CN101751782A - Crossroad traffic event automatic detection system based on multi-source information fusion - Google Patents
Crossroad traffic event automatic detection system based on multi-source information fusion Download PDFInfo
- Publication number
- CN101751782A CN101751782A CN200910238815A CN200910238815A CN101751782A CN 101751782 A CN101751782 A CN 101751782A CN 200910238815 A CN200910238815 A CN 200910238815A CN 200910238815 A CN200910238815 A CN 200910238815A CN 101751782 A CN101751782 A CN 101751782A
- Authority
- CN
- China
- Prior art keywords
- traffic
- information
- video
- audio
- crossroad
- 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
Images
Landscapes
- Traffic Control Systems (AREA)
Abstract
The invention discloses a crossroad traffic event automatic detection system based on multi-source information fusion. The invention aims at urban crossroad traffic management, establishes audio-video traffic information acquisition, traffic information analytical processing and traffic event automatic alarming three subsystems. The invention mainly comprises: designing and realizing an acquisition subsystem of multi-source traffic information (traffic audio information, traffic video information, multi-angle video information, elevation angle video information); realizing a traffic parameter automatic extractive technique based on traffic audio and traffic video; realizing a traffic event (car crash, illegal parking, jam and retrograde motion) automatic detection technique based on the multi-source information fusion; and realizing a software system for traffic event report, traffic event associated video information transmission, and traffic control center display and processing, and forming an integrated crossroad traffic event automatic detection system.
Description
Technical field
The present invention relates to the intelligent transportation system field, specifically, the present invention is that traffic audio-video collection, transport information Treatment Analysis and the traffic events traffic event automatic detection system of warning three big modules has automatically been managed, integrated to a cover at the city crossroad access, mainly comprises: the acquisition subsystem of having designed and Implemented a kind of multi-source traffic information; Realized based on the automatic extractive technique of the traffic parameter of traffic audio frequency and traffic video; Realized traffic events Automatic Measurement Technique based on Multi-source Information Fusion; Realized that traffic events report, traffic events associated video information transmission, traffic control center show and the process software system, have formed the complete crossroad traffic event automatic detection system based on Multi-source Information Fusion.
Background technology
Intelligent transportation system (Intelligent Transportation System, ITS) be that people are in order to realize good life, obtain desirable traffic environment and of proposing to reality and the far-reaching major action of future world, it is the integrated effectively whole traffic management system that applies to such as advanced person's infotech, data communication transmission technology, electronic sensor technology, electron controls technology and Computer Processing technology etc., can be on a large scale interior, real-time, accurate, comprehensive playing a role.Specifically, intelligent transportation system is carried out real-time supervision to the traffic in the roadnet, traffic hazard, meteorological condition and traffic environment, rely on advanced vehicle detection technology and Computerized Information Processing Tech, obtain the information of relevant traffic, and according to the information of collecting traffic is controlled, as signal lamp, issue induction information, road control, accident treatment and rescue etc.
Along with the city dweller has the rapid increase of automobile quantity, urban traffic blocking becomes the world wide problem to be solved of having to go to the toilet.Traffic data shows that in urban transportation, traffic hazard, vehicle cast anchor and often cause secondary traffic hazard and very serious traffic congestion.In more than ten years in past; international academic community detects (Automatic Incident Detection automatically at traffic events; AID) development activities of technical field is very active; but it is to detect and carry out at the traffic events on highway or the national highway that document shows most technical research, then relatively lags behind at the traffic events detection technique research of city road crossing.From present technical merit, with verification and measurement ratio, rate of false alarm and detect average service time these three performance index and weigh, existing crossroad AID technical feature still can not adapt to the traffic monitoring requirement of crossroad, city, carries out at the research and development of the AID technology of crossroad, city all significant to timely discovery traffic events, the aspects such as blocking up, optimize the crossroad design that relieves traffic congestion.
At the crossroad, city, the intelligent transportation system with the automatic measuring ability of traffic events mainly is made up of three parts: crossroad access information acquisition system, information process analysis system, information show and delivery system.Now the background technology with three subsystems is summarized as follows.
1, crossroad access information acquisition system background technology
Traffic information acquisition system is the front end of intelligent transportation system, mainly finish the functions such as collection, transmission, storage of transport information, the means that adopt comprise at present: artificial input, GPS vehicle mounted guidance instrument, GPS navigation mobile phone, vehicle pass-through electronic information card, CCTV video camera, infrared radar detecting device, coil checker, optical detector etc.Through statistics, in comprising the intelligent transportation system that traffic events detects (AID) technology automatically, obtain broad research and application be based on ground induction coil and based on the system of computer vision technique.Based on the technology of ground induction coil, when being installation and maintenance, its major defect must block traffic, excavate the road surface, pre-buried ground induction coil.This technology can only obtain limited traffic parameters such as vehicle flowrate, the speed of a motor vehicle, vehicle commander in addition.By contrast, the AID technology of handling based on video image can obtain more telecommunication flow information, as parameters such as vehicle flowrate, traffic density, the speed of a motor vehicle, vehicle classification, spacing, lane occupancy ratio, queue lengths; At the parting of the ways, can also make a dash across the red light to vehicle, acts of violating regulations such as vehicle drives in the wrong direction, overspeed of vehicle detect and analyze (comprise starting and take pictures and car plate identification) in real time, mishap detects, writes down and reports to crossing traffic, and utilizes the video information of gathering to satisfy the need the oral sex interpreter so carry out forecast analysis etc.This technology biggest advantage is intuitively to observe, and installs simply, and expense is lower, does not need to destroy road during installation and maintenance and does not influence road traffic.Yet folk prescription is subjected to the influence of factors such as vehicle blocks, shade, Changes in weather to, low-angle video through regular meeting, and it detects effect and incident and judges and be affected.Therefore, assist, rationally arrange the layout of crossroad audio ﹠ video equipment, will greatly enrich the collection of transport information, for follow-up traffic events monitoring provides audio/video information reliable, comprehensive, robust by audio system.
2, information process analysis system background technology
The information process analysis system analyzes and handles the transport information that collects, and can judge traffic exactly at last, for information shows and delivery system provides information.Visual monitor system (Visual SurveillanceSystem-VSS) based on Flame Image Process has played very big help for traffic events and traffic congestion in monitoring and minimizing highway and the urban transportation.Therefore, attention and broad research have been subjected to based on the automatic detection algorithm of the traffic events of video.The anomalous event detection system based on image processing techniques of Japan Ikeda team research and development can detect stopped vehicle, slow-moving vehicle, dropped object, the four kinds of unusual traffic events in vehicle continuous transformation track.Its algorithm is extracted by vehicle and the anomalous event detection algorithm is formed, and mainly is background elimination algorithm and the temporal difference method that adopts based on video.The Kimachi team of Japan proposes a kind of event detecting method based on image sequence analysis and fuzzy theory, from image sequence, obtain to cause the vehicle abnormality behavior of traffic events, traveling state of vehicle is estimated, detected vehicle collision, the traffic events of types such as dropped object.Up to the present, carry out at mainly having in addition that city crossroad traffic event Automatic Measurement Technique is studied: the improvement probe algorithm that has proposed a kind of Markov random data model based on space-time with the Tokyo University team headed by the Kamijo, the method of the state of this algorithm by determining each pixel in each width of cloth image frame is set up the model of detection problem, determines simultaneously how pixel carries out state transitions along X-Y image coordinate and time coordinate.This system can discern vehicle collision, pass through and congested three kinds of traffic events; The Michalopoulos of Univ Minnesota-Twin Cities USA has proposed the automatic detection algorithm of traffic events; the video image processing unit of this algorithm need carry out analyzing and processing to whole video stream, detects stationary vehicle, the vehicle that travels very slowly, drive in the wrong direction vehicle and traffic hazard.The sharpest edges of this algorithm are the space time informations that has utilized traffic flow, and the more transport information of horn of plenty can be provided.The Veeraraghavan of Univ Minnesota-Twin Cities USA etc. detects problem automatically at the crossroad incident, propose to adopt single camera to make up supervisory system, utilization is determined Moving Target based on the modules at lower layers coupling of image with based on the high-rise Kalman filtering of direction of traffic bounding box, monitor the traffic at crossing, thereby detect traffic hazard.Its core technology is that background model is set up, target extracts and vehicle tracking technology, vehicle movement parameter acquisition algorithm and based on the incident Detection Algorithm of pattern match and state estimation, can realize understanding and description to the vehicle behavior.Kamijo has proposed the automatic detection algorithm of a kind of crossroad traffic event based on video and hidden Markov, mainly vehicle collision, the beginning three kinds of traffic events such as knock into the back, overtake other vehicles that stop is detected.Employing video techniques such as the Ki of Korea S have researched and developed that complicated crossroad traffic event detects automatically, record and reporting system.The actual test result of this system shows that the correct verification and measurement ratio of its traffic events can arrive 50%, and the traffic events verification and measurement ratio can reach 60%.In addition, problem detects automatically at U.S.'s interstate highway traffic events in Michalopoulos team, according to the video traffic information that the AUTOSCOPE system acquisition arrives, has researched and developed the traffic event automatic detection system based on video image processing technology.This system is made of 39 video cameras, and 7 miles of installation courses adopt successive detection mode, can reach 80% correct verification and measurement ratio; The incident Detection Algorithm based on video image processing technology that adopts France Blosseville has successfully realized the accurate detection to traffic events between highway; The Trivedi of California, USA university etc., event detection at highway, proposed that a kind of new being used to detects and the distributed video collection network framework incident that regulates the traffic, that adopt orthoscopic and omnidirectional vision camera, improve the vehicle tracking precision with this, and proposed the traffic data that adopts data base administration to collect.Summary is got up, and the research that domestic and international research team carries out in the traffic events Automatic Measurement Technique field that relates to the crossroad is also very not enough, still is being in the starting stage aspect the traffic events Automatic Measurement Technique of Audiotechnica.The research document shows both at home and abroad, crossroad real-time traffic incident Automatic Measurement Technique has important application background and value, energetically the crossroad AID system of research and development with the correct verification and measurement ratio of high traffic events, low rate of false alarm will help to strengthen to the crossroad, city monitoring, induce traffic, alleviate traffic congestion, analyze the traffic hazard occurrence cause, the crossroad that helps reducing traffic hazard for design provides support.
3, information shows and the delivery system background technology
Information shows and delivery system mainly adopts modes such as internet, mobile phone, car-mounted terminal, broadcasting, electronic intelligence plate, telephone service platform, when traffic events takes place, traffic control center need obtain the demonstration and the processing capacity of transport information, its related intelligent transportation system can adopt certain or multiple mode to announce transport information, dredges or rescues.Information display mode based on map is present international popular and advanced information display strategy, delivery system is that the intelligent transportation system decision-making is implemented later on and auxiliary module, and the present invention directly utilizes existing city map, existing module to carry out secondary development, internet (comprising webpage and mail notification), SMS, three kinds of information published methods of telephone service have been comprised, belong to the combination of existing means, not as emphasis of the present invention.
Summary of the invention
This hairpin detects problem automatically to the city crossroad traffic event, has developed the traffic event automatic detection system that merges based on audio frequency and video, and invention mainly comprises four parts: the acquisition subsystem of multi-source (audio frequency and video, height video) transport information; Traffic parameter based on traffic audio frequency and traffic video extracts subsystem automatically; Based on the automatic detection subsystem of the traffic events of Multi-source Information Fusion; Traffic events report, traffic events associated video information transmission, traffic control center show and the process software subsystem; Formed complete city crossroad access information acquisition and the traffic events automatic transport detection system of a cover.
1, based on the crossroad access real time information sampling and the processing hardware system architecture of Multi-source Information Fusion
On the basis of having analysed in depth typical urban road crossroad characteristics, the present invention has adopted space layout principle and the strategy of realizing crossroad audio frequency and video acquisition of road traffic information based on the audio-video collection systematic collaboration of omnidirectional video image collection, multiple-camera and linear microphone array, as shown in Figure 1.Be specially: low angle four road video acquisitions, direction are for going out the car direction, and height is contour with the traffic lights flag-rod, and main monitoring range is for going out the car direction forward till the crossroad; High-order one road video acquisition, be positioned on turning, crossroad skyscraper or the high bar, highly be 30-100 rice adjustable (adjusting) with the crossroad size, depression angle becomes the 45-90 degree adjustable with the local horizon, the main satisfied whole crossroad area-of-interest that covers is as the criterion, and monitors the traffic lights queuing situation of certain-length simultaneously; The audio frequency array is positioned at car direction right side, and is adjustable to vehicle queue direction 0-30 rice at vehicle outlet, and audio frequency array height 0.5-2 rice is adjustable, and the microphone array column direction becomes with the local horizon-and the pitching of 45-+45 degree is adjustable.
For realizing functions such as the real-time conduct monitoring at all levels of the multidirectional road traffic in crossroad, traffic data storage, transmission and demonstration, the automatic detection of traffic events, the present invention has designed and Implemented the hardware system of crossroad audio/video information collection and disposal system.Wherein video acquisition system has adopted the solution of video camera Network Based, the information processing platform and network transmission technology, and the audio collection that audio collecting system then adopts many group microphone arrays to combine with the information processing platform is handled experimental system.
1) crossroad video traffic real time information sampling hardware system scheme
At present both at home and abroad the comprehensive real-time monitoring information acquisition system for the city road crossing traffic mainly contains three kinds of solutions: based on the solution of analog video camera and DVR; Solution based on analog video camera and video server; The solution of video camera Network Based and network transmission technology.The present invention adopts the solution of video camera Network Based and network transmission technology towards Traffic monitoring technology of new generation.The system buildup of video camera Network Based then has suitable dirigibility, can cooperate independent memory device, signal handling equipment, and ordinary PC can realize storage and back-end processing to data.Web camera is the novel video camera that integrates digital camera and network function, functions such as A/D conversion, digital compression and Network Transmission that its principal feature has been integrated, represented the development trend of the front-end equipment of monitoring technique of new generation, the digital video signal of its output can directly pass through internet transmission.This scheme has been saved hardware devices such as video frequency collection card, video server and image pick-up card, has good cost performance and actual application value.System schema as shown in Figure 2.
2) the real-time acquisition hardware system schema of crossroad audio traffic information
Use for reference the research approach of main flow in the world, consider that the convenience of installation and system realize reliability, (Array Acoustic Capture Unit AACU) is made of 4 pack modules in the traffic audio collection unit that the present invention adopts.Each group comprises 8 ECM microphones, signal condition module (realizing the amplification and the filter function of feeble signal), analog-to-digital conversion module (realizing the multi-channel synchronous analog-digital conversion function) and FPGA control and processing module (realizing control signal collection, storage, processing, demonstration and transfer function).As shown in Figure 3.Crossroad audio collection subsystem comprises 4 distributed AACU.Each AACU unit will be vertically fixed on the traffic-control device cross bar, monitoring traffic lights dead ahead traffic route.The voice data that the audio signal sample subsystem collects can adopt wired or wireless mode to output to the audio-frequency information processing subsystem.
2, based on the traffic parameter feature extraction and the traffic events detection technique of audio/video information
1) based on the vehicle Automatic Measurement Technique of video information
The present invention adopts mixed Gauss model to realize the automatic detection of moving vehicle.Mixed Gauss model can be described as:
Wherein, P (A
t) be constantly at t, the sampled value A of certain pixel A
tThe probability that occurs.K is the number (span is between 3-5) of different Gaussian distribution during mixed Gaussian distributes.ω
I, tBe the estimation weights of t i Gaussian distribution during constantly mixed Gaussian distributes (Estimate Of Weight, 1≤i≤K, and the weights sum is 1).μ
I, tBe the t moment, the average (Mean Value) of i Gaussian distribution during mixed Gaussian distributes.∑
I, tBe the t moment, the covariance matrix (CovarianceMatrix) of i Gaussian distribution during mixed Gaussian distributes, and
Wherein, in order to reduce data operation quantity, we suppose that the sampled value of each Color Channel of pixel is uncorrelated each other, and have identical variances sigma
I, t 2Therefore, in calculating, we adopt:
In the formula (2), the dimension of unit matrix equals the number of Color Channel (for example: the RGB color space has three Color Channels, so the dimension of unit matrix is 3); η (A
t, μ
I, t, ∑
I, t) being the probability density function (Gaussian Probability Density Function) of i Gaussian distribution in the mixed Gaussian distribution, can be described as:
In mixed Gaussian distributed, each Gaussian distribution had different weights ω respectively
I, tWith distribution priority parameters ω
I, t/ σ
I, t, and they always according to priority parameter is descending sorts.Wherein, distribution priority parameters ω
I, t/ σ
I, tImplication be that the mathematical model of i Gaussian distribution correspondence is represented the degree of approximation of background model.Therefore forward more Gauss model, energy approximate representation video background model more put in order.
In addition, the present invention proposes and has realized eliminating and vehicle detecting algorithm based on the vehicle shadow of texture information, we have compared ratioedge and two kinds of vehicle shadow elimination algorithms of texture-based at research, result of study shows, we propose based on texture information and be aided with brightness and the vehicle dividing method of chrominance information can suppress the influence of shade to testing result effectively.Wherein, the generation of Texture-Likelihood Map is the texture information that obtains image block according to the calculating of texture autocorrelation function R.Studies show that the image-region that shade covers only changes its monochrome information, and texture information is not changed; The texture information that is the image block in shadow image piece and the corresponding background image in the foreground image is identical.Therefore, can eliminate the shade interference by the texture autocorrelation function difference of calculating between foreground image and the background image.But studies show that, only often can't extract vehicle (for example: vehicle color and road color are too approaching) exactly according to texture information.The present invention adopts auxiliary monochrome information and chrominance information to improve correctness and stability that vehicle extracts.Luminance-Likelihood Map then is the vehicle bianry image that relies on monochrome information to set up, and wherein we have realized from rgb space to the conversion in YCbCr space to obtain the monochrome information of separation.Y-channel among the YcbCr is the luminance channel of image; Cb and Cr are two chrominance channels in the image.In addition, we utilize the chrominance information of video, promptly calculate Chrominance-Likelihood Map.This C-Map is the vehicle bianry image that relies on the chrominance information foundation in YCbCr space.The method that the invention allows for a kind of threshold value selection of robust realizes that L-MAP and C-MAP's cuts apart to T-MAP.In addition, the invention still further relates to the automatic extractive technique of a kind of traffic parameter, independently apply for a patent, no longer endure herein and state based on the index entropy.
2) based on the crossroad access scene information extraction algorithm of audio frequency
The traffic characteristic vector based on audio-frequency information that the present invention proposes as shown in Figure 4 forms and traffic event automatic detection system.On the basis of the statistical property of analyzing and study the road sound signal, we adopt the traffic sound signal of 3 seconds length as process object, it are carried out down-sampled (passband frequency range is 500HZ～5000Hz) with the influence of eliminating non-vehicle audio signal, carries out time domain and spatial domain windowing pre-service to reduce truncation effect to lower computation complexity (can require adjust according to systematic sampling rate and computation complexity), to carry out bandpass filtering.The present invention adopts the optimum beam formation device based on microphone array (promptly to adopt microphone array output energy minimum criteria, and the output constant in energy is constraint on target direction), realize spatial filtering and denoising enhancing to the road traffic sound signal of specific direction.Optimum beam forms the input signal of the signal of device output as the traffic route feature extraction.In order to obtain to characterize the feature of road traffic audio event, relevance, sound signal short-time energy, zero-crossing rate direction, wavelet transform (DWT) coefficient, Fourier transform (FFT) coefficient, discrete cosine transform (DCT) coefficient, real cepstrum conversion (the Real Cepstrum Transform of traffic sound signal adopted in the present invention's research and proposition, RCT) coefficient, Mel frequency cepstral (MEL Frequency Cepstral Transform, MFCT) coefficient etc. forms the proper vector of road sound signal, as the input vector of rear end traffic events sorter.The present invention adopts the linear discriminant analysis method, and (LinearDiscriminant Analysis LDA) realizes choose (the carrying out dimensionality reduction) of the feature be associated with event detection.The present invention also adopts the MUSIC algorithm to realize the estimation of vehicle number to vehicle, and adopts cross correlation algorithm to realize speed of a motor vehicle estimation.Studies show that, based on microphone array beam-forming technology the lose interest in road audio signal components in direction or zone of filtering effectively, suppress to disturb, help to realize effective extraction effectively the audio-frequency information feature.
3) traffic events detects automatically
The present invention has at first proposed traffic events and has detected the decision level fusion technology.The fusion of decision level at first makes a policy according to the information to same target that each module obtained, and these are independently made a strategic decision integrates at last, provides final decision.We consider the diversity and the complicacy of traffic events, have designed the multi-level decision-making level and have merged the traffic events detection method.At first, judge current weather condition, be divided into fine day, cloudy, wet weather and night four kinds of situations, every kind of situation has independently sorter.Promptly the situation according to weather adopts different traffic parameter extractive techniques, uses different sorters to carry out the judgement of traffic events.Independently in the sorter, be divided into two-stage again, as shown in Figure 5 at each.The first order judges whether the crossroad blocks up, and the second level then further differentiates whether have parking violation, collision or retrograde.By by the multi-level decision-making level fusing method of integral body to the part, promptly guaranteed the accuracy rate of classification, guaranteed the determinacy that traffic events detects again.
The present invention research and proposed transport information feature level integration technology based on multi-source information, according to the traffic video proper vector of Video processing subsystem output and the audio frequency characteristics vector of Audio Processing subsystem output, merge and form the transport information proper vector that characterizes the crossroad access situation; The generation of transport information vector need be carried out correlation analysis and optimization in vector space.The transport information vector will be used for the training and testing to sorter as the input vector of traffic events automatic categorizer.
The present invention proposes traffic events sorting technique based on support vector machine and hidden Markov model.At first, select typical crossroad, Shenzhen to carry out the audio, video data collection, the movable information that utilizes image difference to obtain has been set up the traffic video database; We adopt transport information proper vector and hidden Markov model (HMM) or support vector machine that four kinds of traffic logical signals and traffic events are carried out classification learning; At last, utilize the actual measurement traffic data that sorter is tested.In order to test the performance of the algorithm that we propose, we have carried out one group of experiment.Fig. 6 shows a kind of only the extraction based on video features and the traffic events sorting algorithm process flow diagram of HMM.In this case, we have at first set up video information data base.For a typical crossroad, mainly contain four signal lamp indications, be respectively Dong-Xi, Xi-Dong, south-north, North-south, each signal lamp comprises turning left and U-shaped is curved turns around.We are defined as the 5th kind of driving situation to traffic events, comprise accident, parking violation, lamp etc. drives in the wrong direction, makes a dash across the red light.Each 200 width of cloth video difference of the driving situation that our picked at random is five types are divided picture totally 1000 width of cloth pictures, and wherein 500 width of cloth are as training sample, and other 500 width of cloth are as test data.Generation about video feature vector, we have adopted principal component analysis (PCA) (PCA) that input picture is carried out dimension-reduction treatment (having kept the low order major component) to reduce computation complexity and to improve characteristic information, adopt hybrid discrete cosine conversion (DCT) and modified discrete Fourier transformation (FFT) to form the input feature value of information characteristics as HMM then.Wherein, DCT has realized the nearly orthogonal conversion of image, keeps the DCT low frequency coefficient and has then kept the information component that characterizes concentrated main energy effectively; Utilize the characteristics such as strange, idol, void, reality of FFT conversion to realize effective compression of video image.Experimental result shows that overall classification can reach 91% accuracy to algorithm shown in Figure 6 to traffic, and wherein the detection to traffic events can reach 95% accuracy.
3, a cover is complete urban transportation monitoring and event detection intelligent transportation system
The present invention has integrated traffic audio-video collection system, audio/video information disposal system, traffic event automatic detection system, and having increased traffic events demonstration and warning system, the original traffic audio/video information that has formed visual crossroad shows, traffic flow parameter information shows and traffic event information detects software (as shown in Figure 7).The present invention adopts the input of the output of front end audio, video data collecting unit and video information process subsystem as this software, design has realized man-machine friendly multifunctional traffic information monitoring and display terminal, be divided into the client and server end, respectively corresponding traffic scene and traffic control center.The major function of this software comprises: 1) control function: transport information input source, traffic events type, sorting algorithm of client etc. can be set; 2) multiwindow traffic information display function: realize windows such as original audio frequency and video transport information playback demonstration, traffic detection system position display, the demonstration of traffic events scene, the demonstration of hyperchannel traffic audio frequency and video parameter, the demonstration of hyperchannel traffic raw data based on map; 3) result exports Presentation Function: realization traffic event data storehouse is upgraded automatically, and the traffic events examining report generates and storage automatically, traffic hazard warning, the output of traffic events induction information; 4) traffic event information transfer function: realized that client is sent to server end with 5 seconds the audio/video information in detected traffic events front and back, screens for the staff.
The display terminal that Fig. 8 designs for seminar (client) interface synoptic diagram.As can be seen from the figure, this interface is divided into eight parts, is respectively real-time monitored space display window, traffic signals display window, background modeling display window, detection zone prospect display window, sound signal display window, logout form, traffic parameter demonstration, detects information setting.Wherein: 1) real-time monitored space display window: display network camera video acquisition subsystem output raw data; When having an accident, provide obvious sign in the specific region; 2) traffic monitoring video display window: the passed through zone that shows the current crossroad of current time; 3) background modeling display window: show the background modeling result who comes from the video information process subsystem based on the mixed Gaussian theory; 4) detection zone prospect display window: show the video image in the area-of-interest (ROI); 5) sound signal display window: the signal waveforms that shows the output of audio collection subsystem; 6) logout form: the time of origin of recording traffic incident, type, place; 7) traffic parameter shows: show parameters such as vehicle flowrate, average speeds, traffic density, time headway; 8) detect information setting: detect the information setting part mainly to the input of video flowing select, functions such as the setting of parameter extraction detection zone, window maximum/minimize, event detection type selecting, server ip setting, help.The display terminal that Fig. 9 designs for seminar (server end) interface client synoptic diagram.Server end is mainly used in warning and the video image of reception from the client that traffic hazard takes place, and notice traffic control central task personnel are confirmed.Major function comprises: video playback, incident is confirmed, warning Email/note transmission etc.
Advantage of the present invention:
1. the omni-directional visual technology is incorporated into the crossroad traffic event automatic detection system of ITS, has overcome the limitation in the traditional cameras visual angle and the visual field; By many orders omni-directional visual video camera, realize that the real-time collection and the high-resolution local traffic scene information of the information of traffic scene on a large scale of low resolution gathered;
2. with advanced person's video image information treatment technology, the crossroad traffic event automatic detection system that multisource information fusion technology draws ITS, proposed to innovation based on Multi-source Information Fusion strategy, based on the robust traffic background modeling of mixed Gauss model, based on the robust vehicle detection of texture and index entropy with traffic parameter accurately extracts and based on Kalman's vehicle tracking technology etc.;
3. the microphone array beam-forming technology with the advanced person is applied to crossroad traffic event automatic detection system, utilize sound signal to the insensitivity of traffic change of background, to the independence of traffic events generation occasion, the advantages such as real-time of Audio Signal Processing, remedied the limitation of video technique, help to realize the detection quick and precisely of traffic events such as vehicle is collided, vehicle suddenly stops, road construction noise, improved the event detection performance of total system;
4. Multi-source Information Fusion and the classifier technique with the advanced person is applied to crossroad traffic event automatic detection system, innovation research and having realized based on traffic event automatic detection system system-level, feature level information fusion strategy, to adapt to the traffic monitoring technical need of complicated weather and traffic scene.The traffic video proper vector that proposes innovation ground to adopt merges traffic audio frequency characteristics vector and forms transport information proper vector based on multi-source information as the input vector of rear end based on support vector machine and hidden Markov chain (HMM) model classification device, realizes advanced real-time traffic event automatic detection system.
Description of drawings:
Fig. 1 traffic event automatic detection system hardware is disposed structural drawing
The solution of Fig. 2 video camera Network Based and network transmission technology
Fig. 3 AACU audio collection cell schematics
Fig. 4 is based on the automatic testing process figure of the traffic events of audio-frequency information
Fig. 5 traffic events detects the decision level fusion process flow diagram
Fig. 6 is based on the traffic events sorting algorithm process flow diagram of video
Urban transportation monitoring and event detection intelligent transportation system that Fig. 7 is complete
Fig. 8 Message Display Terminal (client) interface synoptic diagram
Fig. 9 Message Display Terminal (server end) interface synoptic diagram
Crossroad access monitoring of Figure 10 city and traffic event automatic detection system process flow diagram
Figure 11 demo system
Embodiment:
Be concrete case study on implementation below to the traffic event automatic detection system of the present invention's proposition.The implementation case does not limit the present invention, for those skilled in the art, under the prerequisite that does not break away from the principle of the invention, can also make some improvement and variation, and these improvement and variation also should be considered as within protection scope of the present invention.
The multi-angle video monitoring system comprises: 1) select for use based on one one of catadioptric omnidirectional vision camera, comprise one one of the colour TV camera that rotates the control The Cloud Terrace; 2) intelligent network D1 resolution, 25 frames/second CCD colour TV camera is 4 ones; 3) network DVR video recording apparatus;
The microphone array audio information acquisition system comprises: based on 8 good comprehensive microphones of consistance, 8 passage high-gains, good prime amplifier and the signal processing platform of consistance;
The implementation case operates on the XW4600 graphics workstation of HP and realizes that concrete configuration is as follows:
A kind of traffic event automatic detection system based on Multi-source Information Fusion, its flow process below are the detailed processes that case is implemented as shown in figure 10:
Step 1: select typical case crossroad, city, set up a cover audio frequency and video traffic information acquisition system; Obtain traffic video stream information, traffic audio-frequency information;
Step 2: the original traffic audio, video data that will collect is directly passed traffic control center back and is shown, and stores.Simultaneously audio frequency array information and multi-angle video information are sent to respectively in audio processing modules and the video processing module;
Step 3: in the Video processing subsystem, each road video data is carried out vehicle detection automatically, vehicle tracking and traffic parameter extract, form the traffic video proper vector.Audio-frequency information is carried out pre-service, wave beam formation, audio feature extraction etc. form traffic audio frequency characteristics vector;
Step 4:, carry out sound and look feature level fusion formation transport information proper vector according to the video feature vector and the audio frequency characteristics vector that extract;
Step 5: select different sorters to carry out traffic events automatic categorizer training (can select neural network, support vector machine, HMM isotype recognition classifier).With the transport information proper vector is that the sorter input is trained sorter, obtains the traffic events automatic categorizer towards this traffic cross-road, and sorter is implanted the automatic detection that front end system is used for the real-time traffic incident;
Step 6: utilize the traffic events automatic categorizer that trains that the multi-source traffic information that collects is carried out the automatic detection of traffic events, if no traffic events, then repeating step 2-step 5.If traffic events takes place, then trigger alarm mechanism, carry out step 7;
Step 7: when the traffic events automatic categorizer detects traffic events, to on the map of traffic control center, remind managerial personnel to note the site of an accident in modes such as flickers in this position, crossroad, and pass five seconds audio frequency and video before and after the traffic events back command centre, back up traffic events simultaneously;
Step 8: the point duty personnel judge whether the relevant traffic warning message of needs issue according to video playback and monitoring video, if need issue, then system will realize the crossing event information issue of modes such as mail, phone, note, traffic indication board, webpage automatically, and the relevant traffic department of notice handles;
Step 9: system will wait for that traffic events disposes, and recover normal when traffic, and system then cancels warning message, still with unobstructed information of mode issuing traffic such as mail, phone, note, traffic indication board, webpages.
Demo system of the present invention and test result are by shown in Figure 11.
Above content be in conjunction with concrete preferred implementation to further describing that the present invention did, can not assert that concrete enforcement of the present invention is confined to these explanations.For the general technical staff of the technical field of the invention, without departing from the inventive concept of the premise, can also make some simple deduction or replace, all should be considered as belonging to protection scope of the present invention.
Claims (9)
1. the crossroad traffic event automatic detection system based on Multi-source Information Fusion is characterized in that, described method comprises:
1) selects typical case crossroad, city, set up the system that obtains of a cover multi-source traffic information (hyperchannel traffic audio-frequency information, traffic video information, multi-angle video information, high and low angle video information);
The original traffic audio, video data that 2) will collect is directly passed traffic control center back and is shown, and stores.Simultaneously audio frequency array information and multi-angle video information are sent to respectively in Audio Processing subsystem and the Video processing subsystem;
3) in the Video processing subsystem, each road video data is carried out vehicle detection automatically, vehicle tracking and traffic parameter extract automatically;
4) in the Audio Processing subsystem, the multi-channel audio signal of microphone array output is carried out wave beam form (spatial filtering), feature extraction and traffic parameter and extract automatically;
5) according to the video traffic parameter and the audio frequency traffic parameter that extract, carry out audio/video information and merge and form Multi-source Information Fusion transport information proper vector, adopt corresponding transport information proper vector according to the different weather testing conditions, realize that the feature level of transport information merges;
6) select neural network, HMM, support vector machine isotype recognition classifier, handing over information part proper vector with Multi-source Information Fusion is that the sorter input vector is trained, and obtains robust traffic events recognition classifier;
7) utilize the traffic events recognition classifier that trains to realize the automatic detection of real-time traffic incident, if traffic events takes place, then trigger the report subsystem, send location information and transmission and this traffic video signal of 5 seconds in traffic events front and back of storage that traffic events takes place, and open the check passage;
8) the traffic events report information occurs with flashing mode on the monitoring map of traffic control center, and reminder center managerial personnel carry out hand inspection;
9) the traffic hub managerial personnel confirm traffic events according to the incident video information in 5 seconds, need the traffic events of processing in time if confirm as, then system produces this traffic events automatically and reports that mail, phone, note, webpage, the relevant traffic department of traffic indication board (optional) information notice handle;
10) treat that traffic events disposes, it is normal that traffic is recovered, and the cancellation warning message is still with unobstructed information of mode issuing traffic such as mail, phone, note, traffic indication board, webpages.
2. multi-source traffic information according to claim 1 obtains system, it is characterized in that step 1) is described, low angle four road video acquisitions, direction is for going out the car direction, and height is contour with the traffic lights flag-rod, and main monitoring range is for going out the car direction forward till the crossroad; High angle one road video acquisition, be positioned on the four crossway oral-lateral skyscraper or on the high bar of special-purpose adjustable height, depression angle becomes the 45-90 degree adjustable with the local horizon, satisfy the whole crossroad of covering area-of-interest and be as the criterion, and monitors the traffic lights queuing situation of certain-length simultaneously; The audio frequency array is positioned at car direction right side, and is adjustable to vehicle queue direction 0-30 rice at vehicle outlet, and audio frequency array height 0.5-2 rice is adjustable, and the microphone array column direction becomes with the local horizon-and the pitching of 45-+45 degree is adjustable;
3. audio frequency and video transport module according to claim 1, it is characterized in that, step 2) described, the front end video signal collective adopts web camera, by netting twine or wireless network transmissions, the audio frequency array adopts 8 passage microphone audio acquisition system and Network Transmission with independent intellectual property right;
4. processing system for video according to claim 1, it is characterized in that, step 3) is described, employing forms the traffic video proper vector based on the automatic detection algorithm of the moving vehicle of mixed Gauss model, detect automatically and traffic parameter extracts automatically based on the moving vehicle of image texture information and image information index entropy;
5. audio frequency processing system according to claim 1, it is characterized in that, step 4) is described, employing strengthens denoising based on hyperchannel microphone array beam-forming technology in advance to the traffic signals of specific target areas, adopt wavelet transform (Discrete Wavelet Transform, DWT), real cepstrum conversion (Real Cepstrum Transform, RCT), Mel frequency cepstral (MEL Frequency Cepstral Transform, MFCT), features such as short-time energy spectrum and zero-crossing rate, and adopt the linear discriminant analysis method (Linear Discriminant Analysis, that LDA) realizes the feature that is associated with event detection chooses (carrying out dimensionality reduction) formation traffic audio frequency characteristics vector;
6. audio frequency and video according to claim 1 merge, and it is characterized in that step 5) is described, traffic video proper vector and traffic audio frequency characteristics vector are carried out the fusion of feature level, form the transport information proper vector,, be used for training and testing sorter as the input vector of sorter;
7. traffic events Automatic Measurement Technique according to claim 1 is characterized in that step 6) and step 7) are described, based on the traffic events sorting technique of hidden Markov model and support vector machine.At first set up transport information characteristic vector data storehouse sorter is trained, and the sorter that trains is incorporated in the traffic event automatic detection system, realize the real-time monitoring of crossroad access information, the real time automatic detection of traffic events;
8. traffic control center according to claim 1, it is characterized in that, step 8) is described, traffic audio-video collection system, audio/video information disposal system, traffic event automatic detection system have been integrated, and integrated the traffic events warning system, formed original traffic flow data, traffic flow parameter information and the traffic event information monitoring software of visual crossroad;
9. warning system according to claim 1, it is characterized in that, step 8) and step 9) are described, client software is divided into 8 parts, is respectively real-time monitored space display window, traffic signals display window, background modeling display window, detection zone prospect display window, sound signal display window, logout form, traffic parameter demonstration, detects information setting.The back takes place in staff's affirmation incident and sends warning message in server software, in modes such as mail, phone, note, traffic indication board, webpages the crossing event information is issued; Recovering the unobstructed information of normal back transmission, still carry out in the above described manner.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910238815A CN101751782A (en) | 2009-12-30 | 2009-12-30 | Crossroad traffic event automatic detection system based on multi-source information fusion |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910238815A CN101751782A (en) | 2009-12-30 | 2009-12-30 | Crossroad traffic event automatic detection system based on multi-source information fusion |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101751782A true CN101751782A (en) | 2010-06-23 |
Family
ID=42478708
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200910238815A Pending CN101751782A (en) | 2009-12-30 | 2009-12-30 | Crossroad traffic event automatic detection system based on multi-source information fusion |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101751782A (en) |
Cited By (68)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101917601A (en) * | 2010-08-26 | 2010-12-15 | 四川大学 | Digital video intelligent monitoring equipment based on dual camera and data processing method |
| CN102521983A (en) * | 2011-12-23 | 2012-06-27 | 北京易华录信息技术股份有限公司 | Vehicle violation detection system based on high definition video technology and method thereof |
| CN102945603A (en) * | 2012-10-26 | 2013-02-27 | 青岛海信网络科技股份有限公司 | Method for detecting traffic event and electronic police device |
| CN103280096A (en) * | 2013-03-29 | 2013-09-04 | 苏州数伦科技有限公司 | Traffic monitoring management system |
| CN103295396A (en) * | 2013-04-19 | 2013-09-11 | 福建工程学院 | Method and system for speedy ex-situ evidence collection of traffic accident |
| CN103824454A (en) * | 2014-02-21 | 2014-05-28 | 南京莱斯信息技术股份有限公司 | Multi-zone illegality automatic capture method based on video event detection |
| CN103838772A (en) * | 2012-11-26 | 2014-06-04 | 香港生产力促进局 | Multi-source traffic data fusion method |
| CN103985253A (en) * | 2014-05-26 | 2014-08-13 | 北京易华录信息技术股份有限公司 | System and method for fast finding abnormal stopping in road junction controlled by signal lamp |
| CN104200657A (en) * | 2014-07-22 | 2014-12-10 | 杭州智诚惠通科技有限公司 | Traffic flow parameter acquisition method based on video and sensor |
| CN104217524A (en) * | 2014-09-17 | 2014-12-17 | 国家电网公司 | Smoke monitoring system and method |
| CN104240510A (en) * | 2014-10-17 | 2014-12-24 | 成都四为电子信息股份有限公司 | Urban traffic electronic card port system and detection method thereof |
| CN104240505A (en) * | 2014-09-11 | 2014-12-24 | 胡又宏 | Method for analyzing road traffic video image information |
| CN104408924A (en) * | 2014-12-04 | 2015-03-11 | 深圳北航新兴产业技术研究院 | Detection method for abnormal traffic flow of urban road based on coupled hidden markov model |
| CN105355067A (en) * | 2015-11-25 | 2016-02-24 | 南京莱斯信息技术股份有限公司 | Publishing method of intersection traffic information |
| CN105389977A (en) * | 2014-08-28 | 2016-03-09 | 福特全球技术公司 | Apparatus and System for Generating Emergency Vehicle Record Data |
| CN105513349A (en) * | 2014-12-03 | 2016-04-20 | 中咨泰克交通工程集团有限公司 | Double-perspective learning-based mountainous area highway vehicle event detection method |
| CN105741565A (en) * | 2016-04-21 | 2016-07-06 | 正元地理信息有限责任公司 | Method for automatic determining accident and giving off alarm based on monitoring video |
| CN105761500A (en) * | 2016-05-10 | 2016-07-13 | 腾讯科技（深圳）有限公司 | Traffic accident handling method and traffic accident handling device |
| CN105976616A (en) * | 2016-07-29 | 2016-09-28 | 林玉峰 | Analysis method of traffic event monitoring model based on traffic data |
| CN106251663A (en) * | 2016-09-14 | 2016-12-21 | 深圳市喜悦智慧数据有限公司 | A kind of intelligent traffic robot |
| CN106408940A (en) * | 2016-11-02 | 2017-02-15 | 南京慧尔视智能科技有限公司 | Microwave and video data fusion-based traffic detection method and device |
| CN106851132A (en) * | 2017-01-20 | 2017-06-13 | 安徽达尔智能控制系统股份有限公司 | Traffic accident intelligent checking system |
| CN106919925A (en) * | 2017-03-07 | 2017-07-04 | 南京师范大学 | A kind of Ford Motor's detection method based on Wavelet Entropy Yu artificial neural network |
| CN107016859A (en) * | 2017-05-18 | 2017-08-04 | 肇庆学院 | A kind of crossing traffic management system, methods and applications |
| CN107194490A (en) * | 2016-03-14 | 2017-09-22 | 商业对象软件有限公司 | Predict modeling optimization |
| CN107292386A (en) * | 2016-04-11 | 2017-10-24 | 福特全球技术公司 | Detected using the rainwater of the view-based access control model of deep learning |
| CN107590995A (en) * | 2016-07-08 | 2018-01-16 | 沃尔沃汽车公司 | Include the method and system of the database of the influence traffic events of report for safeguarding |
| CN107665583A (en) * | 2017-11-20 | 2018-02-06 | 中兴软创科技股份有限公司 | The computational methods of track saturation volume rate under the conditions of a kind of different weather |
| CN107845262A (en) * | 2017-11-09 | 2018-03-27 | 山东浪潮云服务信息科技有限公司 | A kind of Predictive Methods of Road Accidents and device |
| CN108198423A (en) * | 2018-01-24 | 2018-06-22 | 哈尔滨工业大学 | A kind of protrusion Crash characteristics recognition methods of two displays signal control cross level-crossing |
| DE102017200153A1 (en) | 2017-01-09 | 2018-07-12 | Ford Global Technologies, Llc | Method for detecting traffic situations |
| CN108304773A (en) * | 2017-12-25 | 2018-07-20 | 广州市高科通信技术股份有限公司 | A kind of vehicle density analysis method, device, electronic equipment and storage medium based on wavelet transformation |
| CN108307153A (en) * | 2017-01-13 | 2018-07-20 | 佳能株式会社 | Video monitoring devices and its control method, storage medium and video monitoring system |
| CN108764042A (en) * | 2018-04-25 | 2018-11-06 | 深圳市科思创动科技有限公司 | A kind of exception traffic information recognition methods, device and terminal device |
| CN109033997A (en) * | 2018-07-02 | 2018-12-18 | 厦门快商通信息技术有限公司 | A kind of lumbering event detecting method and system |
| CN109935078A (en) * | 2019-02-18 | 2019-06-25 | 深兰科技（上海）有限公司 | A kind of automatic evolution type traffic control system |
| WO2019127651A1 (en) * | 2017-12-30 | 2019-07-04 | 惠州学院 | Method and system thereof for identifying malicious video |
| CN110097755A (en) * | 2019-04-29 | 2019-08-06 | 东北大学 | Freeway traffic flow amount state identification method based on deep neural network |
| CN110291477A (en) * | 2016-12-02 | 2019-09-27 | 斯塔斯凯机器人公司 | Vehicle control system and application method |
| CN110415272A (en) * | 2019-07-08 | 2019-11-05 | 东华大学 | A kind of battery truck drives in the wrong direction detection system and method |
| CN110494902A (en) * | 2017-02-03 | 2019-11-22 | 西门子交通有限责任公司 | For managing the system, apparatus and method of the traffic in geographical location |
| CN110545359A (en) * | 2019-08-02 | 2019-12-06 | 国家计算机网络与信息安全管理中心 | Communication line feature extraction method, communication line identification method and device |
| WO2020024447A1 (en) * | 2018-08-01 | 2020-02-06 | 平安科技（深圳）有限公司 | Traffic accident processing method and apparatus, and computer readable storage medium |
| CN110942629A (en) * | 2019-11-29 | 2020-03-31 | 中核第四研究设计工程有限公司 | Road traffic accident management method and device and terminal equipment |
| CN110942011A (en) * | 2019-11-18 | 2020-03-31 | 上海极链网络科技有限公司 | Video event identification method, system, electronic equipment and medium |
| CN111066318A (en) * | 2017-09-15 | 2020-04-24 | 三菱电机株式会社 | Monitoring support device and monitoring support system |
| CN111127876A (en) * | 2019-11-18 | 2020-05-08 | 腾讯科技（深圳）有限公司 | Information extraction method and device for Internet of vehicles |
| CN111145558A (en) * | 2019-12-27 | 2020-05-12 | 北京容联易通信息技术有限公司 | Illegal behavior identification method based on high-point video monitoring |
| CN111192456A (en) * | 2020-01-14 | 2020-05-22 | 泉州市益典信息科技有限公司 | Road traffic operation situation multi-time scale prediction method |
| CN111768625A (en) * | 2020-07-01 | 2020-10-13 | 中国计量大学 | Traffic road event prediction method based on graph embedding |
| CN111882861A (en) * | 2020-06-06 | 2020-11-03 | 浙江工业大学 | Online traffic incident perception system based on edge cloud fusion |
| CN111932427A (en) * | 2020-09-24 | 2020-11-13 | 北京泰策科技有限公司 | Method and system for detecting emergent public security incident based on multi-mode data |
| CN112164223A (en) * | 2020-02-27 | 2021-01-01 | 周国霞 | Intelligent traffic information processing method and device based on cloud platform |
| CN112185110A (en) * | 2020-09-17 | 2021-01-05 | 李志� | Remote monitoring method and device based on Internet |
| CN112637568A (en) * | 2020-12-24 | 2021-04-09 | 中标慧安信息技术股份有限公司 | Distributed security monitoring method and system based on multi-node edge computing equipment |
| CN112634634A (en) * | 2020-11-24 | 2021-04-09 | 深圳技术大学 | Intelligent rod aid decision-making method, intelligent rod and system |
| CN112885086A (en) * | 2021-01-18 | 2021-06-01 | 湖南省交通规划勘察设计院有限公司 | Sudden congestion judging system based on multi-source traffic big data |
| CN112906634A (en) * | 2021-03-18 | 2021-06-04 | 西北大学 | Video segment sequence prediction model establishment and sequence prediction method and system based on VSS |
| CN113299317A (en) * | 2021-05-14 | 2021-08-24 | 东南大学 | Intersection collision audio early warning evaluation method based on MFCC feature extraction |
| CN113537170A (en) * | 2021-09-16 | 2021-10-22 | 北京理工大学深圳汽车研究院(电动车辆国家工程实验室深圳研究院) | Intelligent traffic road condition monitoring method and computer readable storage medium |
| CN113706871A (en) * | 2021-09-13 | 2021-11-26 | 中交第一公路勘察设计研究院有限公司 | Multisource data fusion system in urban traffic big data processing |
| CN114446052A (en) * | 2022-01-17 | 2022-05-06 | 东南大学 | Multi-source data multi-stage fusion-based highway incident detection method and device |
| JP2022099389A (en) * | 2020-12-23 | 2022-07-05 | 三菱電機株式会社 | In-vehicle image information processor, in-vehicle image information transmitter, and image information processing method |
| CN114827590A (en) * | 2022-06-22 | 2022-07-29 | 浙江力嘉电子科技有限公司 | Camera position distribution rationality analysis method and device and electronic equipment |
| CN116266215A (en) * | 2022-08-01 | 2023-06-20 | 广西北投公路建设投资集团有限公司 | Geometric feature extraction method for pavement recessive disease area |
| CN116518989A (en) * | 2023-07-05 | 2023-08-01 | 新唐信通(浙江)科技有限公司 | Method for vehicle navigation based on sound and thermal imaging |
| CN117935559A (en) * | 2024-03-07 | 2024-04-26 | 深圳市新创中天信息科技发展有限公司 | Traffic accident decision system based on multi-mode fusion perception technology |
| CN117935559B (en) * | 2024-03-07 | 2024-05-24 | 深圳市新创中天信息科技发展有限公司 | Traffic accident decision system based on multi-mode fusion perception technology |
-
2009
- 2009-12-30 CN CN200910238815A patent/CN101751782A/en active Pending
Cited By (95)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101917601B (en) * | 2010-08-26 | 2012-07-25 | 四川大学 | Digital video intelligent monitoring equipment based on dual camera and data processing method |
| CN101917601A (en) * | 2010-08-26 | 2010-12-15 | 四川大学 | Digital video intelligent monitoring equipment based on dual camera and data processing method |
| CN102521983A (en) * | 2011-12-23 | 2012-06-27 | 北京易华录信息技术股份有限公司 | Vehicle violation detection system based on high definition video technology and method thereof |
| CN102521983B (en) * | 2011-12-23 | 2013-10-16 | 北京易华录信息技术股份有限公司 | Vehicle violation detection system based on high definition video technology and method thereof |
| CN102945603A (en) * | 2012-10-26 | 2013-02-27 | 青岛海信网络科技股份有限公司 | Method for detecting traffic event and electronic police device |
| CN103838772A (en) * | 2012-11-26 | 2014-06-04 | 香港生产力促进局 | Multi-source traffic data fusion method |
| CN103838772B (en) * | 2012-11-26 | 2017-09-05 | 香港生产力促进局 | A kind of multi-source traffic data fusion method |
| CN103280096A (en) * | 2013-03-29 | 2013-09-04 | 苏州数伦科技有限公司 | Traffic monitoring management system |
| CN103295396A (en) * | 2013-04-19 | 2013-09-11 | 福建工程学院 | Method and system for speedy ex-situ evidence collection of traffic accident |
| CN103295396B (en) * | 2013-04-19 | 2015-11-04 | 福建工程学院 | Speedy ex-situ evidence collection of traffic accident method and system |
| CN103824454A (en) * | 2014-02-21 | 2014-05-28 | 南京莱斯信息技术股份有限公司 | Multi-zone illegality automatic capture method based on video event detection |
| CN103985253A (en) * | 2014-05-26 | 2014-08-13 | 北京易华录信息技术股份有限公司 | System and method for fast finding abnormal stopping in road junction controlled by signal lamp |
| CN103985253B (en) * | 2014-05-26 | 2016-05-25 | 北京易华录信息技术股份有限公司 | The system and method for exception parking in the crossing of a kind of quick discovery signals lamp control |
| CN104200657A (en) * | 2014-07-22 | 2014-12-10 | 杭州智诚惠通科技有限公司 | Traffic flow parameter acquisition method based on video and sensor |
| CN105389977A (en) * | 2014-08-28 | 2016-03-09 | 福特全球技术公司 | Apparatus and System for Generating Emergency Vehicle Record Data |
| CN104240505A (en) * | 2014-09-11 | 2014-12-24 | 胡又宏 | Method for analyzing road traffic video image information |
| CN104217524A (en) * | 2014-09-17 | 2014-12-17 | 国家电网公司 | Smoke monitoring system and method |
| CN104240510A (en) * | 2014-10-17 | 2014-12-24 | 成都四为电子信息股份有限公司 | Urban traffic electronic card port system and detection method thereof |
| CN105513349A (en) * | 2014-12-03 | 2016-04-20 | 中咨泰克交通工程集团有限公司 | Double-perspective learning-based mountainous area highway vehicle event detection method |
| CN105513349B (en) * | 2014-12-03 | 2017-12-05 | 中咨泰克交通工程集团有限公司 | Mountainous area highway vehicular events detection method based on double-visual angle study |
| CN104408924B (en) * | 2014-12-04 | 2016-06-01 | 深圳北航新兴产业技术研究院 | A kind of urban road abnormal traffic stream detection method based on coupled hidden markov model |
| CN104408924A (en) * | 2014-12-04 | 2015-03-11 | 深圳北航新兴产业技术研究院 | Detection method for abnormal traffic flow of urban road based on coupled hidden markov model |
| CN105355067A (en) * | 2015-11-25 | 2016-02-24 | 南京莱斯信息技术股份有限公司 | Publishing method of intersection traffic information |
| CN107194490A (en) * | 2016-03-14 | 2017-09-22 | 商业对象软件有限公司 | Predict modeling optimization |
| CN107292386A (en) * | 2016-04-11 | 2017-10-24 | 福特全球技术公司 | Detected using the rainwater of the view-based access control model of deep learning |
| CN105741565A (en) * | 2016-04-21 | 2016-07-06 | 正元地理信息有限责任公司 | Method for automatic determining accident and giving off alarm based on monitoring video |
| CN105761500A (en) * | 2016-05-10 | 2016-07-13 | 腾讯科技（深圳）有限公司 | Traffic accident handling method and traffic accident handling device |
| CN105761500B (en) * | 2016-05-10 | 2019-02-22 | 腾讯科技（深圳）有限公司 | Traffic accident treatment method and traffic accident treatment device |
| CN107590995A (en) * | 2016-07-08 | 2018-01-16 | 沃尔沃汽车公司 | Include the method and system of the database of the influence traffic events of report for safeguarding |
| CN105976616A (en) * | 2016-07-29 | 2016-09-28 | 林玉峰 | Analysis method of traffic event monitoring model based on traffic data |
| CN106251663A (en) * | 2016-09-14 | 2016-12-21 | 深圳市喜悦智慧数据有限公司 | A kind of intelligent traffic robot |
| CN106408940A (en) * | 2016-11-02 | 2017-02-15 | 南京慧尔视智能科技有限公司 | Microwave and video data fusion-based traffic detection method and device |
| CN110291477B (en) * | 2016-12-02 | 2022-08-16 | 嘉蒂克人工智能公司 | Vehicle control system and method of use |
| CN110291477A (en) * | 2016-12-02 | 2019-09-27 | 斯塔斯凯机器人公司 | Vehicle control system and application method |
| DE102017200153A1 (en) | 2017-01-09 | 2018-07-12 | Ford Global Technologies, Llc | Method for detecting traffic situations |
| US11830251B2 (en) | 2017-01-13 | 2023-11-28 | Canon Kabushiki Kaisha | Video monitoring apparatus, method of controlling the same, computer-readable storage medium, and video monitoring system |
| US11048947B2 (en) | 2017-01-13 | 2021-06-29 | Canon Kabushiki Kaisha | Video monitoring apparatus, method of controlling the same, computer-readable storage medium, and video monitoring system |
| CN108307153A (en) * | 2017-01-13 | 2018-07-20 | 佳能株式会社 | Video monitoring devices and its control method, storage medium and video monitoring system |
| CN106851132A (en) * | 2017-01-20 | 2017-06-13 | 安徽达尔智能控制系统股份有限公司 | Traffic accident intelligent checking system |
| CN110494902A (en) * | 2017-02-03 | 2019-11-22 | 西门子交通有限责任公司 | For managing the system, apparatus and method of the traffic in geographical location |
| CN106919925A (en) * | 2017-03-07 | 2017-07-04 | 南京师范大学 | A kind of Ford Motor's detection method based on Wavelet Entropy Yu artificial neural network |
| CN107016859A (en) * | 2017-05-18 | 2017-08-04 | 肇庆学院 | A kind of crossing traffic management system, methods and applications |
| CN111066318A (en) * | 2017-09-15 | 2020-04-24 | 三菱电机株式会社 | Monitoring support device and monitoring support system |
| CN107845262A (en) * | 2017-11-09 | 2018-03-27 | 山东浪潮云服务信息科技有限公司 | A kind of Predictive Methods of Road Accidents and device |
| CN107665583B (en) * | 2017-11-20 | 2021-07-02 | 浩鲸云计算科技股份有限公司 | Method for calculating lane saturation flow rate under different weather conditions |
| CN107665583A (en) * | 2017-11-20 | 2018-02-06 | 中兴软创科技股份有限公司 | The computational methods of track saturation volume rate under the conditions of a kind of different weather |
| CN108304773A (en) * | 2017-12-25 | 2018-07-20 | 广州市高科通信技术股份有限公司 | A kind of vehicle density analysis method, device, electronic equipment and storage medium based on wavelet transformation |
| WO2019127651A1 (en) * | 2017-12-30 | 2019-07-04 | 惠州学院 | Method and system thereof for identifying malicious video |
| CN108198423A (en) * | 2018-01-24 | 2018-06-22 | 哈尔滨工业大学 | A kind of protrusion Crash characteristics recognition methods of two displays signal control cross level-crossing |
| CN108198423B (en) * | 2018-01-24 | 2020-09-08 | 哈尔滨工业大学 | Method for recognizing salient accident form of crossroad controlled by two display signals |
| CN108764042A (en) * | 2018-04-25 | 2018-11-06 | 深圳市科思创动科技有限公司 | A kind of exception traffic information recognition methods, device and terminal device |
| CN108764042B (en) * | 2018-04-25 | 2021-05-28 | 深圳市科思创动科技有限公司 | Abnormal road condition information identification method and device and terminal equipment |
| CN109033997A (en) * | 2018-07-02 | 2018-12-18 | 厦门快商通信息技术有限公司 | A kind of lumbering event detecting method and system |
| WO2020024447A1 (en) * | 2018-08-01 | 2020-02-06 | 平安科技（深圳）有限公司 | Traffic accident processing method and apparatus, and computer readable storage medium |
| CN109935078A (en) * | 2019-02-18 | 2019-06-25 | 深兰科技（上海）有限公司 | A kind of automatic evolution type traffic control system |
| CN110097755A (en) * | 2019-04-29 | 2019-08-06 | 东北大学 | Freeway traffic flow amount state identification method based on deep neural network |
| WO2020220439A1 (en) * | 2019-04-29 | 2020-11-05 | 东北大学 | Highway traffic flow state recognition method based on deep neural network |
| CN110415272A (en) * | 2019-07-08 | 2019-11-05 | 东华大学 | A kind of battery truck drives in the wrong direction detection system and method |
| CN110545359A (en) * | 2019-08-02 | 2019-12-06 | 国家计算机网络与信息安全管理中心 | Communication line feature extraction method, communication line identification method and device |
| CN111127876A (en) * | 2019-11-18 | 2020-05-08 | 腾讯科技（深圳）有限公司 | Information extraction method and device for Internet of vehicles |
| CN114005277A (en) * | 2019-11-18 | 2022-02-01 | 腾讯科技（深圳）有限公司 | Information extraction method and device of Internet of vehicles and readable medium |
| CN111127876B (en) * | 2019-11-18 | 2021-11-05 | 腾讯科技（深圳）有限公司 | Information extraction method and device for Internet of vehicles |
| CN110942011A (en) * | 2019-11-18 | 2020-03-31 | 上海极链网络科技有限公司 | Video event identification method, system, electronic equipment and medium |
| CN110942011B (en) * | 2019-11-18 | 2021-02-02 | 上海极链网络科技有限公司 | Video event identification method, system, electronic equipment and medium |
| CN110942629A (en) * | 2019-11-29 | 2020-03-31 | 中核第四研究设计工程有限公司 | Road traffic accident management method and device and terminal equipment |
| CN111145558A (en) * | 2019-12-27 | 2020-05-12 | 北京容联易通信息技术有限公司 | Illegal behavior identification method based on high-point video monitoring |
| CN111192456A (en) * | 2020-01-14 | 2020-05-22 | 泉州市益典信息科技有限公司 | Road traffic operation situation multi-time scale prediction method |
| CN112164223B (en) * | 2020-02-27 | 2022-04-29 | 浙江恒隆智慧科技集团有限公司 | Intelligent traffic information processing method and device based on cloud platform |
| CN112164223A (en) * | 2020-02-27 | 2021-01-01 | 周国霞 | Intelligent traffic information processing method and device based on cloud platform |
| CN111882861B (en) * | 2020-06-06 | 2021-08-24 | 浙江工业大学 | Online traffic incident perception system based on edge cloud fusion |
| CN111882861A (en) * | 2020-06-06 | 2020-11-03 | 浙江工业大学 | Online traffic incident perception system based on edge cloud fusion |
| CN111768625A (en) * | 2020-07-01 | 2020-10-13 | 中国计量大学 | Traffic road event prediction method based on graph embedding |
| CN112185110A (en) * | 2020-09-17 | 2021-01-05 | 李志� | Remote monitoring method and device based on Internet |
| CN111932427A (en) * | 2020-09-24 | 2020-11-13 | 北京泰策科技有限公司 | Method and system for detecting emergent public security incident based on multi-mode data |
| CN111932427B (en) * | 2020-09-24 | 2021-01-26 | 北京泰策科技有限公司 | Method and system for detecting emergent public security incident based on multi-mode data |
| CN112634634A (en) * | 2020-11-24 | 2021-04-09 | 深圳技术大学 | Intelligent rod aid decision-making method, intelligent rod and system |
| JP7109528B2 (en) | 2020-12-23 | 2022-07-29 | 三菱電機株式会社 | VEHICLE IMAGE INFORMATION PROCESSING DEVICE, VEHICLE IMAGE INFORMATION TRANSMITTER AND IMAGE INFORMATION PROCESSING METHOD |
| JP2022099389A (en) * | 2020-12-23 | 2022-07-05 | 三菱電機株式会社 | In-vehicle image information processor, in-vehicle image information transmitter, and image information processing method |
| CN112637568A (en) * | 2020-12-24 | 2021-04-09 | 中标慧安信息技术股份有限公司 | Distributed security monitoring method and system based on multi-node edge computing equipment |
| CN112885086A (en) * | 2021-01-18 | 2021-06-01 | 湖南省交通规划勘察设计院有限公司 | Sudden congestion judging system based on multi-source traffic big data |
| CN112906634B (en) * | 2021-03-18 | 2023-09-01 | 西北大学 | Video segment sequence prediction model establishment and sequence prediction method and system based on VSS |
| CN112906634A (en) * | 2021-03-18 | 2021-06-04 | 西北大学 | Video segment sequence prediction model establishment and sequence prediction method and system based on VSS |
| CN113299317A (en) * | 2021-05-14 | 2021-08-24 | 东南大学 | Intersection collision audio early warning evaluation method based on MFCC feature extraction |
| CN113706871A (en) * | 2021-09-13 | 2021-11-26 | 中交第一公路勘察设计研究院有限公司 | Multisource data fusion system in urban traffic big data processing |
| CN113537170A (en) * | 2021-09-16 | 2021-10-22 | 北京理工大学深圳汽车研究院(电动车辆国家工程实验室深圳研究院) | Intelligent traffic road condition monitoring method and computer readable storage medium |
| CN114446052A (en) * | 2022-01-17 | 2022-05-06 | 东南大学 | Multi-source data multi-stage fusion-based highway incident detection method and device |
| CN114446052B (en) * | 2022-01-17 | 2023-05-30 | 东南大学 | Expressway event detection method and device based on multi-level fusion of multi-source data |
| CN114827590A (en) * | 2022-06-22 | 2022-07-29 | 浙江力嘉电子科技有限公司 | Camera position distribution rationality analysis method and device and electronic equipment |
| CN114827590B (en) * | 2022-06-22 | 2022-09-13 | 浙江力嘉电子科技有限公司 | Camera position distribution rationality analysis method and device and electronic equipment |
| CN116266215A (en) * | 2022-08-01 | 2023-06-20 | 广西北投公路建设投资集团有限公司 | Geometric feature extraction method for pavement recessive disease area |
| CN116266215B (en) * | 2022-08-01 | 2024-01-26 | 广西北投公路建设投资集团有限公司 | Geometric feature extraction method for pavement recessive disease area |
| CN116518989A (en) * | 2023-07-05 | 2023-08-01 | 新唐信通(浙江)科技有限公司 | Method for vehicle navigation based on sound and thermal imaging |
| CN116518989B (en) * | 2023-07-05 | 2023-09-12 | 新唐信通(浙江)科技有限公司 | Method for vehicle navigation based on sound and thermal imaging |
| CN117935559A (en) * | 2024-03-07 | 2024-04-26 | 深圳市新创中天信息科技发展有限公司 | Traffic accident decision system based on multi-mode fusion perception technology |
| CN117935559B (en) * | 2024-03-07 | 2024-05-24 | 深圳市新创中天信息科技发展有限公司 | Traffic accident decision system based on multi-mode fusion perception technology |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101751782A (en) | Crossroad traffic event automatic detection system based on multi-source information fusion | |
| CN103258430B (en) | Road traveling time calculating and traffic road condition judging method and road traveling time calculating and traffic road condition judging device | |
| CN102231231A (en) | Area road network traffic safety situation early warning system and method thereof | |
| CN104392610B (en) | Expressway based on distributed video traffic events coverage dynamic monitoring and controlling method | |
| CN107705574A (en) | A kind of precisely full-automatic capturing system of quick road violation parking | |
| CN111798678A (en) | Urban traffic intelligent monitoring coordination management system based on big data | |
| CN106781458A (en) | A kind of traffic accident monitoring method and system | |
| CN101377884A (en) | Intelligent management system for highway traffic | |
| KR20100119476A (en) | An outomatic sensing system for traffic accident and method thereof | |
| CN102945603A (en) | Method for detecting traffic event and electronic police device | |
| TWI649729B (en) | System and method for automatically proving traffic violation vehicles | |
| CN111932876A (en) | Road surface induction system and method based on wide area radar | |
| CN111754786A (en) | System for identifying traffic vehicle passing events on highway | |
| CN106156890B (en) | Detection of passenger flow and Forecasting Methodology and its system in a kind of urban track traffic channel | |
| CN109035749A (en) | A kind of road thunders sound detection early warning application system extremely | |
| CN210721850U (en) | Highway road conditions intellectual detection system | |
| CN111932899A (en) | Traffic emergency control method and device based on traffic simulation | |
| CN102360524B (en) | Automatic detection and confirmation method of dangerous traffic flow characteristics of highway | |
| CN112669596A (en) | Traffic safety situation distinguishing system and method based on big data | |
| Ravish et al. | Intelligent traffic violation detection | |
| Assalam et al. | The Effectiveness Analysis of Frontage Road on Jalan Margonda Raya, Depok | |
| CN110428617A (en) | A kind of traffic object recognition methods based on 5G Portable intelligent terminal and MEC | |
| TWI774984B (en) | Traffic incident detection system and method | |
| CN113128847A (en) | Entrance ramp real-time risk early warning system and method based on laser radar | |
| Hourdos | Crash prone traffic flow dynamics: Identification and real-time detection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Open date: 20100623 |