CN1770204A - Method for extracting barycenter trajectory of motive object from motive video with static background - Google Patents
Method for extracting barycenter trajectory of motive object from motive video with static background Download PDFInfo
- Publication number
- CN1770204A CN1770204A CN 200410086740 CN200410086740A CN1770204A CN 1770204 A CN1770204 A CN 1770204A CN 200410086740 CN200410086740 CN 200410086740 CN 200410086740 A CN200410086740 A CN 200410086740A CN 1770204 A CN1770204 A CN 1770204A
- Authority
- CN
- China
- Prior art keywords
- video
- frame
- motion object
- sport
- static background
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Abstract
The invention discloses a method for getting motion object gravity trace from motion video with static background which comprises steps of: getting a motion video consisted with a plurality of video frames and storing into computer; extracting the static background from motion video; for each video frame, subtracting the video frame from the static background, getting motion foreground of the video frame; extracting two-dimensional figure of motion object from the foreground area of the video frame; calculating the gravity point of the two-dimensional figure of motion object; getting motion objects gravities trace by converting all the gravity points of all video frames to the same image. The invention has the advantages of having good effects.
Description
Technical field
The present invention relates to a kind of motion object center of gravity track extraction method of sport video, particularly a kind of method of from sport video, extracting motion object barycenter trajectory with static background.
Background technology
Utilizing video that the motion motion of objects is analyzed is pattern-recognition, and the hot issue in virtual reality, Intelligent Human-Machine Interface field has great using value.For example, in the sports video athletic motion is analyzed, for instructing trainer and athletic training, improving the level of training and competitiveness has important directive significance.
Wherein, the 2D barycenter trajectory of motion object when carrying out specific action is an extremely important parameter in the motion analysis, if can from video, obtain the 2D barycenter trajectory of motion object in the motion process, just can provide strong foundation for the quality of judging the performed action of motion object.For example,, just can provide effective means,, improve the level of training and competitiveness and have important value improving athletic moving-mass for instructing athletic training if be applied in the sports video.
So-called static background sport video is meant in sequence of video images the video that scene background is constant.For example, in the trampolining video, video camera is actionless, and scene background also is static, so the trampolining video just belongs to the static background sport video.If a kind of method can be arranged, can be under the static background condition, from sport video, find the barycenter trajectory of motion object, as: the centre-of-gravity path that finds the trampolinist, just can study motion motion of objects rule better, can do further research to the motion motion of objects according to motion motion of objects rule.
But in the existing Video processing software at home and abroad, all do not provide the function of motion object barycenter trajectory in this automatic extraction static background video.So far, do not find the patent that above-mentioned functions can be provided yet.
Summary of the invention
The objective of the invention is under the static background condition, from sport video, find the barycenter trajectory of motion object, thereby study motion motion of objects rule better, and a kind of method of extracting motion object barycenter trajectory from the sport video with static background is provided.
For achieving the above object, the present invention proposes a kind ofly to extract the method for motion object barycenter trajectory from the sport video with static background, and this method comprises the steps:
Obtain a sport video of forming by a plurality of frame of video, and deposit computing machine in; Described sport video has static background and the sport foreground on static background, and described sport foreground includes the motion object;
Extract the static background of described sport video;
For each frame of video, it is poor that this frame of video and described static background are made frame, obtains the sport foreground zone of this frame of video;
For each frame of video, go out the two-dimensional silhouette of motion object from the described sport foreground extracted region of frame of video;
For each frame of video, calculate the focus point of described motion object two-dimensional silhouette;
The focus point of the motion object of all frame of video is transformed in the same image, obtains the barycenter trajectory of motion object.
In the technique scheme, the described static background that extracts sport video comprises:
Ask the frame of video difference figure of adjacent video interframe;
Ask the stationary part among the frame of video difference figure, described stationary part comprises at least one segmentation;
Fill relevant position in the background with the pixel of the corresponding frame number of the longest segmentation mid point in the stationary part of frame of video difference figure, obtain static background.
Describedly ask frame of video difference figure between consecutive frame to comprise the luminance component between consecutive frame is subtracted each other.
Describedly ask frame of video difference figure between consecutive frame also to comprise to set a threshold values to remove noise.
In the technique scheme, the two-dimensional silhouette that described sport foreground extracted region from frame of video goes out the motion object comprises denoising and/or removes non-motion object.
Described denoising comprise remove area in the described sport foreground zone less than a threshold values the hole.
The non-motion object of described removal comprises removes the foreground area that is in non-middle section in the frame of video.
In the technique scheme, the focus point of the described motion object of described calculating two-dimensional silhouette is a geometric center of asking motion object two-dimensional silhouette.
The advantage of method of the present invention is:
1, the inventive method has good versatility, all is applicable to the inventive method as long as background is static sport video.
2, the inventive method is in extracting the motion object outline process, adopted the method for progressively removing noise by different level based on the characteristics of motion, even there being big noise, cause under the not satisfactory situation of prospect profile extraction effect, still can access comparatively desirable effect.
Description of drawings
Fig. 1 is that the motion object barycenter trajectory of static background sport video extracts process flow diagram.
Embodiment
Below in conjunction with the drawings and specific embodiments the present invention is described in further detail.
With reference to Fig. 1, be the process flow diagram of present embodiment.Expression operation in the frame of broken lines among Fig. 1, the result that the expression associative operation obtains in the solid box.As shown in Figure 1, present embodiment comprises the following steps:
Step 10: utilize picture pick-up device to obtain original sport video, and deposit computing machine in.This sport video has static background and the sport foreground on static background.This sport foreground includes the motion object, the motion object here typically refer to this sport video the main moving target of taking, the sportsman in the sports video for example.Also may comprise some other non-motion object in this sport foreground, the non-motion object here typically refers to other moving object except that the motion object of sport video, and for example the sportsman is employed such as synkinesia instruments such as ball or rods.
Step 20: pre-service.Promptly adopt conventional smoothing technique that the sport video that is obtained is carried out smoothing processing,, then sport video is cut apart each frame of video of generation with preliminary removal The noise.When sport video is carried out smoothing processing, preferably adopt the local average smoothing technique of medium filtering, the inhibition of interference of its paired pulses and refined salt is effective, can reduce the fog-level at edge when suppressing random noise.Medium filtering is a kind of nonlinear image smoothing method, it carries out gray scale ordering to all pixels in the moving window and forms the sequence of a gray-scale value, with the Mesophyticum in the sequence for the original gray scale of window center pixel (, then getting the average of two intermediate values in the gray value sequence) if in the window even number pixel is arranged.
Step 30: the static background that extracts sport video.Specifically may further comprise the steps:
A1) ask frame of video poor: frame of video difference CDM (Change Detection Mask) is used for reflecting the grey scale change between the consecutive frame, by the situation of change that can know easily between frame and the frame of asking for of gray scale frame difference.Specifically finding the solution of frame of video difference is as follows: (i) expression sport video i frame is in that (wherein i represents frame number (i=1...N) for x, the pixel value of y) locating, and N is the sequence totalframes, and (x y) represents volume coordinate for x, y with image sequence I.(x, y, luminance component i) are I to image sequence I
L(i), then the solution formula of frame of video difference CDM is as follows for x, y:
Wherein, threshold value T is used to the second removal noise, and threshold values T is a known known value, need not the user and sets separately.(x, y), (x, y i) can be expressed as the function of frame number i to CDM, and it has write down, and (x y) locates the change curve of the frame of video difference of pixel along time shaft in the position to fixing coordinate position.
B1) ask stationary part in the frame of video difference: for fixing coordinate position (x, y), according to CDM (x, y, i) whether greater than zero with CDM (x, y, i) curve segmentation, wherein CDM (x, y, value i) is that 0 part is called stationary part, this stationary part is made up of one or more segmentation.
C1) background extracting: for fixing coordinate position (x, y), its CDM (x, y i) pick out the longest segmentation in the stationary part of curve, and the corresponding frame number that writes down this segmentation mid point be M (x, y); With M (x, y) the frame place (x, pixel value y) fill (x, y) position in the video background.Like this, when spread all over coordinate positions all on the frame of video (x, y) after, can form a complete background.This step can be described with following formula:
M(x,y)=(ST(x,y)+EN(x,y))/2
B(x,y)=I(x,y,M(x,y))
Wherein, ST (x, y) and EN (x, y) the denotation coordination position (x, frame of video difference CDM y) (x, y, the starting point frame number and the terminal point frame number of long segmentation in stationary part i), (x y) is (x, the pixel value of y) locating in the static background to B.
Step 40: the foreground area profile that extracts sport video.Concrete operations are to utilize the static background that has obtained, and obtain the background frames difference figure ID of luminance component between each frame in video and the static background with subtraction
L, background frames difference figure ID
LIn non-zero region reflected the foreground area of each frame of video.Background frames difference figure ID
LComputing formula as follows:
Wherein, I
LBe the luminance component of frame of video, B
LIt is the luminance component of background.Finding the solution background frames difference figure ID
LProcess in, utilize threshold values T also can remove partial noise.In actual conditions, background is not fully static, owing to some local noises have been introduced in illumination or other interference, these interference bring very big difficulty to the correct detection and the location of moving target, therefore, can utilize other information such as colourity or morphology methods to eliminate background frames difference figure ID
LIn these noises.
Step 50: to each background frames difference figure ID
LForeground area carry out the layering denoising, obtain two dimension (2D) profile of accurate movement object.Background frames difference figure ID in the step 40
LIn non-zero region reflected the profile of the foreground area of frame of video, but wherein also include noise or non-motion object usually.Noise wherein is usually expressed as the hole, at background frames difference figure ID
LIn, the hole is meant a black region that is surrounded by the white background zone.The processing in hole can be carried out as follows: at first search out background frames difference figure ID
LAll holes that comprised among the figure; Set a threshold values and calculate the area in each hole, if the hole area less than threshold value, just that current hole is all pixel is composed with background value, carries out the hole cancellation; If greater than threshold value, then keep original hole.Because the motion object in the foreground area generally is positioned at the central part of frame of video, the foreground area that therefore is arranged in the non-middle section of frame of video can be thought the non-motion object of sport foreground, the pixel of corresponding region can be composed with background value, carries out cancellation.
To background frames difference figure ID
LIn the hole and after non-motion object handles, can obtain the two-dimensional silhouette of accurate movement object.
Step 60: the calculating of motion object center of gravity.Obtain after the accurate two-dimensional silhouette of motion object,, just can access the two-dimentional center of gravity of motion object by the simple two-dimensional computing.The two-dimentional computing here just is meant the geometric center of calculating the two dimensional motion object outline:
For example: establish G
LBe the motion object outline zone of L two field picture in the video sequence, and (X
1, Y
1), (X
2, Y
2) ..., (X
n, Y
n) be component movement object outline zone G
LAll pixels, then (X is the 2D barycenter trajectory point coordinate of this two field picture just Y), wherein:
Step 70: will from each frame of video obtains same frame of video that motion object focus point is transformed into appointment, each focus point be connected, just obtain a continuous motion object barycenter trajectory.
Claims (8)
1, a kind ofly extract the method for motion object barycenter trajectory from the sport video with static background, this method comprises the steps:
Obtain a sport video of forming by a plurality of frame of video, and deposit computing machine in; Described sport video has static background and the sport foreground on static background, and described sport foreground includes the motion object;
Extract the static background of described sport video;
For each frame of video, it is poor that this frame of video and described static background are made frame, obtains the sport foreground zone of this frame of video;
For each frame of video, go out the two-dimensional silhouette of motion object from the described sport foreground extracted region of frame of video;
For each frame of video, calculate the focus point of described motion object two-dimensional silhouette;
The focus point of the motion object of all frame of video is transformed in the same image, obtains the barycenter trajectory of motion object.
2, according to claim 1ly extract the method for motion object barycenter trajectory, it is characterized in that the described static background that extracts sport video comprises from sport video with static background:
Ask the frame of video difference figure of adjacent video interframe;
Ask the stationary part among the frame of video difference figure, described stationary part comprises at least one segmentation;
Fill relevant position in the background with the pixel of the corresponding frame number of the longest segmentation mid point in the stationary part of frame of video difference figure, obtain static background.
3, according to claim 2ly extract the method for motion object barycenter trajectory, it is characterized in that, describedly ask frame of video difference figure between consecutive frame to comprise the luminance component between consecutive frame is subtracted each other from sport video with static background.
4, according to claim 3ly extract the method for motion object barycenter trajectory, it is characterized in that, describedly ask frame of video difference figure between consecutive frame also to comprise to set a threshold values to remove noise from sport video with static background.
5, the method for extracting motion object barycenter trajectory from sport video according to claim 1 with static background, it is characterized in that the two-dimensional silhouette that described sport foreground extracted region from frame of video goes out the motion object comprises denoising and/or removes non-motion object.
6, according to claim 5ly extract the method for motion object barycenter trajectory from the sport video with static background, it is characterized in that, described denoising comprises removes in the described sport foreground zone area less than the hole of a threshold values.
7, according to claim 6ly extract the method for motion object barycenter trajectory, it is characterized in that the non-motion object of described removal comprises removes the foreground area that is in non-middle section in the frame of video from sport video with static background.
8, according to claim 1ly extract the method for motion object barycenter trajectory, it is characterized in that the focus point of the described motion object of described calculating two-dimensional silhouette is a geometric center of asking motion object two-dimensional silhouette from sport video with static background.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 200410086740 CN1770204A (en) | 2004-10-29 | 2004-10-29 | Method for extracting barycenter trajectory of motive object from motive video with static background |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 200410086740 CN1770204A (en) | 2004-10-29 | 2004-10-29 | Method for extracting barycenter trajectory of motive object from motive video with static background |
Publications (1)
Publication Number | Publication Date |
---|---|
CN1770204A true CN1770204A (en) | 2006-05-10 |
Family
ID=36751480
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN 200410086740 Pending CN1770204A (en) | 2004-10-29 | 2004-10-29 | Method for extracting barycenter trajectory of motive object from motive video with static background |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN1770204A (en) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2008154825A1 (en) * | 2007-06-21 | 2008-12-24 | Tencent Technology (Shenzhen) Company Limited | A method and device for extracting a background image in a motion image |
CN102339625A (en) * | 2011-09-20 | 2012-02-01 | 清华大学 | Video object level time domain editing method and system |
CN102906788A (en) * | 2010-05-21 | 2013-01-30 | 松下电器产业株式会社 | Traffic line creation device and traffic line creation method |
CN103776604A (en) * | 2014-02-13 | 2014-05-07 | 山东理工大学 | Trampoline impacting ball detecting method |
CN104808669A (en) * | 2015-04-28 | 2015-07-29 | 苏州科技学院 | Fully-monitored automatic scoring intelligent trolley competition platform |
WO2018028363A1 (en) * | 2016-08-09 | 2018-02-15 | 深圳光启合众科技有限公司 | Target object tracking method and device |
US10440239B1 (en) | 2018-10-01 | 2019-10-08 | Interra Systems | System and method for detecting presence of a living hold in a video stream |
CN110989285A (en) * | 2014-04-22 | 2020-04-10 | 日本电信电话株式会社 | Video generation device, video generation method, data structure, and program |
-
2004
- 2004-10-29 CN CN 200410086740 patent/CN1770204A/en active Pending
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2008154825A1 (en) * | 2007-06-21 | 2008-12-24 | Tencent Technology (Shenzhen) Company Limited | A method and device for extracting a background image in a motion image |
CN102906788A (en) * | 2010-05-21 | 2013-01-30 | 松下电器产业株式会社 | Traffic line creation device and traffic line creation method |
US8934671B2 (en) | 2010-05-21 | 2015-01-13 | Panasonic Corporation | Traffic line creation device and traffic line creation method |
CN102906788B (en) * | 2010-05-21 | 2016-06-08 | 松下电器产业株式会社 | Moving-wire producing device and moving-wire manufacture method |
CN102339625A (en) * | 2011-09-20 | 2012-02-01 | 清华大学 | Video object level time domain editing method and system |
CN102339625B (en) * | 2011-09-20 | 2014-07-30 | 清华大学 | Video object level time domain editing method and system |
CN103776604A (en) * | 2014-02-13 | 2014-05-07 | 山东理工大学 | Trampoline impacting ball detecting method |
CN103776604B (en) * | 2014-02-13 | 2016-02-10 | 山东理工大学 | Trampoline impacts ball detection method |
CN110989285A (en) * | 2014-04-22 | 2020-04-10 | 日本电信电话株式会社 | Video generation device, video generation method, data structure, and program |
US11036123B2 (en) | 2014-04-22 | 2021-06-15 | Nippon Telegraph And Telephone Corporation | Video presentation device, method thereof, and recording medium |
CN104808669A (en) * | 2015-04-28 | 2015-07-29 | 苏州科技学院 | Fully-monitored automatic scoring intelligent trolley competition platform |
CN107730534A (en) * | 2016-08-09 | 2018-02-23 | 深圳光启合众科技有限公司 | The tracking and device of destination object |
CN107730534B (en) * | 2016-08-09 | 2020-10-23 | 深圳光启合众科技有限公司 | Target object tracking method and device |
WO2018028363A1 (en) * | 2016-08-09 | 2018-02-15 | 深圳光启合众科技有限公司 | Target object tracking method and device |
US10440239B1 (en) | 2018-10-01 | 2019-10-08 | Interra Systems | System and method for detecting presence of a living hold in a video stream |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN1131439C (en) | Target tracking method and device therefor | |
CN102098440B (en) | Electronic image stabilizing method and electronic image stabilizing system aiming at moving object detection under camera shake | |
CN110782477A (en) | Moving target rapid detection method based on sequence image and computer vision system | |
CN110992288B (en) | Video image blind denoising method used in mine shaft environment | |
CN108200432A (en) | A kind of target following technology based on video compress domain | |
CN102917220A (en) | Dynamic background video object extraction based on hexagon search and three-frame background alignment | |
CN111753693B (en) | Target detection method under static scene | |
CN102917217B (en) | Movable background video object extraction method based on pentagonal search and three-frame background alignment | |
CN1770204A (en) | Method for extracting barycenter trajectory of motive object from motive video with static background | |
CN1766928A (en) | A kind of motion object center of gravity track extraction method based on the dynamic background sport video | |
CN115375733A (en) | Snow vehicle sled three-dimensional sliding track extraction method based on videos and point cloud data | |
Pan et al. | Single-image dehazing via dark channel prior and adaptive threshold | |
CN103632373B (en) | A kind of flco detection method of three-frame difference high-order statistic combination OTSU algorithms | |
CN108765463A (en) | A kind of moving target detecting method calmodulin binding domain CaM extraction and improve textural characteristics | |
CN100337472C (en) | Video composing method with motion prospect | |
CN1873656A (en) | Detection method of natural target in robot vision navigation | |
CN107464220B (en) | Highway surface layer disease image enhancement method based on gravity superposition model | |
CN103051893B (en) | Dynamic background video object extraction based on pentagonal search and five-frame background alignment | |
CN105741317A (en) | Infrared moving target detection method based on time-space domain saliency analysis and sparse representation | |
CN102917218B (en) | Movable background video object extraction method based on self-adaptive hexagonal search and three-frame background alignment | |
CN108875630B (en) | Moving target detection method based on video in rainy environment | |
CN102917224B (en) | Mobile background video object extraction method based on novel crossed diamond search and five-frame background alignment | |
CN111862152A (en) | Moving target detection method based on interframe difference and super-pixel segmentation | |
CN111242983A (en) | Moving object detection method adopting statistical significance background subtraction method | |
Shi et al. | An improved method of removing fog and haze effect from images |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
WD01 | Invention patent application deemed withdrawn after publication |