CN110348305A - A kind of Extracting of Moving Object based on monitor video - Google Patents
A kind of Extracting of Moving Object based on monitor video Download PDFInfo
- Publication number
- CN110348305A CN110348305A CN201910490852.6A CN201910490852A CN110348305A CN 110348305 A CN110348305 A CN 110348305A CN 201910490852 A CN201910490852 A CN 201910490852A CN 110348305 A CN110348305 A CN 110348305A
- Authority
- CN
- China
- Prior art keywords
- image
- formula
- moving region
- indicate
- region
- 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.)
- Granted
Links
- 230000011218 segmentation Effects 0.000 claims description 11
- 239000011159 matrix material Substances 0.000 claims description 3
- 230000000694 effects Effects 0.000 description 11
- 238000000034 method Methods 0.000 description 10
- 238000000605 extraction Methods 0.000 description 4
- 239000000284 extract Substances 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000000750 progressive effect Effects 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000004575 stone Substances 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
- G06V20/42—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items of sport video content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- General Health & Medical Sciences (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- Human Computer Interaction (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The present invention provides a kind of Extracting of Moving Object based on monitor video, comprising: according to AnMixed Gauss model obtain AnIn preliminary doubtful moving region Arn;Calculate AnIn doubtful moving region Dn;Calculate AnIn doubtful moving region gather { Drn}M;Calculate Arn{ Drn}MIntersection J;IfCalculate AnIn final moving region;IfUpdate kth frame background image;To kth frame image AkForefoot area recalled;The present invention can be in many complex conditions such as target delay for a long time, experiencing small oscillating movements, to moving target recognition and Background Reconstruction.
Description
Technical field
The invention belongs to technical field of video monitoring, and in particular to a kind of moving target recognition side based on monitor video
Method.
Background technique
Video monitoring is a kind of intuitive effective important means for recording, observing scenario, wherein moving target recognition
It is the committed step for analyzing video monitoring, is the emphasis of current research.Its main algorithm has algorithm pixel-based: according to pixel
The regularity of distribution constructs probabilistic model, to estimate moving region.But in modeling, a large amount of pixel samples, cost of labor need to be provided
It is larger;Algorithm based on region: according to the consistency of regional texture feature, illumination is reduced to the shadow of moving target and background
It rings, but bad in face of minor change and the unconspicuous motion target area detection effect of feature;Algorithm based on picture frame: according to
Moving target detects in light changing rule building world model, but light variation is complicated and changeable, it is difficult to construct unified mould
Type.
Summary of the invention
For the deficiencies in the prior art, the object of the present invention is to provide a kind of movements based on monitor video
Target extraction method solves the technical issues of prior art is unable to satisfy precision and efficiency of detecting.
In order to solve the above-mentioned technical problem, the application, which adopts the following technical scheme that, is achieved:
A kind of Extracting of Moving Object based on monitor video, comprising the following steps:
Step 1, input image sequence { A1,...,An,...,AN, AnN-th frame image is indicated, to AnConstruct mixed Gaussian mould
Type, according to AnMixed Gauss model obtain AnIn preliminary doubtful moving region set Arn, Indicate AnIn i-th of preliminary doubtful moving region;
Step 2, if background frame image sequence is { B1,...,Bn,...,BN, BnIndicate n-th frame background image;
Step 3, n=1, B are enabled1=A1;
Step 4, A is obtained by formula (1)nIn doubtful moving region set Dn, Table
Show AnIn j-th of doubtful moving region, M is integer greater than 1;
In formula (1), TnFor segmentation threshold;
Wherein, Agn(x, y) indicates AnGray level image, Bgn(x, y) indicates BnGray level image;
Step 5, D is traversednIn each doubtful moving region carry out mathematical morphology operation, obtain AnIn preliminary motion area
Domain set Drn, Indicate DrnIn j-th of preliminary motion region;
Step 6, it is calculated by formula (2)With set DrnIn each preliminary motion regionIntersection J;
In formula (2),Indicate DrnIn j-th of preliminary motion region,Indicate AnIn i-th of preliminary doubtful fortune
Dynamic region;
Step 7, A is obtained by formula (3)nIn moving region;
In formula (3),Indicate AnIn i-th of moving region,Indicate AnIn i-th of preliminary doubtful motor area
Domain;
Step 8, if as n=k, kth frame background image is updated by formula (4), obtains updated kth frame background image
Bk={ Bk,1,Bk,2...Bk,h};
In formula (4),Indicate empty set, H be image channel number, color image H={ 1,2,3 }, gray level image H={ 1 },
Ak,h(x, y) is kth frame image AkIn pixel value of the channel h in coordinate (x, y), Bk,h(x, y) is kth frame background image BkIt is logical in h
Pixel value of the road in coordinate (x, y);Tk,hTo update segmentation threshold;
Wherein,ForThe minimum square at place
Shape frame region, H are image channel number, color image H={ 1,2,3 }, gray level image H={ 1 };QkFor AkMiddle satisfactionPixel sum;
Step 9, according to updated kth frame background image Bk, obtain kth frame image AkAll frame images before
{A1,...,Al,...,Ak-1Moving region sequence { Cr before update1′,...,Crl′,...,Crk′-1, it willDate back
K frame image AkAll frame image { A before1,...,Al,...,Ak-1Moving region moving region sequence { Cr1,...,
Crl,...,Crk, wherein
Step 10, n=n+1, Bn=Bn-1, step 4 is repeated to step 9, until n=N is to get arriving image sequence { A1,...,
An,...,ANIn moving region sequence { Cr1,...,Crn,...,CrN, wherein
Further, to A described in step 1nMixed Gauss model is constructed, according to AnMixed Gauss model obtain AnIn
Preliminary doubtful moving region Arn, comprising:
Step 11, mixed Gauss model shown in formula (1) is constructed;
In formula (1), g indicates the number of Gauss model, and η indicates Gaussian probability-density function, ωi,nIndicate n-th frame image pair
I-th of Gauss model weight, ui,nIndicate n-th frame image to i-th of Gauss model mean value, ∑i,nIndicate n-th frame image to i-th
A Gauss model covariance matrix;
Step 12, according to mixed Gauss model, n-th frame image is updated to i-th of Gauss model weight ωi,n;
ωi,n=(1- α) ωi,n-1+α(Mi,n) (2)
In formula (2), α is Studying factors;Mi,nIt is n-th frame image pixel and i-th of Gauss model matching degree,
Step 13, b Gauss model is chosen according to formula (3) construct background model Bb;
In formula (3), T is preset threshold;
Step 14, according to background model Bb, each pixel A in preliminary motion region is obtained by formula (4)nThe pixel of (x, y)
Value Arn(x,y);
By AnMiddle pixel value ArnRegion composed by the pixel that (x, y) is 1 is as preliminary doubtful moving region set
Arn。
Compared with prior art, the present invention beneficial has the technical effect that
1. the present invention blends GMM and frame difference method, it can be achieved that under the conditions of the uncertain factors such as the static, movement of target, transport
Moving-target extracts.
2. the present invention establishes backtracking system again and rebuilds pure background, it can be achieved that under complex background.
3. the present invention establishes local updating system, average treatment speed is fast, reaches 0.35 frame/s.
Detailed description of the invention
Fig. 1 is the flow chart of the method for the present invention;
Fig. 2 is the treatment effect figure of Class1;
Fig. 3 is the treatment effect figure of type 2;
Fig. 4 is the treatment effect figure of type 3;
Fig. 5 is ROI analysis chart;
Explanation is further explained in detail to particular content of the invention below in conjunction with drawings and examples.
Specific embodiment
Specific embodiments of the present invention are given below, it should be noted that the invention is not limited to implement in detail below
Example, all equivalent transformations made on the basis of the technical solutions of the present application each fall within protection scope of the present invention.
Embodiment:
The present embodiment provides a kind of Extracting of Moving Object based on monitor video, comprising the following steps:
Step 1, input image sequence { A1,...,An,...,AN, AnN-th frame image is indicated, to AnConstruct mixed Gaussian mould
Type, according to AnMixed Gauss model obtain AnIn preliminary doubtful moving region set Arn, by ArnIn a connected region
As a preliminary motion region Indicate AnIn i-th of preliminary doubtful motor area
Domain;
Include:
Step 11, mixed Gauss model shown in formula (1) is constructed;
In formula (1), k indicates the number of Gauss model, and η indicates Gaussian probability-density function, ωi,nIndicate n-th frame image pair
I-th of Gauss model weight, ui,nIndicate n-th frame image to i-th of Gauss model mean value, ∑i,nIndicate n-th frame image to i-th
A Gauss model covariance matrix;
Step 12, according to mixed Gauss model, n-th frame image is updated to i-th of Gauss model weight ωi,n;
ωi,n=(1- α) ωi,n-1+α(Mi,n) (2)
In formula (2), α is Studying factors;
Mi,nIt is n-th frame image pixel and i-th of Gauss model matching degree,Wherein, Mi,nMeter
It calculates with reference to " Jin Guangzhi, stone forest lock, Bai Xiangfeng wait to answer based on novel object detection system [J] computer of mixed Gauss model
With 2011,31 (12): the method in 3360-3362. ".
Step 13, b Gauss model is chosen according to formula (3) construct background model Bb;
In formula (3), T is preset threshold;
Step 14, according to background model Bb, each pixel A in preliminary motion region is obtained by formula (4)nThe pixel of (x, y)
Value Arn(x,y);
By AnMiddle pixel value ArnRegion composed by the pixel that (x, y) is 1 is as preliminary doubtful moving region Arn。
It can be obtained by formula (4), the region that pixel value is 1 is preliminary motion regional ensemble, the company in region that pixel value is 1
Logical region is a preliminary motion region.
Step 2, if background frame image sequence is { B1,...,Bn,...,BN, BnIndicate n-th frame background image;
Step 3, n=1, B are enabled1=A1;
Step 4, A is obtained by formula (1)nIn doubtful moving region set Dn,By Dn
In a connected region as a doubtful moving region,Indicate AnIn j-th of doubtful moving region, M be greater than 1
Integer;
In formula (5), TnFor segmentation threshold;
Wherein, Agn(x, y) indicates AnGray level image, Bgn(x, y) indicates BnGray level image;
It is doubtful moving region set, the region for being 1 by pixel value by the way that the region that pixel value is 1 can be obtained in formula (1)
In connected region as doubtful moving region.
Step 5, D is traversednIn each doubtful moving region carry out mathematical morphology operation, obtain AnIn preliminary motion area
Domain set Drn, Indicate DrnIn j-th of preliminary motion region;
Step 6, it is calculated by formula (2)With set DrnIn each preliminary motion regionIntersection J;
In formula (2),Indicate DrnIn j-th of preliminary motion region,Indicate AnIn i-th of preliminary doubtful fortune
Dynamic region;
Step 7, A is obtained by formula (3)nIn moving region;
In formula (3),Indicate AnIn i-th of moving region,Indicate AnIn i-th of preliminary doubtful motor area
Domain;
Step 8, if as n=k, kth frame background image is updated by formula (4), obtains updated kth frame background image
Bk={ Bk,1,Bk,2...Bk,h};
In formula (4),Indicate empty set, H be image channel number, color image H={ 1,2,3 }, gray level image H={ 1 },
Ak,h(x, y) is kth frame image AkIn pixel value of the channel h in coordinate (x, y), Bk,h(x, y) is kth frame background image BkIt is logical in h
Pixel value of the road in coordinate (x, y);Tk,hTo update segmentation threshold;
Wherein,ForThe minimum square at place
Shape frame region, H are image channel number, color image H={ 1,2,3 }, gray level image H={ 1 };QkFor AkMiddle satisfactionPixel sum;
Step 9, according to updated kth frame background image Bk, obtain kth frame image AkAll frame images before
{A1,...,Al,...,Ak-1Moving region sequence { Cr before update1′,...,Crl′,...,Cr′k-1}.It willDate back
K frame image AkAll frame image { A before1,...,Al,...,Ak-1Moving region, moving region sequence { Cr1,...,
Crl,...,Crk}.Wherein
Step 10, n=n+1, Bn=Bn-1, step 4 is repeated to step 9, until n=N is to get arriving image sequence { A1,...,
An,...,ANIn moving region sequence { Cr1,...,Crn,...,CrN, wherein
Experimental verification:
Class1: start no moving target, then target moves always.
Type 2: start target movement, then the target residence some time.
Type 3: beginning target is static, and then target moves.Since gauss hybrid models are progressive learning process, so real
10 frame images are learnt before testing selecting video.
The present invention as shown in Fig. 2, focus to moving region by gauss hybrid models, obtains the treatment effect of Class1
Moving target approximate region, but the little point of localized variation cannot be effectively extracted, it needs to repair, then frame difference method is extracted and is tied
Fruit and gauss hybrid models extraction result blend and accurately extract moving target { Cr50}1, rebuild background B50。
The present invention to the treatment effect of type 2 as shown in figure 3, focus to moving region by gauss hybrid models, with
The passage of time, foreground pixel point can gradually be converted into background pixel point.Moving target is accurately extracted by the algorithm of this programme
{Cr100}1, rebuild background B100.As the time further elapses, final foreground moving region is all fallen into oblivion in the background, but is counted
Calculation machine can not determine that the region is caused by target stops for a long time, still to rest in background image, need context update.Pass through
The algorithm of this programme accurately extracts moving target { Cr1000}1, rebuild background B1000。
The present invention to the treatment effect of type 3 as shown in figure 4, due to gauss hybrid models algorithm, prospect background conversion
It is a progressive learning process, it may appear that trailing phenomenon.Computer can not determine that the region is caused by target stops for a long time, also
It is to rest in background image, needs context update.Moving target { Cr is accurately extracted by the algorithm of this programme1000}1, rebuild
Background B1000.On this basis, removal trailing phenomenon is recalled by background, obtains moving target { Cr100}1, rebuild background B100。
It is as shown in Figure 5 to count ROI distribution situation.Know A816There are a large amount of ROI, compares A815With A816, it is known that A816Occur
Virtualization, proves that the algorithm can effectively detect subtle variation from side.
The present invention can be in many complex conditions such as target delay for a long time, experiencing small oscillating movements, to movement in summary
Objective extraction and Background Reconstruction.
It is proposed method of the present invention and mainstream algorithm are compared according to segmentation precision and processing time.Using area degree of being folded
The evaluation index of (area overlap measure, AOM) as segmentation effect.Its is defined as:
Wherein, AOM is area degree of being folded, and υ is the moving region of handmarking, and ν is the moving region that algorithm is partitioned into, S
() indicates the pixel number of corresponding region, and AOM value shows that more greatly segmentation effect is better.Testing result is as shown in table 1.
1 AOM of table and average time statistics
According to upper table it is found that the case where moving always for target, the algorithm of low-rank pixel-based, judge background and fortune
Moving-target, segmentation precision is high, but the algorithm need to calculate all pixels point, and average handling time is long.And the calculation based on frame
Method only considers Global Information, and the processing time is fast, but precision is not high.And the moving target proposed in this paper based on Gestalt principle
Extraction and Background Reconstruction algorithm, although the slightly inferior algorithm with frame on average handling time, in the case where target moves always
Segmentation precision is slightly inferior to the algorithm of pixel, but from general effect and after target movement, stops for a long time and target starting stops,
After the case where moving, accuracy comparison mainstream algorithm is improved largely.
Claims (2)
1. a kind of Extracting of Moving Object based on monitor video, which comprises the following steps:
Step 1, input image sequence { A1,...,An,...,AN, AnN-th frame image is indicated, to AnMixed Gauss model is constructed,
According to AnMixed Gauss model obtain AnIn preliminary doubtful moving region set Arn, Table
Show AnIn i-th of preliminary doubtful moving region;
Step 2, if background frame image sequence is { B1,...,Bn,...,BN, BnIndicate n-th frame background image;
Step 3, n=1, B are enabled1=A1;
Step 4, A is obtained by formula (1)nIn doubtful moving region set Dn, Indicate An
In j-th of doubtful moving region, M is integer greater than 1;
In formula (1), TnFor segmentation threshold;
Wherein, Agn(x, y) indicates AnGray level image, Bgn(x, y) indicates BnGray level image;
Step 5, D is traversednIn each doubtful moving region carry out mathematical morphology operation, obtain AnIn preliminary motion region collection
Close Drn, Indicate DrnIn j-th of preliminary motion region;
Step 6, it is calculated by formula (2)With set DrnIn each preliminary motion regionIntersection J;
In formula (2),Indicate DrnIn j-th of preliminary motion region,Indicate AnIn i-th of preliminary doubtful motor area
Domain;
Step 7, A is obtained by formula (3)nIn moving region;
In formula (3),Indicate AnIn i-th of moving region,Indicate AnIn i-th of preliminary doubtful moving region;
Step 8, if as n=k, kth frame background image is updated by formula (4), obtains updated kth frame background image Bk=
{Bk,1,Bk,2...Bk,h};
In formula (4),Indicate empty set, H is image channel number, color image H={ 1,2,3 }, gray level image H={ 1 }, Ak,h(x,
It y) is kth frame image AkIn pixel value of the channel h in coordinate (x, y), Bk,h(x, y) is kth frame background image BkIt is being sat in the channel h
Mark the pixel value of (x, y);Tk,hTo update segmentation threshold;
H ∈ H, whereinForThe minimum rectangle frame area at place
Domain, H are image channel number, color image H={ 1,2,3 }, gray level image H={ 1 };QkFor AkMiddle satisfaction
Pixel sum;
Step 9, according to updated kth frame background image Bk, obtain kth frame image AkAll frame image { A before1,...,
Al,...,Ak-1Moving region sequence { Cr before update1′,...,Crl′,...,Crk′-1, it willDate back kth frame image
AkAll frame image { A before1,...,Al,...,Ak-1Moving region moving region sequence { Cr1,...,Crl,...,
Crk, wherein
Step 10, n=n+1, Bn=Bn-1, step 4 is repeated to step 9, until n=N is to get arriving image sequence { A1,...,
An,...,ANIn moving region sequence { Cr1,...,Crn,...,CrN, wherein
2. Extracting of Moving Object as described in claim 1, which is characterized in that A described in step 1nConstruct mixed Gaussian
Model, according to AnMixed Gauss model obtain AnIn preliminary doubtful moving region Arn, comprising:
Step 11, mixed Gauss model shown in formula (1) is constructed;
In formula (1), g indicates the number of Gauss model, and η indicates Gaussian probability-density function, ωi,nIndicate n-th frame image to i-th
A Gauss model weight, ui,nIndicate n-th frame image to i-th of Gauss model mean value, ∑i,nIndicate that n-th frame image is high to i-th
This model covariance matrix;
Step 12, according to mixed Gauss model, n-th frame image is updated to i-th of Gauss model weight ωi,n;
ωi,n=(1- α) ωi,n-1+α(Mi,n) (2)
In formula (2), α is Studying factors;
Mi,nIt is n-th frame image pixel and i-th of Gauss model matching degree,
Step 13, b Gauss model is chosen according to formula (3) construct background model Bb;
In formula (3), T is preset threshold;
Step 14, according to background model Bb, each pixel A in preliminary motion region is obtained by formula (4)nThe pixel value of (x, y)
Arn(x,y);
By AnMiddle pixel value ArnRegion composed by the pixel that (x, y) is 1 is as preliminary doubtful moving region set Arn。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910490852.6A CN110348305B (en) | 2019-06-06 | 2019-06-06 | Moving object extraction method based on monitoring video |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910490852.6A CN110348305B (en) | 2019-06-06 | 2019-06-06 | Moving object extraction method based on monitoring video |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110348305A true CN110348305A (en) | 2019-10-18 |
CN110348305B CN110348305B (en) | 2021-06-25 |
Family
ID=68181576
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910490852.6A Expired - Fee Related CN110348305B (en) | 2019-06-06 | 2019-06-06 | Moving object extraction method based on monitoring video |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110348305B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111259718A (en) * | 2019-10-24 | 2020-06-09 | 杭州安脉盛智能技术有限公司 | Escalator retention detection method and system based on Gaussian mixture model |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103077530A (en) * | 2012-09-27 | 2013-05-01 | 北京工业大学 | Moving object detection method based on improved mixing gauss and image cutting |
CN105261037A (en) * | 2015-10-08 | 2016-01-20 | 重庆理工大学 | Moving object detection method capable of automatically adapting to complex scenes |
CN106780548A (en) * | 2016-11-16 | 2017-05-31 | 南宁市浩发科技有限公司 | moving vehicle detection method based on traffic video |
CN106772304A (en) * | 2016-12-23 | 2017-05-31 | 西北大学 | Doppler's adaptive processing method after airborne MIMO radar based on spatial domain multi-level decomposition |
CN107895379A (en) * | 2017-10-24 | 2018-04-10 | 天津大学 | The innovatory algorithm of foreground extraction in a kind of video monitoring |
CN108537821A (en) * | 2018-04-18 | 2018-09-14 | 电子科技大学 | A kind of moving target detecting method based on video |
-
2019
- 2019-06-06 CN CN201910490852.6A patent/CN110348305B/en not_active Expired - Fee Related
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103077530A (en) * | 2012-09-27 | 2013-05-01 | 北京工业大学 | Moving object detection method based on improved mixing gauss and image cutting |
CN105261037A (en) * | 2015-10-08 | 2016-01-20 | 重庆理工大学 | Moving object detection method capable of automatically adapting to complex scenes |
CN106780548A (en) * | 2016-11-16 | 2017-05-31 | 南宁市浩发科技有限公司 | moving vehicle detection method based on traffic video |
CN106772304A (en) * | 2016-12-23 | 2017-05-31 | 西北大学 | Doppler's adaptive processing method after airborne MIMO radar based on spatial domain multi-level decomposition |
CN107895379A (en) * | 2017-10-24 | 2018-04-10 | 天津大学 | The innovatory algorithm of foreground extraction in a kind of video monitoring |
CN108537821A (en) * | 2018-04-18 | 2018-09-14 | 电子科技大学 | A kind of moving target detecting method based on video |
Non-Patent Citations (1)
Title |
---|
王春兰: "智能视频监控***中运动目标检测方法综述", 《自动化与仪器仪表》 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111259718A (en) * | 2019-10-24 | 2020-06-09 | 杭州安脉盛智能技术有限公司 | Escalator retention detection method and system based on Gaussian mixture model |
Also Published As
Publication number | Publication date |
---|---|
CN110348305B (en) | 2021-06-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2021208275A1 (en) | Traffic video background modelling method and system | |
CN107610087B (en) | Tongue coating automatic segmentation method based on deep learning | |
CN111582294B (en) | Method for constructing convolutional neural network model for surface defect detection and application thereof | |
US20210118144A1 (en) | Image processing method, electronic device, and storage medium | |
CN109145872B (en) | CFAR and Fast-RCNN fusion-based SAR image ship target detection method | |
CN110070091B (en) | Semantic segmentation method and system based on dynamic interpolation reconstruction and used for street view understanding | |
CN103870834B (en) | Method for searching for sliding window based on layered segmentation | |
CN107016664B (en) | A kind of bad needle flaw detection method of large circle machine | |
CN104077577A (en) | Trademark detection method based on convolutional neural network | |
CN109712127B (en) | Power transmission line fault detection method for machine inspection video stream | |
CN104899877A (en) | Image foreground extraction method based on super-pixels and fast three-division graph | |
Li et al. | Application of multi-scale feature fusion and deep learning in detection of steel strip surface defect | |
CN109685045A (en) | A kind of Moving Targets Based on Video Streams tracking and system | |
CN110399840A (en) | A kind of quick lawn semantic segmentation and boundary detection method | |
CN109961070A (en) | The method of mist body concentration is distinguished in a kind of power transmission line intelligent image monitoring | |
CN109191430A (en) | A kind of plain color cloth defect inspection method based on Laws texture in conjunction with single classification SVM | |
CN104063713A (en) | Semi-autonomous on-line studying method based on random fern classifier | |
CN108509950A (en) | Railway contact line pillar number plate based on probability characteristics Weighted Fusion detects method of identification | |
Liang et al. | Automatic defect detection of texture surface with an efficient texture removal network | |
CN116721414A (en) | Medical image cell segmentation and tracking method | |
CN115082776A (en) | Electric energy meter automatic detection system and method based on image recognition | |
CN113177927A (en) | Bone marrow cell classification and identification method and system based on multiple features and multiple classifiers | |
CN117495735A (en) | Automatic building elevation texture repairing method and system based on structure guidance | |
CN115266732A (en) | Carbon fiber tow defect detection method based on machine vision | |
Ding et al. | Classification of chromosome karyotype based on faster-rcnn with the segmatation and enhancement preprocessing model |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for 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 | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20210625 |