and system of the illegal on-board and off-board of Expressway Service car |
| CN110895662A (en) * | 2018-09-12 | 2020-03-20 | 杭州海康威视数字技术股份有限公司 | Vehicle overload alarm method and device, electronic equipment and storage medium |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201733403U (en) * | 2010-08-05 | 2011-02-02 | 师振斌 | Traffic violation dynamic snapshot system |
| CN202563683U (en) * | 2012-02-17 | 2012-11-28 | 江苏全固信息科技有限公司 | Remote warning device for illegal carrying position of passenger service vehicle |
| CN103150920A (en) * | 2013-01-29 | 2013-06-12 | 北京瑞拓电子技术发展有限公司 | Vehicle illegal parking remote monitoring system |
| CN103824458A (en) * | 2014-03-25 | 2014-05-28 | 宁波市江东元典知识产权服务有限公司 | Temporary parking warning system based on image recognition technology |
| CN103903438A (en) * | 2014-03-14 | 2014-07-02 | 深圳先进技术研究院 | Method and system for recognizing illegal parking positions of passenger service vehicles |
| CN104157143A (en) * | 2014-08-15 | 2014-11-19 | 青岛比特信息技术有限公司 | Illegal parking detection system and detection method thereof |
| CN104537840A (en) * | 2015-01-09 | 2015-04-22 | 上海博康智能信息技术有限公司 | System for detecting illegally operated taxis |
| JP2015191264A (en) * | 2014-03-27 | 2015-11-02 | 株式会社ニコン | Autonomous driving vehicle |
| CN105160871A (en) * | 2015-06-15 | 2015-12-16 | 北京拓明科技有限公司 | Highway passenger vehicle temporary get-on/off recognition method |
| CN204965750U (en) * | 2015-08-06 | 2016-01-13 | 苏州市世跃智能科技有限公司 | Device that illegal automated inspection that parks took a candid photograph based on two balls machine |
| CN105608906A (en) * | 2016-03-29 | 2016-05-25 | 成都理工大学 | System for monitoring illegal emergency lane occupancy of expressway motor vehicles and implementation method |
-
2016
- 2016-11-30 CN CN201611083131.6A patent/CN108122414A/en active Pending
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201733403U (en) * | 2010-08-05 | 2011-02-02 | 师振斌 | Traffic violation dynamic snapshot system |
| CN202563683U (en) * | 2012-02-17 | 2012-11-28 | 江苏全固信息科技有限公司 | Remote warning device for illegal carrying position of passenger service vehicle |
| CN103150920A (en) * | 2013-01-29 | 2013-06-12 | 北京瑞拓电子技术发展有限公司 | Vehicle illegal parking remote monitoring system |
| CN103903438A (en) * | 2014-03-14 | 2014-07-02 | 深圳先进技术研究院 | Method and system for recognizing illegal parking positions of passenger service vehicles |
| CN103824458A (en) * | 2014-03-25 | 2014-05-28 | 宁波市江东元典知识产权服务有限公司 | Temporary parking warning system based on image recognition technology |
| JP2015191264A (en) * | 2014-03-27 | 2015-11-02 | 株式会社ニコン | Autonomous driving vehicle |
| CN104157143A (en) * | 2014-08-15 | 2014-11-19 | 青岛比特信息技术有限公司 | Illegal parking detection system and detection method thereof |
| CN104537840A (en) * | 2015-01-09 | 2015-04-22 | 上海博康智能信息技术有限公司 | System for detecting illegally operated taxis |
| CN105160871A (en) * | 2015-06-15 | 2015-12-16 | 北京拓明科技有限公司 | Highway passenger vehicle temporary get-on/off recognition method |
| CN204965750U (en) * | 2015-08-06 | 2016-01-13 | 苏州市世跃智能科技有限公司 | Device that illegal automated inspection that parks took a candid photograph based on two balls machine |
| CN105608906A (en) * | 2016-03-29 | 2016-05-25 | 成都理工大学 | System for monitoring illegal emergency lane occupancy of expressway motor vehicles and implementation method |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110895662A (en) * | 2018-09-12 | 2020-03-20 | 杭州海康威视数字技术股份有限公司 | Vehicle overload alarm method and device, electronic equipment and storage medium |
| CN109447032A (en) * | 2018-11-14 | 2019-03-08 | 中设设计集团股份有限公司 | The detection method and system of the illegal on-board and off-board of Expressway Service car |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108154686A (en) | A kind of vehicle-mounted act of violating regulations capturing system | |
| CN109686031B (en) | Identification following method based on security | |
| CN101533557B (en) | Traffic information detection method and traffic information detection device | |
| CN113256924A (en) | Monitoring system, monitoring method and monitoring device for rail train | |
| CN108986472B (en) | Method and device for monitoring vehicle turning round | |
| CN111353369B (en) | Application method and system of urban road side parking high-level video in auxiliary criminal investigation | |
| CN104583051A (en) | Surveillance of a railway track | |
| KR20100119476A (en) | An outomatic sensing system for traffic accident and method thereof | |
| CN110580808B (en) | Information processing method and device, electronic equipment and intelligent traffic system | |
| CN106503622A (en) | A kind of vehicle antitracking method and device | |
| CN104809887A (en) | Detecting method for vehicle converse running on expressway and automatic alarm device | |
| CN110288738B (en) | Intelligent comprehensive management and control system and method for bicycle lane | |
| KR102491091B1 (en) | Method for producing collection video clip and, integrated unmanned traffic control system for two/four wheeled vehicle therewith | |
| CN103730009A (en) | Method and system for detecting vehicle with one number plate missing | |
| CN113055649A (en) | Tunnel intelligent video monitoring method and device, intelligent terminal and storage medium | |
| KR101926562B1 (en) | System for image analysis of multi lane | |
| CN111161435A (en) | Novel highway cardless charging system | |
| KR101210615B1 (en) | Regulation system of u-turn violation vehicle | |
| CN207938184U (en) | A kind of vehicle-mounted act of violating regulations capturing system | |
| CN108122414A (en) | The detection method and device of car on-board and off-board on highway | |
| CN106251622B (en) | Vehicle and road monitoring evidence collecting method and its system | |
| CN114120506A (en) | Infrastructure field personnel management and control system and method based on 5G network architecture | |
| CN111383248A (en) | Method and device for judging red light running of pedestrian and electronic equipment | |
| CN105206060A (en) | Vehicle type recognition device and method based on SIFT characteristics | |
| CN114373303A (en) | Mobile modularized public security intelligent inspection station system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180605 |
|
| RJ01 | Rejection of invention patent application after publication |