CN102663069A - Method and device for retrieving videos - Google Patents
Method and device for retrieving videos Download PDFInfo
- Publication number
- CN102663069A CN102663069A CN2012100933457A CN201210093345A CN102663069A CN 102663069 A CN102663069 A CN 102663069A CN 2012100933457 A CN2012100933457 A CN 2012100933457A CN 201210093345 A CN201210093345 A CN 201210093345A CN 102663069 A CN102663069 A CN 102663069A
- Authority
- CN
- China
- Prior art keywords
- video
- incident
- detected
- sampled
- module
- 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
Abstract
The invention relates to a method and a device for retrieving videos. Video files are sampled at sampling intervals set in advance, and whether events occur or not is retrieved in sampled images. Compared with a traditional method for retrieving video files frame by frame, the method has the advantages that the quantity of the images to be retrieved is reduced, accordingly retrieving speed is greatly increased, workload is reduced, labor and materials are saved, development of follow-up operation is facilitated, and the method and the device are particularly applicable to conditions with long videos and more videos.
Description
Technical field
The present invention relates to video search technique area, particularly relate to a kind of video retrieval method and device.
Background technology
In one section video, search a certain incident and whether occur having very important meaning; As search criminal and whether on a certain section expressway, pass through; Whether search certain car in certain period certain highway section of process, confirm similar incidents as soon as possible, chasing criminal; Use manpower and material resources sparingly, guarantee that society and traffic safety etc. are significant.For one section video, traditional search method is that this search method significant disadvantage is the length that expends time in, when particularly video file is bigger and many by the frame retrieval.As in the robber with guns's case of Nanjing,, expended several days time of numerous police strength in order to search two sections videos about 10s.Therefore, it is necessary seeking a kind of fast searching method real.
Summary of the invention
The present invention proposes a kind of video retrieval method and device, whether take place the needed time to accelerate the speed of video frequency searching, to reduce searching certain incident.
A kind of video retrieval method comprises step:
According to the predefined SI, video file is sampled;
Image to sampling is retrieved, and judges whether incident to be detected takes place.
Therein in preferred embodiment, according to the predefined SI, before video file sampled, also comprise step in said step:
According to the duration of incident to be detected, set the SI, the SI that sets is smaller or equal to the duration of incident to be detected.
Therein in preferred embodiment, according to the predefined SI, before video file sampled, also comprise step in said step:
Set the initial SI,
When judgement learns that incident to be detected does not take place, according to predetermined gradient the initial SI to be turned down, and again video file is sampled according to the SI after turning down, circulation successively takes place until retrieving incident to be detected.
In preferred embodiment, the SI is a tolerance with time or frame number therein.
A kind of video frequency searching device comprises:
SI is provided with module, is used to preestablish the SI;
The video sampling module is used for according to the predefined SI video file being sampled;
Event checking module is used for the image of said video sampling module samples is retrieved, and judges whether incident to be detected takes place.
In preferred embodiment, the said SI is provided with module and also is used for setting the SI according to the duration of incident to be detected therein, and the SI that sets is smaller or equal to the duration of incident to be detected.
Therein in preferred embodiment, the said SI be provided with the predefined SI of module be initial sampled at interval,
The said SI is provided with module and also is used for when said event checking module detects incident to be detected and do not take place; According to predetermined gradient the initial SI is turned down; Said video sampling module also is used for again video file being sampled according to the SI after turning down, and said event checking module also is used for the image of said video sampling module resampling is retrieved again, judges once more whether incident to be detected takes place; Circulation successively takes place until retrieving incident to be detected.
In preferred embodiment, the SI is a tolerance with time or frame number therein.
Video retrieval method of the present invention and device were sampled to video file according to the predefined SI, and whether retrieval event takes place from the image of sampling again; Compare traditional method that video file is retrieved by frame, owing to reduced amount of images to be retrieved, thus improved retrieval rate greatly; Reduced workload; Saved manpower and materials, provided convenience, be particularly useful for the long and a fairly large number of situation of video length for carrying out of later stage work.
Description of drawings
Fig. 1 is the schematic flow sheet of video retrieval method of the present invention;
Fig. 2 is the structural representation of video frequency searching device of the present invention.
Embodiment
Video retrieval method of the present invention and device are based on following viewpoint:
1, objective material can not take place to move in wink, and a section takes place for it moves and need certain hour, might as well be T the time of establishing.
2, video file is continuous, is not that some mixed and disorderly meaningless pictures slap together.
For example, the surveillance video of certain charging section, the situation in camera ability effective monitoring 20m highway section, the video file frame per second that camera generates is 20 frames/s.At the charge crossing, the maximal rate that the speed of a motor vehicle allows is 5m/s.Then a car needs 4s at least through this crossing.
Need to suppose certain car of retrieval whether to pass through at this crossing.If the search method by routine travels through and searches, then need per second to retrieve 20 two field pictures, when video file was long, number of image frames to be retrieved was numerous, needed the time and efforts of labor.Existing problem is to search owing to all Frames to video travel through in essence, and amount of images to be found causes more greatly, so the present invention starts with raising video frequency searching speed from the angle that reduces amount of images to be found.Below in conjunction with accompanying drawing and the present invention of embodiment illustrated in detail.
Video retrieval method of the present invention, as shown in Figure 1, comprise step:
Step S1, according to the predefined SI, video file is sampled;
Step S2, to the sampling image retrieve, judge whether incident to be detected takes place.
The example of preceding text when adopting method of the present invention to retrieve, from numerous images, is sampled by the predetermined SI, according to conventional method the image of sampling is retrieved again, owing to reduced picture number to be retrieved, required time shortens greatly.Wherein, the work of retrieval can be carried out after sampling is accomplished to the whole video file, also can promptly retrieve a frame by every sampling one frame, after retrieving incident to be detected, then can stop sampling and retrieval, has practiced thrift the time once more.
For the SI, can with frame number or time as tolerance.When understanding the duration of incident to be detected, can the SI be set at the interval that is not more than incident duration to be detected, as above routine, the SI can be set to 3s, and whether two two field pictures of then in 4s, sampling at most can decision event take place.Suppose to have special processor whether one two field picture decision event is taken place, if can accomplish 30 detections in the processor 1s.Then adopt conventional search method, the processor per second can only be handled the video of 30/20=1.5s time span.And adopt search method of the present invention, can handle the video of 30*3=90s time span in the processor 1s.As long as efficient has improved 60 times, just needs the processing time in 1 year originally, now less than a week.
It is following to detect the process whether something take place: as whether detecting target vehicle through certain crossing, A represents target vehicle with license plate number.Each sampled images is carried out the picture analysis, extract license plate number B.Judge whether A equates with B, if equate that then incident takes place, otherwise incident does not take place.
If whether detect the someone, represent this people with photo P through certain crossing.Each sampled images is carried out the picture analysis, carry photo characteristic B.Judge the similarity of P and B,, think that then this incident takes place, otherwise incident does not take place if similarity is higher.
For the SI, when not understanding the duration of incident to be detected, can rule of thumb set an initial SI earlier; Through one take turns the sampling with the retrieval, do not retrieve incident to be detected, then can the initial SI be turned down; Again sample and retrieve; If do not retrieve incident to be detected yet, then turn down the SI once more, until retrieving incident to be detected.
Video frequency searching device of the present invention is the device corresponding with said method, and is as shown in Figure 2, comprising:
SI is provided with module, is used to preestablish the SI;
The video sampling module is used for according to the predefined SI video file being sampled;
Event checking module is used for the image of said video sampling module samples is retrieved, and judges whether incident to be detected takes place.
Can know that by above description and Fig. 2 the SI is provided with module, the video sampling module links to each other with event checking module successively, has formed this device.
As a preferred embodiment, the said SI is provided with module and also is used for setting the SI according to the duration of incident to be detected, and the SI that sets is smaller or equal to the duration of incident to be detected.
As a preferred embodiment, it is the initial sampled interval that the said SI is provided with the predefined SI of module,
The said SI is provided with module and also is used for when said event checking module detects incident to be detected and do not take place; According to predetermined gradient the initial SI is turned down; Said video sampling module also is used for again video file being sampled according to the SI after turning down, and said event checking module also is used for the image of said video sampling module resampling is retrieved again, judges once more whether incident to be detected takes place; Circulation successively takes place until retrieving incident to be detected.
As a preferred embodiment, the SI is a tolerance with time or frame number.
Video retrieval method of the present invention is used in combination with device, can realizes that the present invention accelerates the purpose of video frequency searching speed.
The above embodiment has only expressed several kinds of embodiments of the present invention, and it describes comparatively concrete and detailed, but can not therefore be interpreted as the restriction to claim of the present invention.Should be pointed out that for the person of ordinary skill of the art under the prerequisite that does not break away from the present invention's design, can also make some distortion and improvement, these all belong to protection scope of the present invention.Therefore, the protection domain of patent of the present invention should be as the criterion with accompanying claims.
Claims (8)
1. a video retrieval method is characterized in that, comprises step:
According to the predefined SI, video file is sampled;
Image to sampling is retrieved, and judges whether incident to be detected takes place.
2. video retrieval method according to claim 1 is characterized in that, according to the predefined SI, before video file sampled, also comprises step in said step:
According to the duration of incident to be detected, set the SI, the SI that sets is smaller or equal to the duration of incident to be detected.
3. video retrieval method according to claim 1 is characterized in that, according to the predefined SI, before video file sampled, also comprises step in said step:
Set the initial SI,
When judgement learns that incident to be detected does not take place, according to predetermined gradient the initial SI to be turned down, and again video file is sampled according to the SI after turning down, circulation successively takes place until retrieving incident to be detected.
4. according to claim 1 or 2 or 3 described video retrieval methods, it is characterized in that the SI is a tolerance with time or frame number.
5. a video frequency searching device is characterized in that, comprising:
SI is provided with module, is used to preestablish the SI;
The video sampling module is used for according to the predefined SI video file being sampled;
Event checking module is used for the image of said video sampling module samples is retrieved, and judges whether incident to be detected takes place.
6. video frequency searching device according to claim 5 is characterized in that, the said SI is provided with module and also is used for setting the SI according to the duration of incident to be detected, and the SI that sets is smaller or equal to the duration of incident to be detected.
7. video frequency searching device according to claim 5 is characterized in that,
It is the initial sampled interval that the said SI is provided with the predefined SI of module,
The said SI is provided with module and also is used for when said event checking module detects incident to be detected and do not take place; According to predetermined gradient the initial SI is turned down; Said video sampling module also is used for again video file being sampled according to the SI after turning down, and said event checking module also is used for the image of said video sampling module resampling is retrieved again, judges once more whether incident to be detected takes place; Circulation successively takes place until retrieving incident to be detected.
8. according to claim 5 or 6 or 7 described video frequency searching devices, it is characterized in that the SI is a tolerance with time or frame number.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012100933457A CN102663069A (en) | 2012-03-31 | 2012-03-31 | Method and device for retrieving videos |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012100933457A CN102663069A (en) | 2012-03-31 | 2012-03-31 | Method and device for retrieving videos |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102663069A true CN102663069A (en) | 2012-09-12 |
Family
ID=46772560
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2012100933457A Pending CN102663069A (en) | 2012-03-31 | 2012-03-31 | Method and device for retrieving videos |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102663069A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106598980A (en) * | 2015-10-16 | 2017-04-26 | 阿里巴巴集团控股有限公司 | Information push method and device |
| CN110598048A (en) * | 2018-05-25 | 2019-12-20 | 北京中科寒武纪科技有限公司 | Video retrieval method and video retrieval mapping relation generation method and device |
| CN112163123A (en) * | 2020-11-10 | 2021-01-01 | 中国科学院长春光学精密机械与物理研究所 | Event extraction method and device, electronic equipment and storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101398854A (en) * | 2008-10-24 | 2009-04-01 | 清华大学 | Video fragment searching method and system |
| CN102324024A (en) * | 2011-09-06 | 2012-01-18 | 苏州科雷芯电子科技有限公司 | Airport passenger recognition and positioning method and system based on target tracking technique |
-
2012
- 2012-03-31 CN CN2012100933457A patent/CN102663069A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101398854A (en) * | 2008-10-24 | 2009-04-01 | 清华大学 | Video fragment searching method and system |
| CN102324024A (en) * | 2011-09-06 | 2012-01-18 | 苏州科雷芯电子科技有限公司 | Airport passenger recognition and positioning method and system based on target tracking technique |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106598980A (en) * | 2015-10-16 | 2017-04-26 | 阿里巴巴集团控股有限公司 | Information push method and device |
| CN106598980B (en) * | 2015-10-16 | 2020-03-10 | 阿里巴巴集团控股有限公司 | Information pushing method and device |
| CN110598048A (en) * | 2018-05-25 | 2019-12-20 | 北京中科寒武纪科技有限公司 | Video retrieval method and video retrieval mapping relation generation method and device |
| CN112163123A (en) * | 2020-11-10 | 2021-01-01 | 中国科学院长春光学精密机械与物理研究所 | Event extraction method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Lin et al. | A YOLO-based traffic counting system | |
| CN202077142U (en) | Vehicle-mounted intelligent video detecting and analyzing system | |
| US6442474B1 (en) | Vision-based method and apparatus for monitoring vehicular traffic events | |
| CN103714697B (en) | A kind of method of recognition and tracking criminal vehicle | |
| CN105702048A (en) | Automobile-data-recorder-based illegal lane occupation identification system and method for automobile on highway | |
| CN111353369A (en) | Application method and system of high-order video of urban roadside parking in assisting criminal investigation | |
| CN110738857A (en) | vehicle violation evidence obtaining method, device and equipment | |
| CN105321352A (en) | A motor vehicle license plate blocking violation detection and evidence obtaining method | |
| CN102663069A (en) | Method and device for retrieving videos | |
| CN104809887A (en) | Detecting method for vehicle converse running on expressway and automatic alarm device | |
| CN101593418A (en) | Method for associative search of suspected vehicles | |
| RU2011153165A (en) | METHOD FOR Capturing IMAGES OF VEHICLES | |
| CN104537840A (en) | System for detecting illegally operated taxis | |
| CN107369324A (en) | A kind of overspeed of vehicle on highway monitoring system | |
| CN103279481B (en) | Intelligent Skynet intelligence image investigation system | |
| CN104616236A (en) | Police handheld multifunctional device | |
| CN110390288B (en) | Target intelligent searching, positioning and tracking system and method based on computer vision | |
| US20210042940A1 (en) | Digital twin monitoring systems and methods | |
| CN113934889A (en) | Method and system for tracking suspicious people according to pedestrians based on space-time trajectory | |
| CN103824454A (en) | Multi-zone illegality automatic capture method based on video event detection | |
| CN109326126A (en) | A kind of fixed point screens the method and monitoring system of illegal vehicle | |
| Jung et al. | Smart parking management system using AI | |
| CN109003457A (en) | It is a kind of to record the illegal method and device for occupying Emergency Vehicle Lane behavior of more motor vehicles | |
| Balbin et al. | Cloud based color coding scheme violator plate detection through character recognition using image processing | |
| FR2936627A1 (en) | METHOD OF OPTIMIZING THE SEARCHING OF A SCENE FROM A STREAM OF IMAGES ARCHIVE IN A VIDEO DATABASE. |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20120912 |