CN106991685A - A kind of algorithm being tracked under panning mode to football in ball match video - Google Patents
A kind of algorithm being tracked under panning mode to football in ball match video Download PDFInfo
- Publication number
- CN106991685A CN106991685A CN201610037570.7A CN201610037570A CN106991685A CN 106991685 A CN106991685 A CN 106991685A CN 201610037570 A CN201610037570 A CN 201610037570A CN 106991685 A CN106991685 A CN 106991685A
- Authority
- CN
- China
- Prior art keywords
- football
- track
- ball
- algorithm
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30241—Trajectory
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a kind of algorithm being tracked under panning mode to football in ball match video, 1) generation scene mark;2) detection of candidate's ball;3) generation of short track;4) correction and filtering of short track;5) it is long track to connect short track.A kind of algorithm being tracked under panning mode to football in ball match video that the present invention is provided, utilize some prior information (such as court sports ground positions in court in football, goal position, sportsman's average height etc.), and football several main movement forms (ground pass, pass of hanging oneself, sportsman dribbles), its corresponding motion feature is extracted, these motion features are added in the tracking of football, to improve the accuracy of tracking.
Description
Technical field
The present invention relates to a kind of algorithm for video tracking of competing, more particularly to one kind are right under panning mode
The algorithm that football is tracked in ball match video.
Background technology
When carrying out the unsupervised recording of ball match using common camera, due to the football speed mistake of motion
Hurry up, and the camera lens used quality problems, often result in high-speed motion football exist motion
Fuzzy, diplopia and ghost problem;With the situation that generation ball is contacted with sportsman or blocked by sportsman often,
And the area of football is smaller under panning mode, this causes general method for tracking target to ball match mesopodium
Ball tracking effect is poor.For the defect of football tracking technique in ball match video in the prior art, this hair
It is bright to provide a kind of algorithm being tracked under panning mode to football in ball match video, it is existing to overcome
There is above-mentioned deficiency present in technology.
The content of the invention
The present invention is to provide one kind in panorama mould to solve the technical scheme that above-mentioned technical problem is used
The algorithm being tracked under formula to football in ball match video, wherein:Including following 5 step:
1) generation scene mark, the prior information of scene is acquired using artificial mark or algorithm;
2) detection of candidate's ball;
(1) to panoramic video, background model is set up, prospect is therefrom partitioned into;
(2) the suspicious of candidate is obtained to being filtered at prospect using size, shape, colouring information
Football region;
(3) result being tracked using early stage to ball, predicts the position of present frame football, if should
There is prospect on position, be then added to the suspicious football region of candidate;
(4) prototype soccerballs come out to suspicious football region using offline machine learning, carry out football
Detection, regard the region for detecting football as candidate's football region;
3) generation of short track;
(1) using currently detected candidate's football as region to be tracked, the region is tracked, will
Result formation track in tracking, with the replacement that predicts the outcome if not tracking;If at continuous one section
Between do not track, then terminate the tracking of the track;
(2) to current detection and the extended area in the ball region traced into, human testing is carried out;
4) correction and filtering of short track;Using the prior information and the information of human testing of scene,
With reference to the characteristics of motion of football, the judgement to track carry out activity, according to expected activity to changing track
Corrected and filtered;
5) it is long track to connect short track;Using time sequencing and spatial relationship, short track is connected
For long track, more complete football pursuit path is formed.
A kind of above-mentioned algorithm being tracked under panning mode to football in ball match video, wherein:
Generate in scene annotation step, the prior information obtained to scene includes being partitioned into football pitch from video
Ground, extracts goal position from video, marks out in video pitch at a distance and nearby common ball
The height of member.
A kind of above-mentioned algorithm being tracked under panning mode to football in ball match video, wherein:
In the correction and filtration step of short track, wherein, the judgement to track carry out activity is as follows:
Centre coordinate such as ball in short track is approximately straight line, then is judged to ground pass;
Centre coordinate such as short trace ball is approximately the parabola that Open Side Down, then is judged to pass of hanging oneself;
There is human body in the immediate vicinity such as ball in short track, y-coordinate is far below sportsman's height at this, then
It is judged to sportsman's dribbling.
A kind of above-mentioned algorithm being tracked under panning mode to football in ball match video, wherein:
In the correction and filtration step of short track, wherein, carried out according to expected activity to changing track
Correction and filtering are as follows:
Such as according to short track, judge it is current as ground pass during, then to not meeting ground pass rule
Detection or tracking result, corrected or directly rejected;
Such as according to short track, judge it is current as pass of hanging oneself during, then to not meeting pass rule of hanging oneself
Detection or tracking result, corrected or directly rejected;
Such as according to short track, during judging the current dribbling as sportsman, then to not meeting the inspection that sportsman dribbles
Survey or tracking result, corrected or directly rejected.
The present invention has the advantages that relative to prior art:
Using court in football some prior informations (such as court sports ground position, goal position,
Sportsman's average height etc.), and football several main movement forms (ground pass, pass of hanging oneself,
Sportsman dribble), extract its corresponding motion feature, by these motion features be added to football with
In track, to improve the accuracy of tracking.
Brief description of the drawings
Fig. 1 a kind of is tracked under panning mode for what the present invention was provided to football in ball match video
The principle schematic of algorithm.
Embodiment
The invention will be further described with reference to the accompanying drawings and examples.
The present invention provides a kind of algorithm being tracked under panning mode to football in ball match video, tool
Body scheme includes:
Generate scene mark
The method learnt automatically using artificial mark or algorithm, acquires some prior informations of scene
(pitch is such as partitioned into from video, goal position is extracted from video, is marked out in video
Pitch distant place and the nearby height of court player).
The detection of candidate's ball
(1) to panoramic video, background model is set up, prospect is therefrom partitioned into;
(2) size is utilized, shape, the information such as color, to being filtered at prospect, obtains some times
The suspicious football region of choosing;
(3) result being tracked using early stage to ball, predicts the position of present frame football, if should
There is prospect on position, be then added to the suspicious football region of candidate;
(4) to suspicious football region, the prototype soccerballs come out using offline machine learning carry out foot
The detection of ball, regard the region for detecting football as candidate's football region.
The generation of short track
(1) using currently detected candidate's football as region to be tracked, the region is tracked (real
Border tracking go up+detect be combined obtain tracking result), by tracking result formed track.If
Do not track, with the replacement that predicts the outcome;If but continuous a period of time does not track, then terminates this
The tracking of track;
(2) to current detection and the extended area in the ball region traced into, human testing is carried out.
The correction and filtering of short track
Using the prior information and the information of human testing of scene, with reference to the characteristics of motion of football, to rail
The judgement of mark carry out activity, is corrected and is filtered to changing track according to expected activity.
Wherein, it is as follows to the judgement of track carry out activity:
Centre coordinate such as ball in short track is approximately straight line, then is judged to ground pass;
Centre coordinate such as short trace ball is approximately the parabola that Open Side Down, then is judged to pass of hanging oneself;
There is human body in the immediate vicinity such as ball in short track, y-coordinate is far below sportsman's height at this, then
It is judged to sportsman's dribbling.
Wherein, corrected and filtered to changing track according to according to expected activity, it is as follows:
Such as according to short track, judge it is current as ground pass during, then to not meeting ground pass rule
Detection or tracking result (centre coordinate of ball is approximately straight line), corrected or directly rejected;
Such as according to short track, judge it is current as pass of hanging oneself during, then to not meeting pass rule of hanging oneself
Detection or tracking result (track of ball should be approximately the parabola that Open Side Down), corrected or
Directly reject;
Such as according to short track, during judging the current dribbling as sportsman, then to not meeting the inspection that sportsman dribbles
Survey or tracking result (the center y-coordinate of ball far below this at sportsman's height), corrected or directly
Reject;
Short track is connected for long track
Using time sequencing and spatial relationship, short track is connected as long track, formed more complete
Football pursuit path.
Although the present invention is disclosed as above with preferred embodiment, so it is not limited to the present invention, appoints
What those skilled in the art, without departing from the spirit and scope of the present invention, when a little modification can be made
With it is perfect, therefore protection scope of the present invention is when by being defined that claims are defined.
Claims (4)
1. a kind of algorithm being tracked under panning mode to football in ball match video, its feature exists
In:Including following 5 step:
1) generation scene mark, the prior information of scene is acquired using artificial mark or algorithm;
2) detection of candidate's ball;
(1) to panoramic video, background model is set up, prospect is therefrom partitioned into;
(2) the suspicious of candidate is obtained to being filtered at prospect using size, shape, colouring information
Football region;
(3) result being tracked using early stage to ball, predicts the position of present frame football, if should
There is prospect on position, be then added to the suspicious football region of candidate;
(4) prototype soccerballs come out to suspicious football region using offline machine learning, carry out football
Detection, regard the region for detecting football as candidate's football region;
3) generation of short track;
(1) using currently detected candidate's football as region to be tracked, the region is tracked, will
Result formation track in tracking, with the replacement that predicts the outcome if not tracking;If at continuous one section
Between do not track, then terminate the tracking of the track;
(2) to current detection and the extended area in the ball region traced into, human testing is carried out;
4) correction and filtering of short track;Using the prior information and the information of human testing of scene,
With reference to the characteristics of motion of football, the judgement to track carry out activity, according to expected activity to changing track
Corrected and filtered;
5) it is long track to connect short track;Using time sequencing and spatial relationship, short track is connected
For long track, more complete football pursuit path is formed.
2. one kind as claimed in claim 1 football in ball match video is carried out under panning mode with
The algorithm of track, it is characterised in that:Generate in scene annotation step, obtain the prior information bag to scene
Include and pitch is partitioned into from video, goal position is extracted from video, mark out video mesopodium
Court at a distance and the nearby height of court player.
3. one kind as claimed in claim 2 football in ball match video is carried out under panning mode with
The algorithm of track, it is characterised in that:In the correction and filtration step of short track, wherein, track is entered
The judgement of row activity is as follows:
Centre coordinate such as ball in short track is approximately straight line, then is judged to ground pass;
Centre coordinate such as short trace ball is approximately the parabola that Open Side Down, then is judged to pass of hanging oneself;
There is human body in the immediate vicinity such as ball in short track, y-coordinate is then sentenced less than sportsman's height at this
Dribbled for sportsman.
4. one kind as claimed in claim 3 football in ball match video is carried out under panning mode with
The algorithm of track, it is characterised in that:
In the correction and filtration step of short track, wherein, carried out according to expected activity to changing track
Correction and filtering are as follows:
Such as according to short track, judge it is current as ground pass during, then to not meeting ground pass rule
Detection or tracking result, corrected or directly rejected;
Such as according to short track, judge it is current as pass of hanging oneself during, then to not meeting pass rule of hanging oneself
Detection or tracking result, corrected or directly rejected;
Such as according to short track, during judging the current dribbling as sportsman, then to not meeting the inspection that sportsman dribbles
Survey or tracking result, corrected or directly rejected.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610037570.7A CN106991685B (en) | 2016-01-20 | 2016-01-20 | Algorithm for tracking football in ball game video in panoramic mode |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610037570.7A CN106991685B (en) | 2016-01-20 | 2016-01-20 | Algorithm for tracking football in ball game video in panoramic mode |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106991685A true CN106991685A (en) | 2017-07-28 |
CN106991685B CN106991685B (en) | 2020-09-25 |
Family
ID=59413626
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610037570.7A Active CN106991685B (en) | 2016-01-20 | 2016-01-20 | Algorithm for tracking football in ball game video in panoramic mode |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106991685B (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108009504A (en) * | 2017-12-04 | 2018-05-08 | 深圳市赢世体育科技有限公司 | A kind of recognition methods of moving sphere, device and storage medium |
CN108090421A (en) * | 2017-11-30 | 2018-05-29 | 睿视智觉(深圳)算法技术有限公司 | A kind of sportsman's competitive ability analysis method |
CN110543856A (en) * | 2019-09-05 | 2019-12-06 | 新华智云科技有限公司 | Football shooting time identification method and device, storage medium and computer equipment |
WO2020061875A1 (en) * | 2018-09-27 | 2020-04-02 | Intel Corporation | Highlight moment identification technology in volumetric content creation systems |
CN113781523A (en) * | 2021-09-13 | 2021-12-10 | 浙江大学 | Football detection tracking method and device, electronic equipment and storage medium |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101794451A (en) * | 2010-03-12 | 2010-08-04 | 上海交通大学 | Tracing method based on motion track |
CN102142147A (en) * | 2010-01-29 | 2011-08-03 | 索尼公司 | Device and method for analyzing site content as well as device and method for detecting and tracking target |
CN102819749A (en) * | 2012-07-23 | 2012-12-12 | 西安体育学院 | Automatic identification system and method for offside of football based on video analysis |
CN104881882A (en) * | 2015-04-17 | 2015-09-02 | 广西科技大学 | Moving target tracking and detection method |
-
2016
- 2016-01-20 CN CN201610037570.7A patent/CN106991685B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102142147A (en) * | 2010-01-29 | 2011-08-03 | 索尼公司 | Device and method for analyzing site content as well as device and method for detecting and tracking target |
CN101794451A (en) * | 2010-03-12 | 2010-08-04 | 上海交通大学 | Tracing method based on motion track |
CN102819749A (en) * | 2012-07-23 | 2012-12-12 | 西安体育学院 | Automatic identification system and method for offside of football based on video analysis |
CN104881882A (en) * | 2015-04-17 | 2015-09-02 | 广西科技大学 | Moving target tracking and detection method |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108090421A (en) * | 2017-11-30 | 2018-05-29 | 睿视智觉(深圳)算法技术有限公司 | A kind of sportsman's competitive ability analysis method |
CN108090421B (en) * | 2017-11-30 | 2021-10-08 | 睿视智觉(深圳)算法技术有限公司 | Athlete athletic ability analysis method |
CN108009504A (en) * | 2017-12-04 | 2018-05-08 | 深圳市赢世体育科技有限公司 | A kind of recognition methods of moving sphere, device and storage medium |
CN108009504B (en) * | 2017-12-04 | 2020-06-26 | 深圳市赢世体育科技有限公司 | Moving sphere identification method and device and storage medium |
WO2020061875A1 (en) * | 2018-09-27 | 2020-04-02 | Intel Corporation | Highlight moment identification technology in volumetric content creation systems |
US11610399B2 (en) | 2018-09-27 | 2023-03-21 | Intel Corporation | Highlight moment identification technology in volumetric content creation systems |
CN110543856A (en) * | 2019-09-05 | 2019-12-06 | 新华智云科技有限公司 | Football shooting time identification method and device, storage medium and computer equipment |
CN110543856B (en) * | 2019-09-05 | 2022-04-22 | 新华智云科技有限公司 | Football shooting time identification method and device, storage medium and computer equipment |
CN113781523A (en) * | 2021-09-13 | 2021-12-10 | 浙江大学 | Football detection tracking method and device, electronic equipment and storage medium |
CN113781523B (en) * | 2021-09-13 | 2024-04-26 | 浙江大学 | Football detection tracking method and device, electronic equipment and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN106991685B (en) | 2020-09-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106991685A (en) | A kind of algorithm being tracked under panning mode to football in ball match video | |
CN106991359A (en) | A kind of algorithm being tracked under panning mode to basketball in ball match video | |
CN106991357A (en) | The shooting of automatic identification Basketball Match and the algorithm scored based on panoramic video | |
US9437012B2 (en) | Multiple-object tracking and team identification for game strategy analysis | |
US8542874B2 (en) | Videotracking | |
CN106991358A (en) | The algorithm that automatic identification football match based on panoramic video is scored | |
CN107824940A (en) | Welding seam traking system and method based on laser structure light | |
US20050215337A1 (en) | Golf swing-measuring system | |
CN104063700B (en) | The method of eye center point location in natural lighting front face image | |
CN104715238A (en) | Pedestrian detection method based on multi-feature fusion | |
CN110490161B (en) | Captive animal behavior analysis method based on deep learning | |
CN104992453A (en) | Target tracking method under complicated background based on extreme learning machine | |
CN106570447B (en) | Based on the matched human face photo sunglasses automatic removal method of grey level histogram | |
CN105678749B (en) | A kind of fresh fish discriminating direction method of view-based access control model | |
Teachabarikiti et al. | Players tracking and ball detection for an automatic tennis video annotation | |
CN105205437B (en) | Side face detection method and device based on contouring head verifying | |
CN106780473A (en) | A kind of magnet ring defect multi-vision visual detection method and system | |
JPWO2007037350A1 (en) | Horse position information analysis and display method | |
CN110298860A (en) | A kind of high bar silk ball detection number system based on machine vision | |
Komorowski et al. | Deepball: Deep neural-network ball detector | |
CN111709365A (en) | Automatic human motion posture detection method based on convolutional neural network | |
CN106991356A (en) | The algorithm that sportsman is tracked in a kind of video to ball match | |
JP6902492B2 (en) | Object identification device | |
CN109949344A (en) | It is a kind of to suggest that the nuclear phase of window closes filter tracking method based on color probability target | |
Kurose et al. | Player pose analysis in tennis video based on pose estimation |
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 | ||
TR01 | Transfer of patent right |
Effective date of registration: 20220324 Address after: 200080 room 3001, building A2, No. 121, Zhongshan North 1st Road, Hongkou District, Shanghai Patentee after: Hupu (Shanghai) Information Technology Co.,Ltd. Address before: 200080 room 7179, ground floor, No. 203, Xinjian Road, Hongkou District, Shanghai Patentee before: SHANGHAI HUITI NETWORK TECHNOLOGY CO.,LTD. |
|
TR01 | Transfer of patent right |