CN103945089A - Dynamic target detection method based on brightness flicker correction and IP camera - Google Patents
Dynamic target detection method based on brightness flicker correction and IP camera Download PDFInfo
- Publication number
- CN103945089A CN103945089A CN201410158366.1A CN201410158366A CN103945089A CN 103945089 A CN103945089 A CN 103945089A CN 201410158366 A CN201410158366 A CN 201410158366A CN 103945089 A CN103945089 A CN 103945089A
- Authority
- CN
- China
- Prior art keywords
- brightness
- brightness flicker
- field picture
- flicker
- dynamic object
- 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
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a dynamic target detection method based on brightness flicker correction and an IP camera. Firstly, the brightness flicker correction technology is utilized for correcting a global or local flicker in a frame image with the brightness flicker, and then a dynamic target is classified out through a background model based on the random theory. The reliable dynamic target detection method is provided under the condition of various and changeful light sources in the indoor environment. According to the scheme, the method can be applicable to intelligent monitor on the occasions of various complex light sources like televisions and electric lamps under the indoor environment.
Description
Technical field
The present invention relates to computer vision and digital video recovery technique field, particularly the detection technique of take dynamic object under the indoor environment that IP camera is carrier based on brightness flicker correction.
Background technology
Along with the development of network technology, intelligent video monitoring is in the future development of the monitoring from simulation monitoring to complete IP network intelligent digital.Intelligent video monitoring is in the situation that not needing human intervention, to utilize computer vision technique and mode identification technology that vision signal is processed, analyzed and understands; and on this basis to image sequence automatic analysis; variation in monitoring scene is positioned, identified and follows the tracks of; on this basis the behavior of target is analyzed and anticipation; can when abnormal conditions occur, give the alarm in time or useful information is provided; effectively assist Security Officer to process crisis, and reduce to greatest extent wrong report and fail to report phenomenon.
In intelligent video monitoring application, the interested information relevant to foreground target (as people, automobile, pet etc.) just conventionally of people, these information comprise mutual etc. between outward appearance, position, the direction of motion, movement velocity and the target of moving target.These information are all hidden in the non-semantic video segment of destructuring, can not directly read.In most of actual scene of most time, do not have target only to have powerful connections simultaneously, even if there is moving target in video, background is also occupied relatively large ratio, that is to say that the most of data in monitor video are not that people are interested, but interested information is mostly hidden in a large amount of non-structured pixel that is difficult to obtain, moving object detection utilizes computer vision processing method that these information are extracted exactly, dynamic object is mapped in image sequence, for follow-up identification and behavioural analysis operation provides data.
Dynamic object detection method can be seen as the classification problem into all pixels in two field picture, according to the pixel value of each pixel and corresponding time-space domain information thereof, judges that this pixel belongs to prospect (moving target) or background.In indoor environment, dynamic object detects and disturbed by various factor, wherein one of the most common interference is exactly the overall situation that the use due to dissimilar lighting apparatus causes or local light sudden change, and this distortion phenomenon is called as brightness flicker in digital video recovery technique field.Due to the existence of this phenomenon, in scene, the pixel value of stationary part in two field picture also undergone mutation thereupon relatively, thereby causes classification error.
Existing common background extractive technique, the patent that for example patent No. is WO2009/007198 (visual background extractor, ViBe) is all suitable for for the scene of various video flowings, different colours space, several scenes content and variation.And adopt first random mechanism and the neighborhood mechanism of transmission selected set up and upgrade background model, thereby effectively improved accuracy, increased anti-noise ability and reduced calculation cost.The narration in the article " ViBe:A universal background subtraction algorithm for video sequences " of delivering for 2011 according to this patent inventor M.Van Droogenbroeck and O.Barnich, due to this background extracting technology only utilize a frame in image sequence just can initialization background model, therefore when processing light sudden change, only need reinitialize the impact that background model just can avoid this variation to upgrade background model.In indoor environment, particularly in home environment, the reason that light sudden change forms, except switch lamp, also comprises that these all can impact the light of monitoring environment as television set switch and broadcasting etc.The indoor environment light that the light source scintillation causing due to the multidate information of televising causes changes frequent, if directly utilize ViBe technology to carry out dynamic object detection, may cause because background model cannot be upgraded the correct verification and measurement ratio reduction of the dynamic object causing continuously.
Summary of the invention
The problem existing for above-mentioned existing dynamic object detection technique, the object of the invention is to propose a kind of dynamic object detection method based on brightness flicker correction and IP camera, with this overcome in indoor environment, exist in a large number due to cannot correctly the classify limitation of prospect background of the overall situation or local luminance flicker in the image sequence that lighting apparatus opens, cuts out or the frequent flicker of other light source causes.
In order to achieve the above object, the present invention adopts following technical scheme:
Dynamic object detection method based on brightness flicker correction and IP camera, first described detection method utilizes brightness flicker correction technique, by the overall situation or the winking correction that exist in the two field picture of brightness flicker, then sort out dynamic object by the background model based on random theory.
In the preferred embodiment of this method, described detection method specifically comprises the steps:
(1) preliminary treatment: the integrality to the two field picture by Internet Transmission judges, the two field picture of information completely is carried out to picture quality lifting;
(2) initialization background model: utilize the first two field picture that collects information completely, the parameter of initialization dynamic object detection model, sets up background model, and adjusts the brightness value of this two field picture;
(3) brightness flicker detects and revises: the brightness value with reference to the first two field picture, utilizes brightness flicker correction technique that the brightness flicker that may occur in image sequence is detected and revised;
(4) background model is upgraded: according to the background model of the revised two field picture of brightness flicker and foundation, the pixel in current frame image is classified, and upgrade background model, obtain foreground mask;
(5) reprocessing: the foreground mask obtaining is carried out to the dilation erosion operation of mathematical morphology, improve the integrality of the dynamic object detecting, and get rid of the point that the non-prospect that may exist is dispersed in the classification error at background place.
Further, while carrying out preliminary treatment in described step (1), utilize SAD technology, relatively in vertical direction, three sizes are 36 * 36 faces the sad value between piece mutually, gets rid of the image of the imperfect or decoding error causing due to Internet Transmission reason.
Further, the pretreatment operation in described step (1) comprises: a) contrast strengthens; B) mathematical morphology strengthens; C) edge strengthens; D) gray scale stretches.
Further, while utilizing brightness flicker correction technique to detect and revise in described step (3), Jun Yikuaiwei unit.
Further, the process of described detection and correction is as follows: according to the linear model of flicker, pass through color space conversion, utilize block-based method for fast searching to carry out estimation of motion vectors to moving object, obtain each piece at the motion vector of vertical and horizontal direction, by the brightness average of two field picture corresponding blocks before and after comparing, detect and whether have brightness flicker, if there is brightness flicker, estimate and take advantage of sex factor and add factor in brightness flicker linear model, and on this basis the overall situation of the existence in image sequence or local luminance flicker are revised.
Further, while upgrading background model in described step (4), the background modeling method of utilization based on random theory, according to the revised two field picture of brightness flicker, the background model of having set up is upgraded, more the new stage is in ball and the threshold value of the pre-defined radius of two-dimentional theorem in Euclid space basis, all pixels in each two field picture are classified, and obtaining foreground mask is dynamic object mask.
The overall situation or local luminance flicker that scheme provided by the invention utilizes the brightness flicker technology correction in digital video recovery technique to cause because light source changes, in conjunction with the background extracting technology based on random theory, thereby reduce misclassification result in the correct verification and measurement ratio that improves dynamic object detection.This scheme can be applicable to exist under indoor environment the intelligent monitoring of Various Complex light source (as TV, electric light etc.) occasion.
The invention has the advantages that: the incomplete two field picture of information that utilizes sad value to cause Internet Transmission reason and encoding and decoding reason is got rid of, and avoids causing due to the two field picture of information dropout the error detection of dynamic object.
Utilize brightness flicker in the characteristic of spatial distribution Smooth, the linear model glimmering by foundation, in conjunction with the estimation of motion vectors that adopts three step search methods, by the equal value difference of brightness between frame corresponding blocks before and after computer, judge whether to exist brightness flicker, to what have that the piece of this phenomenon estimates linear model, take advantage of sex factor and add factor, and revise on this basis scintillation, avoid at dynamic object detection-phase because the existence of brightness flicker causes error detection.
Background model and stochastical sampling mechanism that utilization is set up based on random theory are classified to the pixel in two field picture, because the method randomness is stronger, dynamic object testing result is each time all slightly different, therefore this assorting process is carried out twice respectively, again two times result is carried out to logical “and” operation, improve the correctness of target detection.
Accompanying drawing explanation
Below in conjunction with the drawings and specific embodiments, further illustrate the present invention.
Fig. 1 is the schematic flow sheet of the inventive method;
Fig. 2 (a) is the 3rd effect frame figure that loss of data occurs in the image sequence getting when network condition is poor;
Fig. 2 (b) is the 64th effect frame figure that loss of data occurs in the image sequence getting when network condition is poor;
Fig. 2 (c) is the 107th effect frame figure that loss of data occurs in the image sequence getting when network condition is poor;
Fig. 2 (d) is the 180th effect frame figure that loss of data occurs in the image sequence getting when network condition is poor;
Fig. 3 is the sad value contrast schematic diagram that the 180th two field picture of take in the image sequence that IP Camera obtains when network condition is poor is example;
Fig. 4 (a) is the 1st effect frame figure of the image sequence of switch lamp moment of utilizing that IP Camera obtains in indoor environment;
Fig. 4 (b) is the 11st effect frame figure of the image sequence of switch lamp moment of utilizing that IP Camera obtains in indoor environment;
Fig. 4 (c) is the 22nd effect frame figure of the image sequence of switch lamp moment of utilizing that IP Camera obtains in indoor environment;
Fig. 4 (d) is the 24th effect frame figure of the image sequence of switch lamp moment of utilizing that IP Camera obtains in indoor environment;
Fig. 5 is ROC (Receiver Operating Characteristic) the curve comparison schematic diagram that the brightness average of original image and the revised image of brightness flicker is drawn;
Fig. 6 is the schematic diagram based on random theory background model replacement criteria;
Fig. 7 is the foreground mask schematic diagram that utilizes the background extracting method in the present invention to obtain;
Fig. 8 is the foreground mask schematic diagram that utilizes the background extracting method in the present invention to obtain.
Embodiment
For technological means, creation characteristic that the present invention is realized, reach object and effect is easy to understand, below in conjunction with concrete diagram, further set forth the present invention.
Referring to Fig. 1, it is depicted as the flow chart that the present invention carries out dynamic object detection.As seen from the figure, before and after whole testing process, comprise: 1. picture frame preliminary treatment, 2. initialization background model, 3. brightness flicker detects and revises, 4. background model is upgraded and 5. reprocessings, five steps.
Below in conjunction with 1 pair of the solution of the present invention of accompanying drawing, further launch explanation.
The preliminary treatment of step 1. picture frame;
This step starts from the collection of image sequence, because IP camera gathers corresponding view data, and the image sequence data of collection is passed through to Internet Transmission.Owing to passing through Internet Transmission, two field picture may occur because network reason causes the imperfect of the rear view data of decoding in transmitting procedure.Referring to Fig. 2 (a), (b), (c) and (d) this four width accompanying drawing for there is the two field picture design sketch of the 3rd frame, the 64th frame, the 107th frame and the 180th frame of loss of data in the image sequence getting when network condition is poor.
As can be seen from Figure 2, all prolong vertical direction when the information of two field picture makes a mistake, this phenomenon is mainly because most IP camera has adopted code encoding/decoding mode H.264 in order to improve data compression rate.Based on this, the impact for fear of similar two field picture on dynamic object testing result, adopts SAD technology to detect the two field picture by Internet Transmission, judges whether current two field picture exists the incomplete situation of data.Concrete is, and by three sizes in vertical direction relatively, to be 36 * 36 face the sad value between piece mutually, gets rid of the image of the imperfect or decoding error causing due to Internet Transmission reason.
In the present invention, suppose that current frame image is if Fig. 3 is as shown, in figure, piece A, B and C are the piece that size adjacent in vertical direction is N * N, according to following formula, calculate respectively the sad value between A, B and between B, C:
Wherein, δ
pand δ
qrepresent respectively current two pieces that compare, x, y and m, n represent respectively piece δ
pand δ
qthe coordinate figure in the upper left corner, N represents the size of piece, in the present invention N=36.
After obtaining two sad values, utilize following formula to obtain difference D:
D=|d
B,C-d
A,B|
Wherein, d
b,Cand d
a,Brepresent respectively the sad value between B, C and between A, B.In same vertical direction, statistics D is not less than the number of threshold value θ, if be not less than two field picture columns half, thinks that this two field picture exists loss of data, when background model is upgraded, will not consider this two field picture.
Complete after the judgement of two field picture integrality, need to carry out to being judged as complete picture frame the raising of picture quality, concrete operations mainly comprise that contrast strengthens, mathematical morphology strengthens, edge strengthens and gray scale stretches.
Step 2. is according to the complete two field picture of the first frame information, initialization background model;
In this step, utilize the first two field picture of information completely, the parameter of initialization dynamic object detection model, sets up background model, and adjusts the brightness value of this two field picture.
Initialization procedure be take random theory as foundation, and each pixel in background model is carried out to stochastical sampling 20 times in the first frame.As an example, the formula of the background model in the present invention is:
M(x)={v
1,v
2,K,v
N},N=20
The brightness value of step 3. based on former frame image (brightness value of the first two field picture of take be with reference to), carries out brightness flicker detection and correction to two field picture thereafter;
In this step, utilize the two field picture of information completely subsequently and former frame image to compare to judge whether and have brightness flicker (overall situation or local), if exist, this two field picture carried out to brightness flicker correction, for avoiding the undetected survey of local luminance flicker, the detection of brightness flicker and correction Jun Yikuaiwei unit.
Its concrete scheme is the brightness flicker correction technique utilizing in digital video reparation field, according to the linear model of brightness flicker, by color space conversion piecemeal, recycle block-based method for fast searching moving object is carried out to estimation of motion vectors, obtain each piece at the motion vector of vertical and horizontal direction; Finally, by the brightness average of two field picture corresponding blocks before and after comparing, detect and whether exist brightness flicker (concrete by the brightness average between two field picture corresponding blocks before and after calculating, and compare with corresponding threshold value, judge whether to be greater than this threshold value), if there is brightness flicker, estimate and take advantage of sex factor and add factor in brightness flicker linear model, and on this basis the overall situation of the existence in image sequence or local luminance flicker are revised.
Such scheme when specific implementation, brightness flicker detect with revise based on the linear model of brightness flicker:
Y
t(x,y)=α
t(x,y)I
t(x,y)+β
t(x,y)
Wherein, x, y represent the coordinate figure of pixel, and subscript t represents current frame number, Y
t(x, y) represents to exist the image of brightness flicker, I
t(x, y) represents the revised image of brightness flicker.α
t(x, y) and β
twhat (x, y) represented respectively flicker parameter takes advantage of sex factor and add factor.
Because flicker has the characteristic of space smoothing, can think that flicker parameter is constant in subrange.Two field picture is divided into the non overlapping blocks of M * N, and uses Φ
m,nrepresent the top left corner apex coordinate that m, n are piece, α
m, n, t=α
t(x, y), β
m, n, t=β
t(x, y),
.Wherein, α
m, n, tand β
m, n, tfor constant, according to the linear model of brightness flicker, can obtain:
E[Y
t(x,y)]=E[α
m,n,tI
t(x,y)+β
m,n,t]=α
m,n,tE[I
t(x,y)]+β
m,n,t；
σ
2[Y
t(x,y)]=E{Y
t(x,y)-E[Y
t(x,y)]}
2=α
2 m,n,tσ
2[I
t(x,y)]；
Wherein, E and σ
2represent respectively mathematic expectaion and variance, by above two formula, two flicker parameters in linear model can be expressed as:
β
m,n,t=E[Y
t(x,y)]-α
m,n,tE[I
t(x,y)]；
Obviously in continuous image sequence for the region of motion, the x in t-1 frame, y be replaced to x'=x-V
hand y'=y-V
v, V wherein
hand V
vrepresent respectively horizontal motion vector and vertical motion vector, for region V relatively static between consecutive frame image
h=0, V
v=0, the formula of the parameter of glimmering so is just further converted to:
After flicker parameter is determined, the correction of just can substitution brightness flicker linear model glimmering, formula is as follows:
Because above operation is all block-based, if directly utilize the flicker parameter estimating to carry out correction to image, can cause image to occur blocking effect.For keeping image smoothing, flicker parameter need to be carried out to 3 * 3 mean filters, template is as follows:
For fear of to there not being the correction of also glimmering of the region of brightness flicker, in the present invention, define Δ M=|M
i-M
i-1|, wherein, M represents Φ
m,nthe brightness average in region, subscript i and i-1 represent frame number.If the value of Δ M is greater than predefined threshold value T (T=5.0), think piece region Φ
m,nthere is brightness flicker and carry out brightness flicker correction.
Referring to Fig. 4 (a), (b), (c) and (d), shown in it, be comparatively obvious four frames of brightness Change in Mean in the image sequence of switch lamp moment of utilizing that IP Camera obtains in indoor environment: the design sketch of the 1st frame, the 11st frame, the 22nd frame and the 24th frame.
Referring to Fig. 5, it is depicted as and utilizes ROC curve to contrast the brightness Change in Mean schematic diagram of revising front and back image, in figure, the larger curve of fluctuation is by utilizing the brightness average of the image sequence of brightness flicker to draw, and this image sequence is 50 two field pictures of switch lamp moment of utilizing that IP camera gets under indoor environment; Another comparatively mild curve is that the brightness average of the image sequence by obtaining after brightness flicker correction is drawn.From these two curves, can find out, larger on the impact of two field picture brightness average such as the light source sudden change of switch lamp, the fluctuation of brightness Mean curve is also more violent, through the revised curve of brightness flicker, tends towards stability, and illustrates that the sudden change of brightness average is corrected.
Step 4. background model is upgraded;
It is mainly according to the background model of the revised two field picture of brightness flicker and foundation, the pixel in current frame image to be classified that this step is carried out background model renewal, and upgrades background model, obtains foreground mask.
Concrete scheme is, the background modeling method (dynamic object that thus can adapt to complex background detect) of utilization based on random theory, according to the revised two field picture of brightness flicker, the background model of having set up is upgraded, more the new stage is in ball and the threshold value of the pre-defined radius of two-dimentional theorem in Euclid space basis, all pixels in each two field picture are classified, and obtaining foreground mask is dynamic object mask.
Hence one can see that, and the renewal process in the present invention is the similarity of each pixel and background model in comparison current frame image.
Suppose that current pixel to be sorted is v (x), comparative approach is a two-dimentional theorem in Euclid space (C of definition
1, C
2), and on this space, define one and take the ball that radius is R centered by v (x), according to background model M (x)={ v setting up in advance
1, v
2, K, v
n, N=20, calculates by following formula:
λ=S
R(v(x))∩{v
1,v
2,K?v
N}
If the result obtaining is greater than 2, v (x) is divided into background, otherwise is prospect, and this process as shown in Figure 6.
The background model of obtaining due to the method is based on random theory, and the foreground mask therefore obtaining is each time all slightly different, and in the present invention, the correctness detecting in order to improve dynamic object, utilizes the foreground mask obtaining for twice to obtain logical “and” result.
5. reprocessing;
Finally, it is that dynamic object is processed to the foreground mask obtaining that the present invention utilizes the dilation erosion operation in mathematical morphology, improves the integrality of the dynamic object detecting, and gets rid of the point that the non-prospect that may exist is dispersed in the classification error at background place.
The present invention is solved and be take IP camera as video capture device by above five steps, and the dynamic object under indoor environment detects.This programme in the specific implementation tool has the following advantages:
The incomplete two field picture of information that utilizes sad value to cause Internet Transmission reason and encoding and decoding reason is got rid of, and avoids may, because the two field picture of information dropout is imperfect, causing the problem of the error detection of dynamic object in Internet Transmission;
Utilize brightness flicker in the characteristic of spatial distribution Smooth, the linear model glimmering by foundation, in conjunction with the estimation of motion vectors that adopts three step search methods, by the equal value difference of brightness between frame corresponding blocks before and after computer, judge whether to exist brightness flicker, to what have that the piece of this phenomenon estimates linear model, take advantage of sex factor and add factor, and revise on this basis scintillation, realization in image sequence may due to light source suddenly or the gradually change overall situation that causes or the correction of local luminance flicker, avoid at dynamic object detection-phase because the existence of brightness flicker causes error detection,
Background model and stochastical sampling mechanism that utilization is set up based on random theory are classified to the pixel in two field picture, because the method randomness is stronger, dynamic object testing result is each time all slightly different, therefore this assorting process is carried out twice respectively, again two times result is carried out to logical “and” operation, improve correctness, the reduction false detection rate of target detection.Moreover background model is upgraded based on the random mechanism of selecting, the dynamic object that can adapt to complex background detects.
Participation Fig. 7 and Fig. 8 are respectively the overall situation and local luminance flicker two field picture utilizes background extracting technology to obtain the schematic diagram of dynamic object mask after brightness flicker correction.
Wherein, Fig. 7 is the foreground mask schematic diagram that utilizes the background extracting method in this programme to obtain, the image of its input is the 54th frame that utilizes the image sequence that IP Camera gets under indoor environment, comparing with other two field picture in image sequence, there is obvious overall brightness flicker in this two field picture.
Fig. 8 is the foreground mask schematic diagram that utilizes the background extracting method in this programme to obtain, and the image of its input is the 62nd frame that utilizes the image sequence that IP Camera gets under indoor environment.Comparing with other two field picture in image sequence, there is obvious local luminance flicker in this two field picture.
More than show and described basic principle of the present invention, principal character and advantage of the present invention.The technical staff of the industry should understand; the present invention is not restricted to the described embodiments; that in above-described embodiment and specification, describes just illustrates principle of the present invention; without departing from the spirit and scope of the present invention; the present invention also has various changes and modifications, and these changes and improvements all fall in the claimed scope of the invention.The claimed scope of the present invention is defined by appending claims and equivalent thereof.
Claims (7)
1. the dynamic object detection method based on brightness flicker correction and IP camera, it is characterized in that, first described detection method utilizes brightness flicker correction technique, by the overall situation or the winking correction that exist in the two field picture of brightness flicker, then sort out dynamic object by the background model based on random theory.
2. the dynamic object detection method based on brightness flicker correction and IP camera according to claim 1, is characterized in that, described detection method specifically comprises the steps:
(1) preliminary treatment: the integrality to the two field picture by Internet Transmission judges, the two field picture of information completely is carried out to picture quality lifting;
(2) initialization background model: utilize the first two field picture that collects information completely, the parameter of initialization dynamic object detection model, sets up background model, and adjusts the brightness value of this two field picture;
(3) brightness flicker detects and revises: the brightness value with reference to the first two field picture, utilizes brightness flicker correction technique that the brightness flicker that may occur in image sequence is detected and revised;
(4) background model is upgraded: according to the background model of the revised two field picture of brightness flicker and foundation, the pixel in current frame image is classified, and upgrade background model, obtain foreground mask;
(5) reprocessing: the foreground mask obtaining is carried out to the dilation erosion operation of mathematical morphology, improve the integrality of the dynamic object detecting, and get rid of the point that the non-prospect that may exist is dispersed in the classification error at background place.
3. the dynamic object detection method based on brightness flicker correction and IP camera according to claim 2, it is characterized in that, while carrying out preliminary treatment in described step (1), utilize SAD technology, relatively in vertical direction, three sizes are 36 * 36 faces the sad value between piece mutually, gets rid of the image of the imperfect or decoding error causing due to Internet Transmission reason.
4. according to the dynamic object detection method based on brightness flicker correction and IP camera described in claim 2 or 3, it is characterized in that, the pretreatment operation in described step (1) comprises: a) contrast strengthens; B) mathematical morphology strengthens; C) edge strengthens; D) gray scale stretches.
5. the dynamic object detection method based on brightness flicker correction and IP camera according to claim 2, is characterized in that, while utilizing brightness flicker correction technique to detect and revise in described step (3), and Jun Yikuaiwei unit.
6. the dynamic object detection method based on brightness flicker correction and IP camera according to claim 5, it is characterized in that, the process of described detection and correction is as follows: according to the linear model of flicker, pass through color space conversion, utilize block-based method for fast searching to carry out estimation of motion vectors to moving object, obtain each piece at the motion vector of vertical and horizontal direction, by the brightness average of two field picture corresponding blocks before and after comparing, detect and whether have brightness flicker, if there is brightness flicker, estimate and take advantage of sex factor and add factor in brightness flicker linear model, and on this basis the overall situation of the existence in image sequence or local luminance flicker are revised.
7. the dynamic object detection method based on brightness flicker correction and IP camera according to claim 2, it is characterized in that, while upgrading background model in described step (4), the background modeling method of utilization based on random theory, according to the revised two field picture of brightness flicker, the background model of having set up is upgraded, more the new stage is in ball and the threshold value of the pre-defined radius of two-dimentional theorem in Euclid space basis, all pixels in each two field picture are classified, and obtaining foreground mask is dynamic object mask.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410158366.1A CN103945089A (en) | 2014-04-18 | 2014-04-18 | Dynamic target detection method based on brightness flicker correction and IP camera |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410158366.1A CN103945089A (en) | 2014-04-18 | 2014-04-18 | Dynamic target detection method based on brightness flicker correction and IP camera |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103945089A true CN103945089A (en) | 2014-07-23 |
Family
ID=51192558
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410158366.1A Pending CN103945089A (en) | 2014-04-18 | 2014-04-18 | Dynamic target detection method based on brightness flicker correction and IP camera |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103945089A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105574891A (en) * | 2015-12-11 | 2016-05-11 | 上海兴芯微电子科技有限公司 | Method and system for detecting moving object in image |
| WO2016110223A1 (en) * | 2015-01-08 | 2016-07-14 | Gean Technology Co. Limited | Method for detecting electronic lighting flickering |
| CN107644431A (en) * | 2017-08-16 | 2018-01-30 | 孙战里 | Effective method for concentration of video monitoring in the case of a kind of low-light (level) |
| CN108875759A (en) * | 2017-05-10 | 2018-11-23 | 华为技术有限公司 | A kind of image processing method, device and server |
| CN109857907A (en) * | 2019-02-25 | 2019-06-07 | 百度在线网络技术（北京）有限公司 | Video locating method and device |
| EP3506621A1 (en) * | 2017-12-28 | 2019-07-03 | Canon Kabushiki Kaisha | Image processing apparatus, image processing method, and computer program |
| CN110381276A (en) * | 2019-05-06 | 2019-10-25 | 华为技术有限公司 | A kind of video capture method and electronic equipment |
| CN112528794A (en) * | 2020-12-03 | 2021-03-19 | 北京百度网讯科技有限公司 | Signal lamp fault identification method and device and road side equipment |
| CN113936579A (en) * | 2021-12-17 | 2022-01-14 | 深圳精智达技术股份有限公司 | Method and related device for measuring flicker value of LCD display screen |
| WO2024119511A1 (en) * | 2022-12-09 | 2024-06-13 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Image processing device, method, and program |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102014240A (en) * | 2010-12-01 | 2011-04-13 | 深圳市蓝韵实业有限公司 | Real-time medical video image denoising method |
| CN102289819A (en) * | 2011-07-27 | 2011-12-21 | 中国电子科技集团公司第二十八研究所 | Method for detecting infrared motion target in real time for background adaptive estimation |
| CN103456028A (en) * | 2013-08-30 | 2013-12-18 | 浙江立元通信技术有限公司 | Moving object detection method |
-
2014
- 2014-04-18 CN CN201410158366.1A patent/CN103945089A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102014240A (en) * | 2010-12-01 | 2011-04-13 | 深圳市蓝韵实业有限公司 | Real-time medical video image denoising method |
| CN102289819A (en) * | 2011-07-27 | 2011-12-21 | 中国电子科技集团公司第二十八研究所 | Method for detecting infrared motion target in real time for background adaptive estimation |
| CN103456028A (en) * | 2013-08-30 | 2013-12-18 | 浙江立元通信技术有限公司 | Moving object detection method |
Non-Patent Citations (5)
| Title |
|---|
| 孟介成: "运动目标检测的研究与DSP实现", 《中国优秀硕士学术论文电子期刊网》 * |
| 张然然: "面向影像资料的视频修复技术的研究", 《中国优秀博士学术论文电子期刊网》 * |
| 张磊 等: "基于HSV颜色空间和Vibe算法的运动目标检测", 《计算机工程与应用》 * |
| 谌安军: "图像序列中亮度闪烁去除", 《电子与信息学报》 * |
| 赵培栋: "基于多帧参考的分块视频闪烁消除算法", 《上海大学学报（自然科学版）》 * |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016110223A1 (en) * | 2015-01-08 | 2016-07-14 | Gean Technology Co. Limited | Method for detecting electronic lighting flickering |
| CN105574891A (en) * | 2015-12-11 | 2016-05-11 | 上海兴芯微电子科技有限公司 | Method and system for detecting moving object in image |
| CN105574891B (en) * | 2015-12-11 | 2019-04-09 | 上海兴芯微电子科技有限公司 | The method and system of moving target in detection image |
| CN108875759A (en) * | 2017-05-10 | 2018-11-23 | 华为技术有限公司 | A kind of image processing method, device and server |
| CN108875759B (en) * | 2017-05-10 | 2022-05-24 | 华为技术有限公司 | Image processing method and device and server |
| CN107644431A (en) * | 2017-08-16 | 2018-01-30 | 孙战里 | Effective method for concentration of video monitoring in the case of a kind of low-light (level) |
| EP3506621A1 (en) * | 2017-12-28 | 2019-07-03 | Canon Kabushiki Kaisha | Image processing apparatus, image processing method, and computer program |
| US11144762B2 (en) | 2017-12-28 | 2021-10-12 | Canon Kabushiki Kaisha | Image processing apparatus, image processing method, and medium |
| CN109857907A (en) * | 2019-02-25 | 2019-06-07 | 百度在线网络技术（北京）有限公司 | Video locating method and device |
| CN110381276A (en) * | 2019-05-06 | 2019-10-25 | 华为技术有限公司 | A kind of video capture method and electronic equipment |
| CN112528794A (en) * | 2020-12-03 | 2021-03-19 | 北京百度网讯科技有限公司 | Signal lamp fault identification method and device and road side equipment |
| CN113936579A (en) * | 2021-12-17 | 2022-01-14 | 深圳精智达技术股份有限公司 | Method and related device for measuring flicker value of LCD display screen |
| CN113936579B (en) * | 2021-12-17 | 2022-03-04 | 深圳精智达技术股份有限公司 | Method and related device for measuring flicker value of LCD display screen |
| WO2024119511A1 (en) * | 2022-12-09 | 2024-06-13 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Image processing device, method, and program |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103945089A (en) | Dynamic target detection method based on brightness flicker correction and IP camera | |
| Elhabian et al. | Moving object detection in spatial domain using background removal techniques-state-of-art | |
| US9247155B2 (en) | Method and system for robust scene modelling in an image sequence | |
| CN105096321B (en) | A kind of low complex degree Motion detection method based on image border | |
| KR102153607B1 (en) | Apparatus and method for detecting foreground in image | |
| CN109544592B (en) | Moving object detection algorithm for camera movement | |
| CN106778712B (en) | Multi-target detection and tracking method | |
| US7982774B2 (en) | Image processing apparatus and image processing method | |
| CN106325485B (en) | A kind of gestures detection recognition methods and system | |
| CN110610150B (en) | Tracking method, device, computing equipment and medium of target moving object | |
| CN111368771A (en) | Tunnel fire early warning method and device based on image processing, computer equipment and computer readable storage medium | |
| CN102469302B (en) | Background model learning system for lighting change adaptation utilized for video surveillance | |
| CN105184818A (en) | Video monitoring abnormal behavior detection method and detections system thereof | |
| CN112488071B (en) | Method, device, electronic equipment and storage medium for extracting pedestrian features | |
| CN105574891A (en) | Method and system for detecting moving object in image | |
| CN110705412A (en) | Video target detection method based on motion history image | |
| Ling et al. | A background modeling and foreground segmentation approach based on the feedback of moving objects in traffic surveillance systems | |
| Woo et al. | Environmentally robust motion detection for video surveillance | |
| CN101996307A (en) | Intelligent video human body identification method | |
| Chen et al. | Research on moving object detection based on improved mixture Gaussian model | |
| CN103413149A (en) | Method for detecting and identifying static target in complicated background | |
| Ali et al. | Moving object detection using self adaptive Gaussian Mixture Model for real time applications | |
| Zhu et al. | A hybrid algorithm for automatic segmentation of slowly moving objects | |
| CN106611165B (en) | A kind of automotive window detection method and device based on correlation filtering and color-match | |
| CN107729811B (en) | Night flame detection method based on scene modeling |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20140723 |