CN104680521B - A kind of improved background modeling and foreground detection method - Google Patents
A kind of improved background modeling and foreground detection method Download PDFInfo
- Publication number
- CN104680521B CN104680521B CN201510065105.XA CN201510065105A CN104680521B CN 104680521 B CN104680521 B CN 104680521B CN 201510065105 A CN201510065105 A CN 201510065105A CN 104680521 B CN104680521 B CN 104680521B
- Authority
- CN
- China
- Prior art keywords
- background
- pixel
- point
- dot
- foreground
- 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
Landscapes
- Image Analysis (AREA)
Abstract
The present invention proposes a kind of improved background modeling and foreground detection method, initially set up background model, then judge that each pixel of present frame belongs to background dot or foreground point, then background collection is updated, finally each picture frame is judged pixel-by-pixel using above-mentioned flow, the bianry image of prospect and background separation is finally given, so as to split foreground moving object from monitor video.The method of the present invention overcomes some shortcomings of conventional target detection method, improves detection adaptability, stability and real-time etc..
Description
Technical field
The present invention relates to image intelligent detection, identification technology field, more particularly to a kind of background modeling and foreground target inspection
The method of survey.
Background technology
In intelligent monitoring, moving object detection is significant in the daily production and living of people, and the purpose is to will
Moving foreground object is split from background.But in a practical situation, monitored environment is filled with many uncertain factors, the back of the body
Scape is often highly susceptible to various influences, such as illumination variation, dynamic background, shade, partial occlusion etc. influence, and these are unfavorable
Factor brings very big challenge to anaphase movement Target detection and identification.Therefore, to various complex situations analysis modelings, try one's best
Reduce extraneous complex environment factor to adversely affect to caused by target detection recognition result, reduce erroneous judgement, improve the adaptation of algorithm
Property, robustness etc. just become particularly significant.
In foreground detection field, main flow algorithm mainly has frame difference method, optical flow method and background subtraction method etc..Frame difference method have compared with
Good real-time, its background is not assembled, and renewal speed is fast, algorithm is easy, amount of calculation is small.The deficiency of algorithm is to make an uproar to environment
Sound is more sensitive, and the selection of threshold value is quite crucial, selects the too low noise being not enough in oppressive image, too high, has neglected image
In useful change.It is possible to occur to be abstracted in target internal, can not comprehensively extracts action target.Optical flow method need not be advance
Obtain scene information, it becomes possible to detect target, but noise, multiple light courcess, shade and the factor such as block can be to the calculating of optical flow field
Convergence impacts；And optical flow method calculates complexity, it is difficult to reaches the purpose handled in real time.Background subtraction method must have background
Image, reference background image is subtracted using present image, and background image must be the change with illumination or external environment condition
And real-time update, therefore the key of background subtraction method is background modeling and its renewal, the quality of segmentation result is largely
Dependent on selected background modeling method.
Due to the property complicated and changeable of actual environment, many traditional background modeling methods have its limitation, therefore many
The method for building up for being proved to the background model of robust is suggested.These methods not only largely improve Detection results,
And improve adaptability, stability and calculating speed of algorithm etc..
The content of the invention
In order to solve the problems of the prior art, the invention provides a kind of improved background modeling and foreground detection side
Method, foreground moving object is split from monitor video.Some shortcomings of conventional target detection method are the method overcome, are improved
Detect adaptability, stability and real-time etc..
The present invention is achieved through the following technical solutions：
A kind of improved background modeling and foreground detection method, it is characterised in that：It the described method comprises the following steps：
Step A1：N two field pictures before selection, calculate the average I of each pixel positionN(x, y), established for each pixel
Respective background collection, the background collection element of each pixel by from preceding N two field pictures with the pixel value of the pixel same position
Composition is randomly selected, each background is concentrated and includes M element.
Step A2：Judge that each pixel of present frame belongs to background dot or foreground point, if current pixel background is concentrated
Number of the similarity less than a given threshold th1 of element and the pixel be more than T1, then the point is judged as background dot, otherwise
It is determined as foreground point.
Step A3：Background collection is not updated if certain pixel is detected as foreground point, updates this if background dot is detected as
The background collection of position pixel.
Further, the step A3 specifically includes following steps：
Step B1：If certain point is judged to background dot, the speed of pixel change is first determined whether, that is, before finding out present frame
Continuous K two field pictures in the position pixel be judged to the pixel of background dot, K<N, and count these background dots and average IN(x,y)
Difference, if to exceed T2 individual for the number of point of the absolute difference more than given threshold th2, can determine whether that the current background dot belongs to change
Change fast background dot, otherwise current background point, which belongs to, changes slow background dot.
Step B2：If the pixel for being detected as background dot, which belongs to background, changes slower point, its back of the body is replaced with the pixel value
Scape concentrates some element randomly selected；If belonging to background changes faster point, replace its background with the pixel value and concentrate
The m element randomly selected.
Brief description of the drawings
Fig. 1 is the flow chart of the background modeling and foreground detection method of the present invention.
Embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, it is right below in conjunction with drawings and Examples
The present invention is further elaborated.It should be appreciated that specific embodiment described herein is only to explain the present invention, not
For limiting the present invention.
Accompanying drawing 1 is the flow chart of the background modeling and foreground detection method of the present invention.
Initially set up background model：N two field pictures before selection, respective background collection, each pixel are established for each pixel
Background collection element by being formed from preceding N two field pictures with being randomly selected in the pixel value of the pixel same position, each background
Concentration includes M element.
Then judge that each pixel of present frame belongs to background dot or foreground point：First judge that current pixel background is concentrated
Element and the pixel similarity (Euclidean distance i.e. between pixel, such as between 2 points (r1, g1, b1) and (r2, g2, b2)
Euclidean distance beIf the pixel concentrates the similar of each element to its background
Number of the degree less than a given threshold th1 is more than T1, then the point is judged as background dot, is otherwise determined as foreground point.
Then background collection is updated：If current pixel is judged as background dot, it is to belong to change to continue to detect the background dot
Change still changes slow point soon, if belonging to background changes slower point, replaces its background with the pixel value and concentrates what is randomly selected
Some element；If belonging to background changes faster point, with the pixel value replace its background concentrate m randomly selected it is first
Element.
Each picture frame is carried out pixel-by-pixel using the background modeling of the invention shown in Fig. 1 and foreground detection method flow
Judge, finally give the bianry image of prospect and background separation.
Above content is to combine specific preferred embodiment further description made for the present invention, it is impossible to is assert
The specific implementation of the present invention is confined to these explanations.For general technical staff of the technical field of the invention,
On the premise of not departing from present inventive concept, some simple deduction or replace can also be made, should all be considered as belonging to the present invention's
Protection domain.
Claims (1)
1. a kind of improved background modeling and foreground detection method, it is characterised in that：It the described method comprises the following steps：
Step A1：N two field pictures before selection, calculate the average I of each pixel positionN(x, y), established for each pixel respective
Background collection, the background collection element of each pixel by from preceding N two field pictures with it is random in the pixel value of the pixel same position
Extract and form, each background is concentrated and includes M element；
Step A2：Judge that each pixel of present frame belongs to background dot or foreground point, if the member that current pixel background is concentrated
Element is more than T1 with number of the similarity of the pixel less than a given threshold th1, then the point is judged as background dot, otherwise judges
For foreground point；
Step A3：Background collection is not updated if certain pixel is detected as foreground point, and the position is updated if background dot is detected as
The background collection of pixel, is specifically included：
Step B1：If certain point is judged to background dot, the speed of pixel change is first determined whether, that is, finds out the company before present frame
The position pixel is judged to the pixel of background dot, K in continuous K two field pictures<N, and count these background dots and average IN(x, y) it
Difference, if the number of point of the absolute difference more than given threshold th2 can determine whether that the current background dot belongs to change more than T2
Fast background dot, otherwise current background point, which belongs to, changes slow background dot；
Step B2：If the pixel for being detected as background dot, which belongs to background, changes slower point, its background collection is replaced with the pixel value
In some element for randomly selecting；If belonging to background changes faster point, replace its background with the pixel value and concentrate at random
M element of selection；The value of the m is 3~5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510065105.XA CN104680521B (en) | 2015-02-06 | 2015-02-06 | A kind of improved background modeling and foreground detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510065105.XA CN104680521B (en) | 2015-02-06 | 2015-02-06 | A kind of improved background modeling and foreground detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104680521A CN104680521A (en) | 2015-06-03 |
| CN104680521B true CN104680521B (en) | 2018-04-06 |
Family
ID=53315514
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510065105.XA Expired - Fee Related CN104680521B (en) | 2015-02-06 | 2015-02-06 | A kind of improved background modeling and foreground detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104680521B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP6754610B2 (en) * | 2016-05-18 | 2020-09-16 | 株式会社デンソーアイティーラボラトリ | Arithmetic processing unit, arithmetic processing method, and program |
| CN106157332A (en) * | 2016-07-07 | 2016-11-23 | 合肥工业大学 | A kind of motion inspection optimization method based on ViBe algorithm |
| CN106548488B (en) * | 2016-10-25 | 2019-02-15 | 电子科技大学 | A kind of foreground detection method based on background model and inter-frame difference |
| CN108537799B (en) * | 2018-03-21 | 2021-03-23 | 广西师范大学 | Color sampling method based on pixel type and weight statistics |
| CN111209771A (en) * | 2018-11-21 | 2020-05-29 | 晶睿通讯股份有限公司 | Neural network identification efficiency improving method and relevant identification efficiency improving device thereof |
| CN110864412B (en) * | 2019-08-12 | 2021-02-12 | 珠海格力电器股份有限公司 | Air conditioner control method and system |
| CN112669294B (en) * | 2020-12-30 | 2024-04-02 | 深圳云天励飞技术股份有限公司 | Camera shielding detection method and device, electronic equipment and storage medium |
| CN113850133A (en) * | 2021-08-24 | 2021-12-28 | 中国船舶重工集团公司第七0九研究所 | Ship line-crossing detection method and system for ship lock video monitoring |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101621615A (en) * | 2009-07-24 | 2010-01-06 | 南京邮电大学 | Self-adaptive background modeling and moving target detecting method |
| CN101777186A (en) * | 2010-01-13 | 2010-07-14 | 西安理工大学 | Multimodality automatic updating and replacing background modeling method |
| CN103971386A (en) * | 2014-05-30 | 2014-08-06 | 南京大学 | Method for foreground detection in dynamic background scenario |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2015252B1 (en) * | 2007-07-08 | 2010-02-17 | Université de Liège | Visual background extractor |
-
2015
- 2015-02-06 CN CN201510065105.XA patent/CN104680521B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101621615A (en) * | 2009-07-24 | 2010-01-06 | 南京邮电大学 | Self-adaptive background modeling and moving target detecting method |
| CN101777186A (en) * | 2010-01-13 | 2010-07-14 | 西安理工大学 | Multimodality automatic updating and replacing background modeling method |
| CN103971386A (en) * | 2014-05-30 | 2014-08-06 | 南京大学 | Method for foreground detection in dynamic background scenario |
Non-Patent Citations (1)
| Title |
|---|
| 基于实时视频的运动目标检测算法;邱祯艳;《中国优秀硕士学位论文全文数据库 信息科技辑》;20140215;第14-16、22页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104680521A (en) | 2015-06-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104680521B (en) | A kind of improved background modeling and foreground detection method | |
| CN107085714B (en) | Forest fire detection method based on video | |
| CN103971386B (en) | A kind of foreground detection method under dynamic background scene | |
| CN106548488B (en) | A kind of foreground detection method based on background model and inter-frame difference | |
| CN106780565B (en) | Multi-student sitting-up detection method based on optical flow and k-means clustering | |
| JP2013196684A (en) | Object counting method and object counting device | |
| CN104318266B (en) | A kind of image intelligent analyzes and processes method for early warning | |
| CN101738394A (en) | Method and system for detecting indoor smog | |
| CN102609724B (en) | Method for prompting ambient environment information by using two cameras | |
| CN105243356B (en) | A kind of method and device that establishing pedestrian detection model and pedestrian detection method | |
| WO2015024257A8 (en) | Unstructured road boundary detection | |
| TWI382360B (en) | Object detection method and device threrof | |
| CN103310444A (en) | Method of monitoring pedestrians and counting based on overhead camera | |
| Wang et al. | A new fire detection method based on flame color dispersion and similarity in consecutive frames | |
| CN104835145A (en) | Foreground detection method based on self-adaptive Codebook background model | |
| CN111160203A (en) | Loitering and lingering behavior analysis method based on head and shoulder model and IOU tracking | |
| CN105469054B (en) | The model building method of normal behaviour and the detection method of abnormal behaviour | |
| Vlaminck et al. | Obstacle detection for pedestrians with a visual impairment based on 3D imaging | |
| WO2017193679A1 (en) | Method for automatically detecting whether bicycle has fallen onto the ground | |
| CN109711256A (en) | A kind of low latitude complex background unmanned plane target detection method | |
| CN103646242B (en) | Extended target tracking based on maximum stable extremal region feature | |
| CN105184308A (en) | Remote sensing image building detection and classification method based on global optimization decision | |
| CN108737785B (en) | Indoor automatic detection system that tumbles based on TOF 3D camera | |
| CN107729811B (en) | Night flame detection method based on scene modeling | |
| Huang et al. | A method for fast fall 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180406 Termination date: 20210206 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |