CN101848377B - Device and method for intelligent linkage of multi-video recording device based on cloud computing and mass video searching - Google Patents
Device and method for intelligent linkage of multi-video recording device based on cloud computing and mass video searching Download PDFInfo
- Publication number
- CN101848377B CN101848377B CN2010101831767A CN201010183176A CN101848377B CN 101848377 B CN101848377 B CN 101848377B CN 2010101831767 A CN2010101831767 A CN 2010101831767A CN 201010183176 A CN201010183176 A CN 201010183176A CN 101848377 B CN101848377 B CN 101848377B
- Authority
- CN
- China
- Prior art keywords
- moving
- target
- video
- input device
- information
- 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.)
- Expired - Fee Related
Links
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Closed-Circuit Television Systems (AREA)
Abstract
The invention relates to a device and a method for the intelligent linkage of a multi-video recording device based on cloud computing and mass video searching. The device comprises a video recording device, a moving target dividing and tracking device, an OCR device, a moving target feature extraction device, a moving target storage device, a searching device and a visual user operation interface. A moving target is divided and tracked by recording the video information and the geographic information of the moving target, then the features of the moving target are classified and extracted, the information of the features of the moving target is stored, and abnormal event information is displayed to a user by user searching. A human-in-circulation strategy is adopted, the problem of accurate tracking of a cross-video recording device is solved by using a mode of assisting an operator by a computer, and the problems of low human information processing speed, weak memory capacity and high labor cost are also solved.
Description
Technical field
The present invention relates to carry out the method and the device of target following at multiple-camera span video camera based on cloud computing and magnanimity video retrieval technology.
Background technology
Video monitoring is widely used in the security application of public place and critical facility etc. as a kind of very effective security control means.Because being limited in scope of single camera monitoring, various securities are used generally all need set up a large amount of video cameras in the place of being paid close attention to.For example, " safe city " project of various places often all will be set up tens thousand of to hundreds thousand of video cameras in the public place in a city.So a large amount of video camera twenty four hours monitor full time, the data volume of its acquisition is considerably beyond the accessible scope of manpower.So utilizing computer to carry out intelligent monitoring is exactly inevitable choice.Yet existing intelligent monitoring technology is generally stressed cutting apart in real time, following the tracks of and detect specific simple event moving-target.Because each cameras view distance, angle and colour balance, illumination possibly exist than big-difference, the precision of existingly cut apart in real time, tracking technique being striden the video camera tracking target between non-overlapped video camera is very low, is difficult in practical application, obtain satisfied effect.Present supervisory control system still mainly relies on manpower to solve the problems such as video camera tracking, playback of striding; But because video monitoring data is too huge; And video monitoring is too high to the requirement of artificial intelligence, so, very difficult through manual detection or playback than complicated incident.Certainly, because human cost is expensive, reaction speed is low, the contemplated way that detects all incidents, object in real time, follows the tracks of of conventional video security monitoring will not have feasibility.
Summary of the invention
Phenomenon based on above-mentioned existence; Technical problem to be solved by this invention provides a kind of tracking and device of man-machine combination; Utilize the magnanimity video retrieval technology; Take the strategy of " people is in circulation ", solve the accurate tracking problem of striding the video input device with computer assistance operator's mode.
For solving the problems of the technologies described above, technical scheme provided by the invention is a kind of device of the many videos input device intelligent linkage based on cloud computing and magnanimity video frequency searching, and it comprises:
The video input device, the video information that its typing monitored, and send the geographical location information at said video information and said video input device place to the moving-target segmenting device;
The moving-target segmenting device, it is partitioned into the moving target video information on every frame video, and sends said moving target video information to the tracking of maneuvering target device;
The tracking of maneuvering target device, it carries out interframe follows the tracks of, and the activity of each moving-target in said single video input device is together in series, and forms the moving target video information of a space-time uniformity; Send the Word message that OCR literal extraction element extracts said moving-target to simultaneously and with said moving target video information;
Moving-target signature analysis extraction element, it is analyzed classification and it is carried out Visual Feature Retrieval Process said moving target video information;
The moving-target storage device is integrated into the complete characterization information of this moving target through classification, visual signature and Word message with above-mentioned each moving-target, stores in the said moving-target storage device through the high dimensional indexing device;
Indexing unit; It also comprises the anomalous event checkout gear; Through said indexing unit the said moving target video information that each said video input device obtains is associated; Detect anomalous event according to said anomalous event checkout gear, and said anomalous event is stored in the said moving-target storage device;
Visual user interface; In conjunction with GIS geography information and said moving target video information; Be shown to the operator to possible anomalous event visually; And according to operator's feedback, the moving target video information that is monitored each said video input device is together in series, and tracking, the playback function of striding said video input device are provided.
For solving the problems of the technologies described above, the present invention also provide a kind of many videos input device intelligent linkage based on cloud computing and magnanimity video frequency searching method, it comprises the steps:
The geographical location information at a, video information that typing monitored and said video place;
B, on every frame video of institute's typing, be partitioned into the moving target video information;
C, carry out interframe to the said moving target video information that is partitioned into and follow the tracks of, and the activity of each moving target in said single video input device is together in series, form the moving target video information of a space-time uniformity; Send the Word message that OCR literal extraction element extracts said moving target to simultaneously and with said moving target video information;
D, said moving target video information is analyzed classification and it is carried out Visual Feature Retrieval Process;
E, through with the classification of above-mentioned each moving-target, the visual signature of extraction and the complete characterization information that Word message is integrated into this moving target, store in the said moving-target storage device through the high dimensional indexing device;
F, the said moving target video information that each said video input device is obtained associate, and detect anomalous event according to said anomalous event checkout gear, and said anomalous event is stored in the said moving-target storage device;
G, combination GIS geography information and said moving target video information; Be shown to the operator to possible anomalous event visually; And according to operator's feedback; The moving target video information that is monitored each said video input device is together in series, and tracking, the playback function of striding said video input device fast and accurately are provided.
The present invention is based on cloud computing technology and magnanimity video retrieval technology, take the strategy of " people is in circulation ", solve the accurate tracking problem of striding the video input device with computer assistance operator's mode.The present invention has designed special operation flow; Only require that computer carries out the tracking and the record of certain precision to the target in the single video input device; Allow computer tracking that certain error rate is arranged; Needs the time standby operator manual correction system mistake, computer provides the high tracking aid, thereby finally reaches at a high speed, high-precisionly stride the video camera accurate tracking.The present invention lets computer and people accomplish the work of oneself being good at separately; Make full use of the shortcoming that people and computer advantage separately remedies the other side; Both solved the low problem of computer tracking precision, solved again that people's information processing rate is slow, memory capability is weak and the high problem of cost of labor.
Description of drawings
Fig. 1 single unit system of the present invention constitutes block diagram
The apparatus structure flow chart of Fig. 2 specific embodiment of the invention
Fig. 3 inventor feature extraction key diagram
Fig. 4 car feature extraction of the present invention key diagram
Other feature extraction key diagrams of Fig. 5 the present invention
The visual user interface sketch map of Fig. 6 the present invention
Embodiment
Lift a preferred embodiment now and combine accompanying drawing that the present invention is specifically set forth:
The present invention is based on the device of many videos input device intelligent linkage of cloud computing and magnanimity video frequency searching, referring to Fig. 1:
The video input device, the video information that its typing monitored, and send the geographical location information at video information and video input device place to the moving-target segmenting device; This video input device is a video camera;
On every frame video, be partitioned into the moving target video information through the moving-target segmenting device, and send the moving target video information to the tracking of maneuvering target device;
The tracking of maneuvering target device, it carries out interframe follows the tracks of, and the activity of each moving target in single video input device is together in series, and forms the moving target video information of a space-time uniformity; Send the Word message that OCR literal extraction element extracts said moving target to simultaneously and with the moving target video information;
Moving-target signature analysis extraction element, it is analyzed classification and it is carried out Visual Feature Retrieval Process said moving target video information; Extract in order to carry out signature analysis to different moving-targets, moving-target signature analysis extraction element is subdivided into moving-target rough sort device, people's feature deriving means, car feature deriving means and other feature deriving means; Other characteristics are except people and the moving-target characteristic the Che; Through extracting to being directed against people, car and other characteristics thereof after the moving-target rough sort again;
The moving-target storage device is integrated into the complete characterization information of this moving target through classification, visual signature and Word message with above-mentioned each moving-target, stores in the said moving-target storage device through the high dimensional indexing device;
Indexing unit; It also comprises the anomalous event checkout gear; Through said indexing unit the said moving target video information that each said video input device obtains is associated; Detect anomalous event according to said anomalous event checkout gear, and said anomalous event is stored in the said moving-target storage device;
Visual user interface; In conjunction with GIS geography information and said moving target video information; Be shown to the operator to possible anomalous event visually; And according to operator's feedback, the moving target video information that is monitored each said video input device is together in series, and tracking, the playback function of striding said video input device are provided.
In order better to describe; Lift an embodiment again to the described apparatus of the present invention block diagram of Fig. 1; The present invention is set forth in more detail, and referring to Fig. 2, the video input device is video camera D2; Action through all moving-targets in the video camera D2 guarded region; And send the geographical coordinate D1 of moving-target video information that is monitored and video camera position to moving-target segmenting device S1, on every frame video, be partitioned into moving-target through moving-target segmenting device S1, and send the moving-target information that is partitioned into to tracking of maneuvering target device S2 and carry out interframe and follow the tracks of; Be together in series the activity of each moving-target in single camera, form the moving-target video data information of a space-time uniformity;
The moving-target video data information that then tracking of maneuvering target device S2 is traced into passes to target rough sort device S4; Also send it to OCR literal extraction element S3 simultaneously and extract Word message; Owing to Word message (like clothes, luggage, car plate etc.) all possibly occur on a lot of objects; And literal is the best trace information of very useful association, retrieval and description moving-target; And be printed on literal through regular meeting on the moving-target body of paying close attention to most in the security monitorings such as clothes, luggage, car, like the number-plate number, work mark etc.Extract the literal on the moving-target video information as one of characteristic of describing moving-target through OCR literal extractive technique and can greatly improve identification capability moving-target;
Moving-target rough sort device S4 carries out preliminary classification with received moving-target video information, and the moving-target classification is people, car and other three types of moving-targets; In security monitoring; The characteristic and the behavior difference that have highly significant between the objects such as the people who pays close attention to most, car, luggage; Utilizing the characteristic of same type to describe and follow the tracks of all these objects obviously is poor efficiency; To follow the tracks of efficient and accuracy in order improving, to utilize the significant otherness between these several type objects, can detect the people in the moving-target video information, car and other classified informations comparatively exactly through utilizing machine learning techniques; Preliminary characteristic according to the moving-target video information is classified, and extracting the different classes characteristic respectively according to classification then will greatly provide the identification capability of moving-target and the high efficiency of tracking; In order better to carry out feature extraction of all categories, after the classification, be respectively arranged with people's feature deriving means, car feature deriving means and other feature deriving means again, extract characteristic of all categories respectively; Wherein:
People's feature deriving means S5 extracts others characteristic of this type of people; Referring to Fig. 3; Because the people has significant head, trunk and leg three parts, is divided into a D6, trunk D7 and leg D8 three parts to target so at first cut apart S12 according to moving-target video information D5 through body part.And sign people's best visual signature is a face, extracts face characteristic so send into face characteristic extraction element S13 to the header data that is partitioned into, if target has people's face information preferably, then comes out corresponding human face and feature extraction thereof as clarification of objective.Certainly, in monitor video, often can't see people's face, so also labour contractor D6, trunk D7 and leg D8 each several part are sent into color characteristic extraction element S14, extract its color histogram respectively and describe its color characteristic.Like this, even in video, can't see people's face, also can come the most of incoherent targets of filtering with the color of cap or hair, clothes and trousers.Can roughly extract the face characteristic that is used for Tracking Recognition that to summarize the body configuration and the color characteristic of a trunk leg through face characteristic extraction element S13 and color characteristic extraction element S14;
Car feature deriving means S6 extracts the car characteristic; Referring to Fig. 4; The moving-target that is classified as " motor vehicle " is owing to the difference except color; Characteristics such as its lines, structure are its principal characters, and the SIFT characteristic is owing to keeping these information, so the SIFT characteristic also is extracted out the characteristic as motor-driven vehicle moving-target preferably; Carry out the car color characteristic according to moving-target video information D5 and extract S14 and SIFT feature extraction S15, extract other color characteristic of vehicle and SIFT characteristic; Because the type of car and model also are very Useful Informations of sign motor vehicle; Can be used to get rid of a large amount of uncorrelated targets; Certainly, be unusual difficulty through the concrete model that video monitoring image identifies motor vehicle, so the present invention is categorized as four types of passenger vehicle, lorry, car and container cars to motor vehicle; To improve identification capability to motor vehicle; After car feature deriving means S6 is extracted the car characteristic, through car disaggregated classification device S8 car is carried out disaggregated classification again,, can improve fast and effectively and follow the tracks of identification efficient through car disaggregated classification device S8;
Other feature deriving means S7 extracts the characteristic of other types, and referring to Fig. 5, except that people and motor vehicle, all other moving-targets are used as same type objects and carry out feature extraction.In security monitoring is used; This type objects mainly is luggage, handcart etc.; So this type objects is except characteristics such as color, lines, structures, its shape information also has certain value; So the present invention except extracting its color and SIFT characteristic, also extracts its shape facility as tracking characteristics to this type objects; According to moving-target video information D5 other classifications are carried out color characteristic and extract S14, SIFT feature extraction S15 and Shape Feature Extraction S16, extract color characteristic, SIFT characteristic and the shape facility thereof of other classifications respectively;
The moving-target Word message that classification, visual signature information and OCR literal extraction element thereof through moving-target that above-mentioned people, car, other feature deriving means are extracted extracts is integrated into the complete characterization information of this moving-target, stores among the moving-target storage device D3 through high dimensional indexing device S9;
Indexing unit S10; Also comprise an anomalous event checkout gear S11; It can associate the moving-target video information that different cameras obtains through indexing unit S10; The anomalous event that detection at any time is possible, and among the moving-target storage device D3 that stores testing result as the adeditive attribute of relevant moving-target into;
Visual user interface F1; In conjunction with GIS geographical coordinate information and moving-target information; Be shown to the operator to possible anomalous event visually; And according to operator's feedback, be together in series the moving-target information that different cameras monitored, provide and stride fast and accurately that video camera is followed the tracks of, playback function.
Because each video camera exposure parameter, installation site, photoenvironment etc. all have than big-difference, it is very difficult will realizing automatically striding accurately the video camera tracking of maneuvering target with present artificial intelligence and computer vision technique, and amount of calculation is huge.In view of this, the present invention does not attempt to realize the full automatic video camera tracking of maneuvering target of striding, but adopts the semi-automatic mode of operator's artificial assistance.The careful design work flow process of the present invention only lets the tracing process of operator's manual intervention system when being necessary, neither needs too many manual operation can greatly improve the purpose of tracking accuracy again thereby reach.Its course of work is following:
1) apparatus of the present invention are not striden the video camera tracking of maneuvering target and are calculated at ordinary times, and only in single camera, moving-target are followed the tracks of, and collect all moving-targets and characteristic storage thereof in the moving-target storage device, set up good index, so that subsequent access;
2) device of the present invention provides abundant retrieval method to supply the operator to retrieve all moving-targets.The operator can carry out integrated retrieval to moving-target according to target type, color, the information such as time, place of literal, people's face and appearance on it.In general, when the incident that merits attention took place, time, place can be in comparatively narrow scopes, even can unique definite video camera and even moving-target.If return results is not unique, the operator can be through checking more related data, and discovery needs the initial moving-target of tracking.
3), in the moving-target storage device, retrieve forward direction, back to candidate's moving-target in case the operator has confirmed the initial moving-target of following the tracks of, and device of the present invention will utilize the moving-target characteristic of storing in the moving-target storage device, comprise information such as its time of occurrence, place.
4) apparatus of the present invention are play the video pictures of the initial moving-target of A in visual user interface central authorities, and arrange afterwards the key frame to B and forward direction C candidate moving-target respectively on its left side and the right side.Arrangement position is according to its characteristic matching degree; With initial moving-target more matched candidate come position intermediate more; As shown in Figure 6; The operator selects a tracking direction D as required then, such as promptly beginning to play corresponding video by selected direction to, apparatus of the present invention according to time forward direction or back.In playing process, if finding certain candidate's moving-target, the operator is complementary, can directly click corresponding key frame.So; Apparatus of the present invention are put into bottommost to the key frame of in progress moving-target; The moving-target of selecting the operator is put into central authorities' broadcast, and both sides are listed in to candidate's moving-target in the forward direction of retrieving novel, back in the driven target storage of the moving-target that utilizes the operator to select again.The operator continues operation, can select the continuous matching moving-target, accomplishes whole tracing process.
In general; Use existing similarity retrieval technology; 5-10 candidate's moving-target all can comprise the moving-target of coupling in most cases, and the operator from 5-10 candidate's moving-target, to select a correct coupling be to be relatively easy to, only need check that usually key frame can uniquely confirm.Above workflow has fully combined computer and people's advantage separately, can stride the video camera tracking by the accurate moving-target of Rapid Realization.And above workflow is not needing operator's intervention at ordinary times, only striding the participation that just needs the operator when video camera is followed the tracks of moving-target really, has greatly reduced manipulation strength and computation complexity.
In addition; After using workflow proposed by the invention; If in single camera to the tracking of certain moving-target exist be interrupted also not serious; As long as add the moving-target storage device to separated moving-target as different moving-targets, in tracing process, apparatus of the present invention will be listed in it in candidate's moving-target through characteristic matching and let the operator select.So, use the present invention can also reduce requirement to tracking of maneuvering target precision in the single camera.Under complex environment comparatively, even the tracking of maneuvering target precision is not high yet in the single camera.So present tracking of maneuvering target technology is difficult to obtain extensive use.After using the present invention,, can adapt to the more applications environment because required precision reduces.
Because the present invention needs and will carry out centralized index, storage and retrieval from the magnanimity moving-target of a large amount of video cameras, its memory space, amount of calculation are all very huge, so the mode that conventional unit is handled can't realize.The present invention utilizes cloud computing platform to solve this problem.Cloud computing platform of the present invention comprises devices such as high dimensional indexing device, moving-target storage device, anomalous event checkout gear and indexing unit, realizes a large amount of storages and calculating based on this cloud computing platform.Utilize the magnanimity distributed storage function of cloud platform, bottom problems such as the consistency that the moving-target storage device can management data, reliability, concurrency, backup only need give cloud computing platform moving-target and characteristic thereof.High dimensional indexing device, anomalous event checkout gear and indexing unit then can directly read the moving-target characteristic through the cloud computing interface, and utilize the cloud computing platform interface concurrent to handle all moving-target data.Because cloud computing platform has guaranteed the high concurrency of reading and writing data, calculating, system will obtain the ability of MPP automatically, parallel storage and calculation element that need not developing special.
Claims (8)
1. device based on many videos input device intelligent linkage of cloud computing and magnanimity video frequency searching, it comprises:
The video input device, the video information that its typing monitored, and send the geographical location information at said video information and said video input device place to the moving-target segmenting device;
The moving-target segmenting device, it is partitioned into the moving target video information on every frame video, and sends said moving target video information to the tracking of maneuvering target device;
The tracking of maneuvering target device, it carries out interframe follows the tracks of, and the activity of each moving target in said single video input device is together in series, and forms the moving target video information of a space-time uniformity; Send the Word message that OCR literal extraction element extracts said moving target to simultaneously and with said moving target video information;
Moving-target signature analysis extraction element, it is analyzed classification and it is carried out Visual Feature Retrieval Process said moving target video information;
The moving-target storage device;
The high dimensional indexing device will be by the complete characterization information stores of this integrated moving target of classification, visual signature and the Word message of each moving-target in said moving-target storage device;
Indexing unit; It also comprises the anomalous event checkout gear; Through said indexing unit the said moving target video information that each said video input device obtains is associated; Detect anomalous event according to said anomalous event checkout gear, and said anomalous event is stored in the said moving-target storage device;
Visual user interface device; In conjunction with GIS geography information and said moving target video information; Be shown to the operator to possible anomalous event visually; And according to operator's feedback, the moving target video information that is monitored each said video input device is together in series, and tracking, the playback function of striding said video input device are provided.
2. the device of the many videos input device intelligent linkage based on cloud computing and magnanimity video frequency searching according to claim 1 is characterized in that said moving-target signature analysis extraction element comprises moving-target rough sort device, people's feature deriving means, car feature deriving means and other feature deriving means; Carry out the video features information extraction respectively through sending the sorted moving-target video information of said moving-target rough sort device to said people's feature deriving means, car feature deriving means and other feature deriving means.
3. the device of the many videos input device intelligent linkage based on cloud computing and magnanimity video frequency searching according to claim 2 is characterized in that including in the said car feature deriving means car disaggregated classification device.
4. according to the device of the arbitrary described many videos input device intelligent linkage based on cloud computing and magnanimity video frequency searching of claim 1 to 3, it is characterized in that said video input device is a video camera.
5. method based on many videos input device intelligent linkage of cloud computing and magnanimity video frequency searching, it comprises the steps:
The geographical location information at a, video information that typing monitored and said video place;
B, on every frame video of institute's typing, be partitioned into the moving target video information;
C, carry out interframe to the said moving target video information that is partitioned into and follow the tracks of, and the activity of each moving target in said single video input device is together in series, form the moving target video information of a space-time uniformity; Send the Word message that OCR literal extraction element extracts said moving target to simultaneously and with said moving target video information;
D, said moving target video information is analyzed classification and it is carried out Visual Feature Retrieval Process;
E, through with the classification of above-mentioned each moving-target, the visual signature of extraction and the complete characterization information that Word message is integrated into this moving target, store in the moving-target storage device through the high dimensional indexing device;
F, the said moving target video information that each said video input device is obtained associate, and detect anomalous event according to the anomalous event checkout gear, and said anomalous event is stored in the said moving-target storage device;
G, combination GIS geography information and said moving target video information; Be shown to the operator to possible anomalous event visually; And according to operator's feedback; The moving target video information that is monitored each said video input device is together in series, and tracking, the playback function of striding said video input device fast and accurately are provided.
6. the method for the many videos input device intelligent linkage based on cloud computing and magnanimity video frequency searching according to claim 5 is characterized in that Visual Feature Retrieval Process in the said d step comprises respectively the feature extraction to people's characteristic, car characteristic and other characteristics.
7. the method for the many videos input device intelligent linkage based on cloud computing and magnanimity video frequency searching according to claim 6 is characterized in that said car feature extraction is for carrying out carrying out the car disaggregated classification again after the car feature extraction according to classification earlier.
8. according to the method for the arbitrary described many videos input device intelligent linkage based on cloud computing and magnanimity video frequency searching of claim 5 to 7, it is characterized in that said video input device is a video camera.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010101831767A CN101848377B (en) | 2010-05-26 | 2010-05-26 | Device and method for intelligent linkage of multi-video recording device based on cloud computing and mass video searching |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010101831767A CN101848377B (en) | 2010-05-26 | 2010-05-26 | Device and method for intelligent linkage of multi-video recording device based on cloud computing and mass video searching |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101848377A CN101848377A (en) | 2010-09-29 |
| CN101848377B true CN101848377B (en) | 2012-06-06 |
Family
ID=42772811
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2010101831767A Expired - Fee Related CN101848377B (en) | 2010-05-26 | 2010-05-26 | Device and method for intelligent linkage of multi-video recording device based on cloud computing and mass video searching |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101848377B (en) |
Families Citing this family (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102752574B (en) * | 2011-04-18 | 2015-01-28 | 中兴通讯股份有限公司 | Video monitoring system and method |
| TWI448975B (en) * | 2011-07-15 | 2014-08-11 | Chunghwa Telecom Co Ltd | Dispersing-type algorithm system applicable to image monitoring platform |
| CN102355572A (en) * | 2011-09-29 | 2012-02-15 | 南京云创存储科技有限公司 | Large high-definition video monitoring system based on cloud computing |
| 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 |
| CN103198110A (en) * | 2013-03-28 | 2013-07-10 | 广州中国科学院软件应用技术研究所 | Method and system for rapid video data characteristic retrieval |
| CN103402079A (en) * | 2013-07-31 | 2013-11-20 | 广东威创视讯科技股份有限公司 | VW (Video Wall)-based intelligent tracking, storing and displaying method for abnormal events |
| CN103761295B (en) * | 2014-01-16 | 2017-01-11 | 北京雅昌文化发展有限公司 | Automatic picture classification based customized feature extraction method for art pictures |
| CN104023205B (en) * | 2014-05-27 | 2018-04-27 | 北京华兴宏视技术发展有限公司 | Intelligent safety defense monitoring system |
| CN104601964B (en) * | 2015-02-06 | 2018-09-21 | 武汉大学 | Pedestrian target tracking and system in non-overlapping across the video camera room of the ken |
| CN108664478B (en) * | 2017-03-27 | 2021-07-20 | 华为技术有限公司 | Target object retrieval method and device |
| CN107871019B (en) * | 2017-12-01 | 2020-09-15 | 湖北微模式科技发展有限公司 | Man-vehicle association search method and device |
| CN108600698A (en) * | 2018-04-24 | 2018-09-28 | 东莞理工学院 | A kind of safety defense monitoring system and its method based on cloud platform |
| CN112528716B (en) * | 2019-09-19 | 2024-01-05 | 杭州海康威视数字技术股份有限公司 | Event information acquisition method and device |
| CN111836014B (en) * | 2020-07-14 | 2021-07-09 | 北京仁光科技有限公司 | Information security display system |
| CN113570574B (en) * | 2021-07-28 | 2023-12-01 | 北京精英系统科技有限公司 | Scene feature detection device, scene feature search device and scene feature search method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201204665Y (en) * | 2008-03-18 | 2009-03-04 | 浙江树人大学 | Apparatus for detecting and tracking multi-source movable target |
| CN101572804A (en) * | 2009-03-30 | 2009-11-04 | 浙江大学 | Multi-camera intelligent control method and device |
| CN101576926A (en) * | 2009-06-04 | 2009-11-11 | 浙江大学 | Monitor video searching method based on geographic information system |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7746378B2 (en) * | 2004-10-12 | 2010-06-29 | International Business Machines Corporation | Video analysis, archiving and alerting methods and apparatus for a distributed, modular and extensible video surveillance system |
-
2010
- 2010-05-26 CN CN2010101831767A patent/CN101848377B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201204665Y (en) * | 2008-03-18 | 2009-03-04 | 浙江树人大学 | Apparatus for detecting and tracking multi-source movable target |
| CN101572804A (en) * | 2009-03-30 | 2009-11-04 | 浙江大学 | Multi-camera intelligent control method and device |
| CN101576926A (en) * | 2009-06-04 | 2009-11-11 | 浙江大学 | Monitor video searching method based on geographic information system |
Non-Patent Citations (3)
| Title |
|---|
| 张莉.多摄像机人体跟踪技术的研究.《中国博士学位论文全文数据库 信息科技辑》.2009,(第7期),全文. * |
| 曲琳.分布式智能视觉监控行为分析及语义检索技术研究.《中国博士学位论文全文数据库 信息科技辑》.2009,(第11期),全文. * |
| 李志华.智能视频监控系统目标跟踪与分类算法研究.《中国博士学位论文全文数据库 信息科技辑》.2009,(第11期),全文. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101848377A (en) | 2010-09-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101848377B (en) | Device and method for intelligent linkage of multi-video recording device based on cloud computing and mass video searching | |
| CN108388888B (en) | Vehicle identification method and device and storage medium | |
| CN108596277B (en) | Vehicle identity recognition method and device and storage medium | |
| Cai et al. | Deep learning-based video system for accurate and real-time parking measurement | |
| US8401252B2 (en) | Video retrieval system for human face content | |
| CN110751022A (en) | Urban pet activity track monitoring method based on image recognition and related equipment | |
| CN101840422A (en) | Intelligent video retrieval system and method based on target characteristic and alarm behavior | |
| CN104574954B (en) | A kind of vehicle auditing method, control device and system based on free streaming system | |
| CN109740420A (en) | Vehicle illegal recognition methods and Related product | |
| TW201227629A (en) | Method, system and computer program product for reconstructing moving path of vehicle | |
| KR101645959B1 (en) | The Apparatus and Method for Tracking Objects Based on Multiple Overhead Cameras and a Site Map | |
| CN107103615A (en) | A kind of monitor video target lock-on tracing system and track lock method | |
| CN111353369A (en) | Application method and system of high-order video of urban roadside parking in assisting criminal investigation | |
| CN108710827A (en) | A kind of micro- police service inspection in community and information automatic analysis system and method | |
| CN114937293B (en) | GIS-based agricultural service management method and system | |
| CN104952273A (en) | Warning method, device and system aiming at vehicle behaviors | |
| CN111539257B (en) | Person re-identification method, device and storage medium | |
| CN112836683A (en) | License plate recognition method, device, equipment and medium for portable camera equipment | |
| CN113256731A (en) | Target detection method and device based on monocular vision | |
| Jiao et al. | Traffic behavior recognition from traffic videos under occlusion condition: a Kalman filter approach | |
| Balali et al. | Video-based highway asset recognition and 3D localization | |
| CN110889971A (en) | Intelligent vehicle finding method and device for parking lot | |
| CN114581990A (en) | Intelligent running test method and device | |
| CN112268554A (en) | Regional range loitering detection method and system based on path trajectory analysis | |
| CN113128540B (en) | Method and device for detecting vehicle theft behavior of non-motor vehicle and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| DD01 | Delivery of document by public notice |
Addressee: SUZHOU ANGERAY ELECTRONIC TECHNOLOGY Co.,Ltd. Document name: Notification to Pay the Fees |
|
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20120606 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |