CN104506804B - Motor vehicle abnormal behaviour monitoring device and its method on a kind of through street - Google Patents
Motor vehicle abnormal behaviour monitoring device and its method on a kind of through street Download PDFInfo
- Publication number
- CN104506804B CN104506804B CN201410810367.XA CN201410810367A CN104506804B CN 104506804 B CN104506804 B CN 104506804B CN 201410810367 A CN201410810367 A CN 201410810367A CN 104506804 B CN104506804 B CN 104506804B
- Authority
- CN
- China
- Prior art keywords
- motor vehicle
- image
- monitoring
- ball machine
- high definition
- 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.)
- Active
Links
Landscapes
- Traffic Control Systems (AREA)
- Closed-Circuit Television Systems (AREA)
- Studio Devices (AREA)
Abstract
The present invention provides motor vehicle abnormal behaviour monitoring method and its device on a kind of through street, including carries out division setting to monitoring region；Single-frame images is decoded as to two road split screen video image streams of panoramic high-definition video camera；Image is carried out to eliminate fuzzy, background refresh process；Motor vehicle is detected, and to it into line trace；Motor-driven vehicle going behavior in image is analyzed；Position where making high definition ball machine go to motor vehicle captures image；The automobile license plate number and its abnormal behaviour that is identified, and will identify that automobile license plate number in image are combined into a caveat and carry out live play, and same alarm.The present invention automatically identifies abnormal traveling behavior of the motor vehicle on various road type through streets and carries out the accurately candid photographs such as detection and image, data; and caveat is broadcasted by wireless transmission and live loudspeaker; driver's normally travel is reminded in time; it is possible to prevente effectively from the generation of major traffic accidents, protects people's life and property safety.
Description
Technical field
The present invention relates to the motor vehicle abnormal behaviour monitoring devices in field of road traffic safety more particularly to through street
And its method.
Background technology
The fast development of social economy realizes interconnecting for highway between city substantially, and urban inner also goes out
Now many expressways, such as enclosed type road, viaduct, motor vehicle travel on these through streets, not only to limit its highest
Speed, and to limit its minimum speed, i.e., the speed that motor vehicle must keep certain travels.Since motor vehicle is on through street
Quickly, slightly mistake is easy for that major traffic accidents occur speed, leads to car crash when driving.Statistics is shown, motor-driven
Parking of the vehicle on through street, the abnormal behaviours such as go slowly, move backward, it has also become after driving over the speed limit, cause after fatigue driving
Another principal element of road traffic accident.Presently mainly by means of the automobile overspeed inspection violating the regulations on through street
Recording equipment is surveyed, overspeed violation motor vehicle can be captured and be punished, efficiently reduce motor vehicle from speeding behavior；It borrows
Help the Policy Conditions limited continuous driving time and vehicle carried driving people's fatigue detecting equipment, can efficiently reduce and drive
Sail the traveling behavior of people's fatigue.But some sections of motor vehicle on through street, as near ring road mouth, near the crossing that crosses, wind
Sight spot nearby etc., is susceptible to the similar abnormal behaviours such as stop, go slowly, moving backward, driving in the wrong direction, due to lacking necessary monitoring and warning
Equipment, these abnormal behaviours have become the main hidden danger of Expressway Traffic safety.
Invention content
The present invention is directed to the deficiency of existing through street traffic safety precautionary technology, provides motor-driven on a kind of through street
Vehicle abnormal behaviour monitoring device and its method.It is based on panoramic high-definition video camera and carries out a wide range of video monitoring to through street,
And video image is handled in system monitoring host, the abnormal row such as identify the parking of motor vehicle, go slowly, move backward, driving in the wrong direction
Sail behavior, while obtaining the position of abnormal motor vehicle, then control integrated high definition ball machine to abnormal motor vehicle carry out videoscanning,
It captures, obtains the clear image of abnormal motor vehicle and its license plate, and generate specific aim using literary periodicals technology and warn term,
By high-pitched speaker in road live play, abnormal motor vehicle normally travel is guided, to eliminate hidden peril of accident.
A kind of motor vehicle abnormal behaviour monitoring method on through street, includes the following steps：
(1) monitoring region is divided into 254 monitoring sub-districts, each monitoring sub-district is numbered, and make monitoring sub-district
Number is consistent with the preset position number of high definition ball machine；
(2) two road split screen video image streams of panoramic high-definition video camera are decoded as single-frame images；
(3) single-frame images in (2) is carried out eliminating fuzzy, background refresh process；
(4) motor vehicle in image after processing is detected, and to it into line trace；
(5) the motor-driven vehicle going behavior in image is analyzed；If motor vehicle is in a certain position remains stationary state
More than 5 seconds, then it is assumed that it belongs to abnormal " STOP " behavior；If motor vehicle motion direction is differed with road regulation travel direction
It causes, then it is assumed that it belongs to abnormal " retrograde " behavior；If the direction of motion of motor vehicle changes, then it is assumed that it belongs to abnormal
" reversing " behavior；If the travel speed of motor vehicle is less than 10 kilometers/hour, then it is assumed that it belongs to abnormal " going slowly " behavior；
(6) it is used as the position of the motor vehicle according to the above-mentioned monitoring sub-district number being abnormal at the motor vehicle of behavior
Data are set, and the position data is converted to the preset position number of high definition ball machine, high definition ball machine is made to go to machine from initial position
Motor-car position；
(7) a frame image is captured from the high clear video image stream of high definition ball machine, you can obtain the exception of a high-resolution
Motor vehicle image；
(8) automobile license plate number in image is identified, the automobile license plate number that will identify that and its abnormal behaviour
It is combined into a caveat, and is simultaneously emitted by sound and light alarm signal；Caveat is broadcasted by IP network again and is showed in through street
It plays out, while high definition ball machine is captured into image in relation to the exception motor vehicle, time, position data, the number-plate number, different
Normal behavior type is stored；If the motor vehicle number-plate number cannot be identified, sound and light alarm signal is sent out.
Further scheme, the middle Fuzzy Processing of eliminating of the step (3) is to eliminate video camera using laplacian spectral radius algorithm
Image caused by shake is fuzzy, and eliminates the image mould caused by misty rain thick weather using histogram enhancement defogging algorithm
Paste；Background refresh process is refreshed to image background using sliding average background model algorithm, to eliminate image light and shade, bright
The influence spent variation and brought.
Further scheme is detected motor vehicle in the step (4), is to use mixture Gaussian background model algorithm pair
Motor vehicle is detected in image；To motor vehicle into line trace be using Blob Tracking algorithms to the motor vehicle that detects
Into line trace.
Further scheme, it is to use Kalman to carry out analysis to the motor-driven vehicle going behavior in image in the step (5)
Filtering algorithm analyzes motor-driven vehicle going behavior.
Further scheme, position where high definition ball machine goes to motor vehicle from initial preset position in the step (6)
Time is 2-3.5 seconds；
Further scheme captures a frame image conduct in the step (7) from the high clear video image stream of high definition ball machine
Abnormal motor vehicle image is to postpone to be captured again at least 4.5 seconds when rotating initial preset position since high definition ball machine.
Another goal of the invention of the present invention is to provide the monitoring device for realizing above-mentioned monitoring method, including field monitoring list
Member and backstage Supervisory Unit, the field monitoring unit includes on the cross bar for the rod end for being mounted on through street roadside
Panoramic high-definition video camera, high definition ball machine and high-pitched speaker, cabinet and the first wireless bridge, the machine are installed in the upright bar
IP network broadcast power amplifier and first network interchanger are equipped in case；First wireless bridge, panoramic high-definition video camera, high definition
Ball machine and IP network broadcast power amplifier are connect with the first network interchanger interactive mode respectively, and the IP network broadcasts power amplifier
Output end connect with the high-pitched speaker；
The backstage Supervisory Unit includes monitoring host computer, and the monitoring host computer is wireless by second network switch and second
Bridge interactive mode connects；
Nothing is carried out by the first wireless bridge, the second wireless bridge between the field monitoring unit and backstage Supervisory Unit
Line data transmission.
Further scheme, the backstage Supervisory Unit further include the ball machine being connect respectively with the second network switch interactive mode
Monitoring station and panoramic camera monitoring station, the ball machine monitoring station pass sequentially through second network switch, the second wireless bridge,
One wireless bridge and first network interchanger drive high definition ball machine；The panoramic camera monitoring station passes sequentially through the second network
Interchanger, the second wireless bridge, the first wireless bridge and first network interchanger drive panoramic camera to work.
Further scheme, second network switch interactive mode are connected with server and disk array.
The present invention carries out interframe sliding window filtering using Kalman filter algorithm to obtained vehicle position coordinate, i.e., motor-driven
The frame test position of vehicle is equal to the average value of its front and back frame test position in time series, and deviation is missed less than its position estimation value
2 times of difference.This method can eliminate influence of the outside noise to motor vehicle detecting position, accurately detect that the position of motor vehicle is sat
Mark provides accurate presetting bit information for the scanning candid photograph of high definition ball machine, whether there is for motor vehicle and stop, go slowly, driving in the wrong direction, falling
The behavioural analyses such as vehicle provide reliable position coordinate value.
In the present invention panoramic high-definition video camera, high definition ball machine and the first wireless bridge be mounted on through street roadside it is vertical
On the crossbeam of bar：Wherein panoramic high-definition video camera can acquire the live 360 degree of full-view video images of through street, and video image is differentiated
Rate is 2048 × 1536, and after H.264 protocol code, the panoramic camera in the Supervisory Unit of backstage is transmitted to 15 frames/second
Monitoring station.High definition ball machine receives the scanning that monitoring host computer is sent in the Supervisory Unit of backstage and captures order and abnormal vehicle position number
According to control high definition ball machine holder realizes that pick-up lens is directed toward abnormal motor vehicle by horizontal rotation or flip vertical, then controls camera shooting
The high clear video image of lens focus acquisition amplification obtains clearly abnormal motor vehicle and its license plate image, video image
Resolution ratio is 1920 × 1080, and after H.264 protocol code, the prison of the ball machine in the Supervisory Unit of backstage is transmitted to 30 frames/second
Control station.Under normal conditions, meeting auto-returned pick-up lens is preset after the abnormal Vehicle video image of high definition ball machine acquisition 3 seconds
Position；High definition ball machine can be controlled by monitoring host computer, manual control can also be realized on ball machine monitoring station, and artificial
Auto-returned preset position after control terminate 30 seconds, and the priority of manual control is higher than automatically controlling.First wireless network
Bridge and the second wireless bridge for being installed on backstage Supervisory Unit constitute radio communication chain circuit, and wireless bridge Visual communications distance is up to 50
Kilometer, transmission rate is more than 150MBPS, can be completed well between road field unit equipment and backstage Supervisory Unit equipment
HD video, HD Audio and control data Wireless transceiver.
IP network broadcasts power amplifier and first network interchanger is installed in the cabinet in the upright bar of through street roadside, IP network
Network broadcast power amplifier receives the high-fidelity phonetic warning information that monitoring host computer is sent, and after carrying out Audio power amplifier, and is sent to
High-pitched speaker plays out, the driver for warning abnormal motor vehicle on through street；Phonetic warning information includes monitoring host computer
The voice of synthesis and propagandaing directed to communicate for backstage operator on duty.Second network switch of first network interchanger and backstage Supervisory Unit is all
With good environmental suitability, realizes the transmission of the data such as the video between each equipment, audio and interconnect.
Panoramic camera monitoring station handles 360 degree of full-view video images that panoramic camera is passed back, and spherical surface is regarded
Frequency image spread is 90 degree of four split screen video image, and two screen video images of corresponding through street monitoring are transmitted to monitoring master
Machine is handled；Four split screen video images are stored in disk array simultaneously, convenient for consulting and reprinting later；Panoramic camera
Monitoring station can be adjusted the parameters such as the video image resolution ratio of panoramic camera, brightness, saturation degree, contrast.
The high clear video image that ball machine monitoring station passes high definition ball machine back is shown, while high clear video image is preserved
In disk array, convenient for consulting and reprinting later；The holder and camera lens of high definition ball machine can be manually controlled simultaneously, it is right
Specifically monitored region carries out picture amplification display；Using 255 presetting bits of ball machine, wherein 1 is used for ball machine initial preset position,
Remaining 254 presetting bit corresponds to 254 monitoring sub-districts on panoramic shooting machine monitoring road surface, realizes that high definition ball machine supervises whole road surfaces
Control the high clear video image covering in region.When work, ball machine monitoring station receives the scanning sent of monitoring host computer and captures order and different
Normal vehicle position data, for vehicle position data with ball machine presetting bit at one-to-one relationship, ball machine monitors stand control high definition
Ball machine is directed toward corresponding presetting bit, and the high definition video image capture of abnormal motor vehicle is returned, system monitoring host is transmitted to, adopts
It is 3 seconds to collect the time, then auto-returned initial preset position.
Monitoring host computer is the key process unit of whole system, and pending video figure is received from panoramic camera monitoring station
As stream, motor vehicle is detected in the monitoring area that video image delimited, and analyze motor-driven vehicle going behavior, if detection has
The similar abnormal behaviours such as stop, go slowly, moving backward, driving in the wrong direction, then output abnormality motor vehicle behavior type and position data progress be subsequently
Processing；In remaining processing sequences, monitoring host computer collects the high-definition image of abnormal motor vehicle, and profit by ball machine monitoring station
Abnormal motor vehicle trade mark is obtained with motor vehicle trade mark identification software, according to " license plate number " and " abnormal behaviour type ", using literary language
Switch technology synthesizes a caveat, and the high-pitched speaker for being sent to road scene is broadcasted, and reminds abnormal motor vehicle driving
Member carries out normal driving, while the information such as abnormal motor vehicle photo, license plate number, abnormal behaviour type are recorded in system server
In.
When server occurs for the photo of long-term preservation exception motor vehicle, license plate number, abnormal behaviour type, abnormal behaviour
Between, the information such as place, can be for statistical analysis to motor vehicle abnormal behaviour, and can easily be retrieved, be consulted, Ke Yiwei
Further punishment provides data.Disk array is used to preserve panoramic high-definition video camera, the high definition of integrated high definition ball machine acquisition regards
Frequency image, in case retrieving and consulting from now on.
So it is normal on through street to automatically identify motor vehicle on the basis of realizing road video monitoring by the present invention
See abnormal traveling behavior, and caveat can be broadcasted by loudspeaker, driver's normally travel is reminded in time, it is possible to prevente effectively from great
People's life and property safety are protected in traffic accident.
The present invention adds manual intervention step in the program of monitoring device automatic work, can greatly avoid system
The motor vehicle abnormal behaviour occurred due to failure, environment influence when work is failed to judge, and it is different to handle some unconventional motor vehicles
Chang Hangwei improves the reliability and applicability of equipment work.
The data records such as photo, time, position, abnormal behaviour type that the present invention captures motor vehicle abnormal behaviour preserve
In server and disk array, it is convenient for big data statistical analysis, scientific forecasting is done for through street traffic safety；Meanwhile it utilizing
The data of record can punish vehicle driver violating the regulations, improve the cost violating the regulations of driver.
The present invention realizes that through street scene is supervised using the wireless transmission between the first wireless bridge and the second wireless bridge
High clear video image, high fidelity speech and the transmission for controlling data between unit and backstage Supervisory Unit are surveyed, without being laid with electricity
Cable or optical fiber, difficulty of construction is small, easy to maintain.
The present invention does global video monitor using panoramic high-definition video camera, and is divided to monitoring area in monitoring host computer
Sub-districts are monitored for 254, and each monitoring sub-district is numbered, and make the preset position of monitoring sub-district number and high definition ball machine
Number is consistent, is used as the positional number of the motor vehicle further according to the monitoring sub-district number being abnormal at the motor vehicle of behavior
According to the preset position that the position data is converted to high definition ball machine is numbered, and so that high definition ball machine is gone to from initial preset position motor-driven
Position where vehicle；It is abnormal to the motor vehicle of various road types (such as "-" type road, tee T road, cross road) to realize
Behavior carries out the accurately candid photographs such as detection and image, data, and using flexible, monitoring is accurately.
Be to the algorithm of image procossing and analysis in the present invention it is existing, it is specific as follows：
1, laplacian spectral radius algorithm
Laplacian spectral radius algorithm refers to being sharpened processing to image using Laplace operator, and essence is that second order is micro-
Partite transport is calculated, and can protrude the details that details or enhancing in image are blurred, common Laplace operator has：
Diagonal field operator, that is, operator 2 is used to make motor vehicle in video background to handle image in the present invention
Edge in image is sharpened, and motor vehicle profile becomes more fully apparent, and improves the Detection accuracy of motor vehicle.
2, histogram enhancement defogging algorithm
" histogram enhancement defogging " algorithm utilizes adaptive histogram based on the big feature of haze sky image scene change in depth
Figure equalization method carries out local detail information enhancement to haze sky image, and the ratio degree by enhancing image eliminates haze weather pair
The influence of image definition.
It is practical that panorama is taken the photograph using the variation for reflecting local scene contrast using partial histogram equalization in the present invention
Camera shooting picture is divided into sky areas, pavement of road region and roadside region, each square to the pavement of road region of monitoring
To 4 × 4 fritters are divided into, i.e., each road direction has 16 histograms, the side of the boundary linear interpolation of 16 fritters
Method realizes that smooth transition, other regions are not processed so that motor vehicle on pavement of road can also be clearly in haze sky
It is apparent in picture, since picture pixel to be treated greatly reduces, improves the real-time of algorithm work.
3, sliding average background model algorithm
Sliding average background model algorithm is the one kind for carrying out real-time update to background image in moving-target detection process
Adaptive algorithm reduces influence when moving object is passed through to background as far as possible using the background model being shown below：
BackgroudIMGn(i, j)=BackgroundIMGn-1α in (i, j) * (1- α)+InputIMG (i, j) * α above formulas
For context update rate, when α values are larger, context update is very fast, and when α values are smaller, context update is slower.
In the present invention, background image is divided into different zones by us, and each region is assigned to different α values：There to be fortune
The region alpha value that animal body passes through obtains smaller so that background image is not easy to change since moving object is passed through；There to be moving object
Body by move switch to static or stationary object by it is static switch to movement region α values obtain it is larger, the former to stop
Moving object can rapidly go to background, the latter allows the object of setting in motion to be removed from background rapidly.Not same district
After domain is assigned to different α values so that background image can adjust renewal rate according to image actual conditions subregion, obtain more preferable
Background extracting effect.
By adaptive updates background modeling, and by systems such as threshold process, Morphological scale-space and foreground image fusions
Row post-processing, can obtain more clearly, the less moving object image of noise, be used for pictures subsequent processing.
4, mixture Gaussian background model algorithm
Mixture Gaussian background model algorithm goes to accurately reflect motor vehicle and other moving-targets using two groups of multistage Gauss models
Difference reduce False Rate to improve motor vehicle detecting accuracy rate.In the monitoring area of panoramic camera, in addition to motor vehicle,
Further include having other similar to moving-target as vehicle using motor, pedestrian.
Mixture Gaussian background model algorithm in the present invention is to be divided into two stages when detecting motor vehicle, including learn rank
Section and detection-phase.In the study stage, for each pixel of associated picture, establish two groups of multistage Gauss models go to reflect it is motor-driven
The shape and distribution of movement of vehicle obtain classification thresholds according to " minimum classification error principle "；In detection-phase, in detection motor vehicle
In the region of covering, a cascade grader is for sorting out each pixel of motor vehicle, and then, secondary use is " more
Number rule " provides final testing result.I.e. one for fusion detection target coverage, all pixel classifications as a result, another
One testing result for merging multiframe.
5, Blob Tracking algorithms
Blob Tracking algorithms are to identify the detection algorithm of target using not bending moment, graphics area and perimeter ratio,
We obtain the boundary information of Blob using the Blob line processing methods for meeting requirement of real-time, and Blob is demarcated and gathered
It closes, and then detects the center of Blob.
In the present invention, after extracting motor vehicle in background image, Blob analyses is carried out to its bianry image, are obtained motor-driven
Vehicle profile parameters and location information obtain speed, the traveling side of motor vehicle to motor vehicle by Kalman filter into line trace
To multidate informations such as, driving traces.This method can when night, illumination condition was bad detect and track motor vehicle well.
6, Kalman filter algorithm
Kalman filter algorithm is the algorithm that linear minimum-variance estimation is carried out to the status switch of dynamical system, using passing
Return filtering method.The present invention carries out interframe sliding window filtering using Kalman filter algorithm to obtained vehicle position coordinate, i.e.,
The frame test position of motor vehicle is equal to the average value of its front and back frame test position in time series, and deviation is less than its location estimation
It is worth 2 times of error.This method can eliminate influence of the outside noise to motor vehicle detecting position, accurately detect the position of motor vehicle
Coordinate is set, accurate presetting bit information is provided for the scanning candid photograph of high definition ball machine, whether there is for motor vehicle and stop, go slowly, is inverse
The behavioural analyses such as row, reversing provide reliable position coordinate value.
Description of the drawings
Fig. 1 is principle schematic diagram of the present invention；
Fig. 2 is field monitoring unit structural schematic diagram；
Fig. 3 is monitoring method flow chart of the present invention.
Specific implementation mode
As shown in Figure 1, 2, monitoring device includes field monitoring unit 1 and backstage Supervisory Unit 2, the field monitoring unit
1 includes the panoramic high-definition video camera 101 being mounted on the cross bar 4 on 3 top of through street roadside upright bar, high definition ball machine 102
With high-pitched speaker 104, cabinet 5 and the first wireless bridge 106 are installed in the upright bar 3, it is wide to be equipped with IP network in the cabinet 5
Broadcast power amplifier 103 and first network interchanger 105；First wireless bridge 106, panoramic high-definition video camera 101, high definition ball machine
102 and IP network broadcast power amplifier 103 respectively with the first network interchanger 105 is interactive connect, the IP network is broadcasted
The output end of power amplifier 103 is connect with high-pitched speaker 104；
Backstage Supervisory Unit 2 includes monitoring host computer 203, ball machine monitoring station 202, panoramic camera monitoring station 201, the second nothing
Gauze bridge 206, second network switch 205, server 204 and disk array 207；Second network switch 205 is distinguished
With monitoring host computer 203, ball machine monitoring station 202, panoramic camera monitoring station 201, the second wireless bridge 206, server 204 and magnetic
The interactive connection of disk array 207；
Between field monitoring unit 1 and backstage Supervisory Unit 2 by the first wireless bridge 106, the second wireless bridge 206 into
Row wireless data transmission realizes high clear video image, high fidelity speech and the control between field monitoring unit and backstage Supervisory Unit
The mutual transmission of data processed.
This monitoring device is by the live 360 degree of full-view video images of panoramic high-definition camera acquisition through street, by the
One network switch and the first wireless bridge are transmitted in panoramic camera monitoring station and the monitoring host computer in the Supervisory Unit of backstage,
Monitoring host computer reads in the video image stream of panoramic high-definition video camera, and is single-frame images according to H.264 protocol-decoding, then to figure
As carrying out eliminating fuzzy and background refresh process, and detect that the motor vehicle in image is detected and tracks；Simultaneously to image
In motor-driven vehicle going behavior by Kalman filter algorithm carry out analysis motor vehicle stop, drive in the wrong direction, move backward, go slowly etc. it is abnormal
Behavior；Monitoring sub-district number at motor vehicle according to abnormal behaviour is used as the position data and high definition ball of the motor vehicle
The preset position of machine is numbered, and scan control order and preset bit number are sent to ball machine monitoring station, ball machine monitoring by monitoring host computer
Position where station drives high definition ball machine to go to motor vehicle from initial preset position is amplified high clear video image to motor vehicle
Camera shooting, will obtain clearly abnormal motor vehicle and its license plate image by sending ball machine monitoring station to；Monitoring master
Automobile license plate number in image is identified in machine, and the automobile license plate number that will identify that and its abnormal behaviour are combined into one
Sentence caveat, and it is simultaneously emitted by sound and light alarm signal, prompt operator on duty to pay attention to；Caveat is broadcasted by IP network quick again
Road scene plays out, while high definition ball machine is captured image, time, position data, car plate in relation to the exception motor vehicle
Number, abnormal behaviour class record are stored in server or disk array.
As shown in figure 3, a kind of the step of motor vehicle abnormal behaviour monitoring method on through street, is as follows：
(1) monitoring region is divided into 254 monitoring sub-districts, and each monitoring sub-district is numbered, and keep monitoring sub
Number consistent with the preset position of high definition ball machine number in area；
(2) monitoring host computer reads in the video image stream of panoramic high-definition video camera, and is single frames figure according to H.264 protocol-decoding
Picture；
(3) single-frame images in (2) image caused by DE Camera Shake is eliminated using laplacian spectral radius algorithm to obscure, with
And using histogram enhancement defogging algorithm eliminate the image caused by misty rain thick weather it is fuzzy eliminate it is fuzzy；Again using cunning
Dynamic average background model algorithm refreshes image background, to eliminate image light and shade, brightness change and the influence brought；
(4) motor vehicle in image is detected using mixture Gaussian background model algorithm；And use Blob
Tracking algorithms are to the motor vehicle that detects into line trace；
(5) the motor-driven vehicle going behavior in image is analyzed using Kalman filter algorithm；If motor vehicle is at certain
One position remains stationary state is more than 5 seconds, then it is assumed that it belongs to abnormal " STOP " behavior；If motor vehicle motion direction and road
Provide that travel direction is inconsistent, then it is assumed that it belongs to abnormal " retrograde " behavior；If the direction of motion of motor vehicle changes,
Think that it belongs to abnormal " reversing " behavior；If the travel speed of motor vehicle is less than 10 kilometers/hour, then it is assumed that it belongs to abnormal
" going slowly " behavior；
(6) it is used as the position of the motor vehicle according to the above-mentioned monitoring sub-district number being abnormal at the motor vehicle of behavior
Data are set, and the position data is converted to the preset position number of high definition ball machine, monitoring host computer is according to vehicle position data
Scan control order and preset bit number are sent to ball machine monitoring station, high definition ball machine is made to go to motor vehicle from initial preset position
The position at place takes 2-3.5 seconds；The motor vehicle detected is completely in some monitoring subregion, then monitoring
The number in region is the position data of motor vehicle；
If the motor vehicle detected is in more than one monitoring sub-district, with a distance from panoramic high-definition video camera most
Position data of the number of close monitoring sub-district as motor vehicle；
(7) postpone to regard from the high definition of high definition ball machine at least 4.5 seconds when rotating initial preset position since high definition ball machine
A frame image is captured in frequency image stream, it is ensured that ball machine camera lens has stablized the position for being directed toward abnormal motor vehicle, and guarantee obtains a height
Image clearly in relation to abnormal motor vehicle；
(8) automobile license plate number in image is identified in monitoring host computer, and the automobile license plate number that will identify that and
Its abnormal behaviour is combined into a caveat, and is simultaneously emitted by sound and light alarm signal, and operator on duty is prompted to pay attention to；Caveat is led to again
IP network broadcast is crossed to play out at through street scene, at the same by high definition ball machine capture image in relation to the exception motor vehicle,
Time, position data, the number-plate number, abnormal behaviour type are stored；If the motor vehicle number-plate number cannot be identified, send out
Sound and light alarm signal.Monitoring host computer by second network switch by the photo of the abnormal motor vehicle of crawl, crawl the time, crawl
The data such as position, motor vehicle trade mark, abnormal behaviour type are recorded in server or disk array, are returned again to and are detected again.
Wherein caveat is made of " automobile license plate number " and " abnormal behaviour type ", the stress of sentence is placed on motor-driven
In the action that vehicle license plate number, abnormal behaviour type and needs are done, a fixed warning word is synthesized a with strong points, language
The clear naturally deterrence language of meaning.The motor vehicle trade mark such as identified is " Anhui A ZW302 ", and abnormal behaviour type is " STOP ", then
The caveat of synthesis is that " Anhui A ZW302, Anhui A ZW302, your present parking offense, parking offense please sail out of as early as possible！It sails as early as possible
From！”；The motor vehicle trade mark such as identified is " Guangdong C 77506 ", and abnormal behaviour type is " reversing ", then the caveat synthesized is
" Guangdong C 77506, Guangdong C 77506, you break rules and regulations to move backward now, and reversing of breaking rules and regulations please is sailed out of as early as possible！It sails out of as early as possible！”.
Above example is not limited only to protection scope of the present invention, all to be repaiied based on basic thoughts of the invention
Change or what is changed belongs to protection scope of the present invention.
Claims (9)
1. the motor vehicle abnormal behaviour monitoring method on a kind of through street, it is characterised in that：Include the following steps：
（1）Monitoring region is divided into 254 monitoring sub-districts, and each monitoring sub-district is numbered, monitoring sub-district is made to number
It is consistent with the preset position number of high definition ball machine；
（2）Panoramic camera monitoring station handles 360 degree of full-view video images that panoramic camera is passed back, spherical video
Image spread is 90 degree of four split screen video image, and two screen video images of corresponding through street monitoring are transmitted to monitoring host computer
It is decoded as single-frame images；
（3）It will（2）Middle single-frame images carries out eliminating fuzzy, background refresh process；
（4）Motor vehicle in image after processing is detected, and to it into line trace；
（5）Motor-driven vehicle going behavior in image is analyzed；If motor vehicle is more than 5 in a certain position remains stationary state
Second, then it is assumed that it belongs to abnormal " STOP " behavior；If motor vehicle motion direction and road regulation travel direction are inconsistent, recognize
Belong to abnormal " retrograde " behavior for it；If the direction of motion of motor vehicle changes, then it is assumed that it belongs to abnormal " reversing " row
For；If the travel speed of motor vehicle is less than 10 kilometers/hour, then it is assumed that it belongs to abnormal " going slowly " behavior；
（6）It is used as the position data of the motor vehicle according to monitoring sub-district number at the motor vehicle of above-mentioned abnormal behaviour, and
The preset position number that the position data is converted to high definition ball machine, makes high definition ball machine go to motor vehicle institute in place from initial position
It sets；
（7）A frame image is captured from the high clear video image stream of high definition ball machine, you can the exception for obtaining a high-resolution is motor-driven
Vehicle image；
（8）Automobile license plate number in image is identified, the automobile license plate number that will identify that and its abnormal behaviour combine
At a caveat, and it is simultaneously emitted by sound and light alarm signal；The caveat is broadcasted by IP network at through street scene again
It plays out, while high definition ball machine is captured into image in relation to the exception motor vehicle, time, position data, the number-plate number, exception
Behavior type is stored；If the motor vehicle number-plate number cannot be identified, sound and light alarm signal is sent out.
2. monitoring method according to claim 1, it is characterised in that：The step（3）Middle elimination Fuzzy Processing is to use
Image is fuzzy caused by laplacian spectral radius algorithm eliminates DE Camera Shake, and using histogram enhancement defogging algorithm eliminate because
Image caused by misty rain thick weather is fuzzy；Background refresh process be using sliding average background model algorithm to image background into
Row refreshes, to eliminate image light and shade, brightness change and the influence brought.
3. monitoring method according to claim 1, it is characterised in that：The step（4）In motor vehicle is detected, be
Motor vehicle in image is detected using mixture Gaussian background model algorithm；To motor vehicle into line trace be use Blob
Tracking algorithms are to the motor vehicle that detects into line trace.
4. monitoring method according to claim 1, it is characterised in that：The step（5）In to the motor line in image
The behavior of sailing, which carries out analysis, to be analyzed motor-driven vehicle going behavior using Kalman filter algorithm.
5. monitoring method according to claim 1, it is characterised in that：The step（6）Middle high definition ball machine is from initial preset
The time that position goes to the position where motor vehicle is 2-3.5 seconds.
6. monitoring method according to claim 1, it is characterised in that：The step（7）In regarded from the high definition of high definition ball machine
It is when being rotated initial preset position since high definition ball machine that a frame image is captured in frequency image stream as abnormal motor vehicle image
Delay is captured at least 4.5 seconds again.
7. a kind of realizing the monitoring device such as any one of the claims 1-6 monitoring method, it is characterised in that：It is supervised including scene
It includes the cross for the rod end for being mounted on through street roadside to survey unit and backstage Supervisory Unit, the field monitoring unit
Panoramic high-definition video camera, high definition ball machine on bar and high-pitched speaker are equipped with cabinet and the first wireless bridge, institute in the upright bar
It states and is equipped with IP network broadcast power amplifier and first network interchanger in cabinet；First wireless bridge, panoramic high-definition video camera,
High definition ball machine and IP network broadcast power amplifier are connect with the first network interchanger interactive mode respectively, and the IP network broadcasts work(
The output end for putting machine is connect with the high-pitched speaker；
The backstage Supervisory Unit includes monitoring host computer, and the monitoring host computer passes through second network switch and the second wireless bridge
Interactive mode connection；
It is carried out without line number by the first wireless bridge, the second wireless bridge between the field monitoring unit and backstage Supervisory Unit
According to transmission.
8. monitoring device according to claim 7, it is characterised in that：The backstage Supervisory Unit further includes respectively with second
The ball machine monitoring station and panoramic camera monitoring station, the ball machine monitoring station of network switch interactive mode connection pass sequentially through second
The network switch, the second wireless bridge, the first wireless bridge and first network interchanger drive high definition ball machine；The panorama is taken the photograph
Camera monitoring station, which passes sequentially through second network switch, the second wireless bridge, the first wireless bridge and first network interchanger, to be come
Drive panoramic camera work.
9. monitoring device according to claim 7, it is characterised in that：Second network switch interactive mode connection is useful
In the server and disk array of storage.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410810367.XA CN104506804B (en) | 2014-12-22 | 2014-12-22 | Motor vehicle abnormal behaviour monitoring device and its method on a kind of through street |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410810367.XA CN104506804B (en) | 2014-12-22 | 2014-12-22 | Motor vehicle abnormal behaviour monitoring device and its method on a kind of through street |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104506804A CN104506804A (en) | 2015-04-08 |
| CN104506804B true CN104506804B (en) | 2018-08-10 |
Family
ID=52948518
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410810367.XA Active CN104506804B (en) | 2014-12-22 | 2014-12-22 | Motor vehicle abnormal behaviour monitoring device and its method on a kind of through street |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104506804B (en) |
Families Citing this family (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106355900A (en) * | 2015-07-23 | 2017-01-25 | 环达电脑(上海)有限公司 | System for actively transmitting dangerous driving alert |
| CN105185109A (en) * | 2015-08-06 | 2015-12-23 | 苏州市世跃智能科技有限公司 | Double-dome-camera-based automatic detection and snapshot apparatus for illegal parking |
| CN106485730A (en) * | 2015-08-31 | 2017-03-08 | 上海悠络客电子科技有限公司 | Method for tracking target in video |
| CN105957312A (en) * | 2016-06-06 | 2016-09-21 | 重庆微标科技股份有限公司 | Prompting method and system for abnormal clearance of vehicle |
| CN106530746A (en) * | 2016-10-28 | 2017-03-22 | 江苏信息职业技术学院 | Vehicle speed detection apparatus and detection method based on video technology |
| CN106791645A (en) * | 2016-12-20 | 2017-05-31 | 长春理工大学 | The monitor video system that a kind of abnormal behaviour is captured |
| CN107180533A (en) * | 2017-06-22 | 2017-09-19 | 北京中交华安科技有限公司 | Road method for early warning and device |
| CN107507424B (en) * | 2017-07-05 | 2019-12-06 | 湖北鑫美企业发展股份有限公司 | Device and method for automatically identifying driving information during vehicle violation |
| CN107704841A (en) * | 2017-10-25 | 2018-02-16 | 云南电网有限责任公司电力科学研究院 | A kind of image/video abnormal state detection identifies system |
| CN108335489A (en) * | 2018-03-11 | 2018-07-27 | 西安电子科技大学 | Vehicle on highway behavior semantic analysis and abnormal behaviour monitoring system and method |
| CN108520212A (en) * | 2018-03-27 | 2018-09-11 | 东华大学 | Method for traffic sign detection based on improved convolutional neural networks |
| CN109191862A (en) * | 2018-10-23 | 2019-01-11 | 长沙荣茂信息技术工程有限公司 | A kind of traffic offence broadcasts equipment and its data processing method automatically |
| CN110222856B (en) * | 2019-06-12 | 2022-03-25 | 中国神华能源股份有限公司 | Method and device for processing damage to train wheel tread and storage medium |
| CN112150795B (en) * | 2019-06-26 | 2022-05-06 | 杭州海康威视数字技术股份有限公司 | Method and device for detecting vehicle track abnormity |
| CN110310493B (en) * | 2019-07-02 | 2022-01-07 | 亳州职业技术学院 | Traffic retrograde constraint system based on Internet of things |
| CN110782677A (en) * | 2019-11-25 | 2020-02-11 | 湖南车路协同智能科技有限公司 | Illegal vehicle snapshot warning method and device |
| CN110933318A (en) * | 2019-12-12 | 2020-03-27 | 天地伟业技术有限公司 | Moving object snapshot method |
| CN111314657B (en) * | 2019-12-31 | 2021-05-14 | 安徽清新互联信息科技有限公司 | Image acquisition method and system based on cluster camera equipment |
| CN112449111A (en) * | 2020-11-13 | 2021-03-05 | 珠海大横琴科技发展有限公司 | Monitoring equipment processing method and device |
| CN113822902A (en) * | 2021-09-24 | 2021-12-21 | 泰州市雷信农机电制造有限公司 | Block chain vehicle body state monitoring system |
| CN115824318B (en) * | 2023-02-27 | 2023-08-15 | 国网江西省电力有限公司电力科学研究院 | Dynamic monitoring system and method for digital control of working state of pole erecting machine |
| CN116752471B (en) * | 2023-08-23 | 2023-10-27 | 湖北工业大学 | Direction-changeable blocking ground lock and warning guide system for tidal lane |
| CN118280124A (en) * | 2024-06-04 | 2024-07-02 | 浙江大华技术股份有限公司 | Traffic abnormality detection method, apparatus, and storage medium |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10011247B2 (en) * | 1996-03-27 | 2018-07-03 | Gtj Ventures, Llc | Control, monitoring and/or security apparatus and method |
| CN1725208A (en) * | 2004-07-19 | 2006-01-25 | 上海市市政工程管理处 | Taffic information processing system for urban through street |
| JP4984915B2 (en) * | 2006-03-27 | 2012-07-25 | セイコーエプソン株式会社 | Imaging apparatus, imaging system, and imaging method |
| CN101551940A (en) * | 2009-04-29 | 2009-10-07 | 上海电器科学研究所（集团）有限公司 | Urban high-speed road traffic state judging and issuing system and method thereof |
| CN202584422U (en) * | 2012-03-23 | 2012-12-05 | 罗普特（厦门）科技集团有限公司 | Apparatus for snapshotting vehicles violating traffic rules based on intelligent analysis and video monitoring |
| CN102665071B (en) * | 2012-05-14 | 2014-04-09 | 安徽三联交通应用技术股份有限公司 | Intelligent processing and search method for social security video monitoring images |
| CN103605956A (en) * | 2013-11-04 | 2014-02-26 | 魏文红 | Emergency intelligence identification system based on video abstraction |
| CN103929592A (en) * | 2014-04-22 | 2014-07-16 | 杭州道联电子技术有限公司 | All-dimensional intelligent monitoring equipment and method |
| CN104123815B (en) * | 2014-08-08 | 2016-09-14 | 浙江高速信息工程技术有限公司 | A kind of highway quick disposal system of self-starting emergency management and rescue |
-
2014
- 2014-12-22 CN CN201410810367.XA patent/CN104506804B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN104506804A (en) | 2015-04-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104506804B (en) | Motor vehicle abnormal behaviour monitoring device and its method on a kind of through street | |
| CN107705574A (en) | A kind of precisely full-automatic capturing system of quick road violation parking | |
| WO2020122986A1 (en) | Driver attention detection using heat maps | |
| CN109191829B (en) | road safety monitoring method and system, and computer readable storage medium | |
| KR100835085B1 (en) | A real time processing system and method of a traffic accident using an unmanned recording device | |
| CN102081844A (en) | Traffic video behavior analyzing and alarming server | |
| CN101964145A (en) | Automatic license plate recognition method and system | |
| CN113593250A (en) | Illegal parking detection system based on visual identification | |
| CN114898297B (en) | Target detection and target tracking-based non-motor vehicle illegal behavior judgment method | |
| KR102353724B1 (en) | Apparatus and method for monitoring city condition | |
| CN114781479A (en) | Traffic incident detection method and device | |
| Odeh | Management of an intelligent traffic light system by using genetic algorithm | |
| JP7305965B2 (en) | Video surveillance system parameter setting method, device and video surveillance system | |
| CN107277301A (en) | The image analysis method and its system of monitor video | |
| CN104766480A (en) | Crossroad red light running capturing device | |
| CN107316463A (en) | A kind of method and apparatus of vehicle monitoring | |
| CN204291236U (en) | Motor vehicle abnormal behaviour monitoring device on a kind of through street | |
| CN111311942A (en) | Road condition display method and system based on V2X technology, V2X terminal and V2X server | |
| JP2020518165A (en) | Platform for managing and validating content such as video images, pictures, etc. generated by different devices | |
| CN106205135A (en) | A kind of detection method of vehicle behavior that turns around violating the regulations, Apparatus and system and a kind of ball machine | |
| CN117152971A (en) | AI traffic signal optimization method based on high-altitude panoramic video | |
| WO2022179124A1 (en) | Image restoration method and apparatus | |
| CN108280898A (en) | It is a kind of can shared information automobile data recorder | |
| CN113112813B (en) | Illegal parking detection method and device | |
| CN104599510A (en) | Dual-camera velocimeter |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |