CN105139418A - Novel video tracking method based on partitioning policy - Google Patents
Novel video tracking method based on partitioning policy Download PDFInfo
- Publication number
- CN105139418A CN105139418A CN201510471102.6A CN201510471102A CN105139418A CN 105139418 A CN105139418 A CN 105139418A CN 201510471102 A CN201510471102 A CN 201510471102A CN 105139418 A CN105139418 A CN 105139418A
- Authority
- CN
- China
- Prior art keywords
- fritter
- color histogram
- hist
- candidate blocks
- represent
- 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
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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30232—Surveillance
Landscapes
- Image Analysis (AREA)
Abstract
The present invention relates to a novel video tracking method based on a partitioning policy. The method comprises: dividing a target region into multiple patches; using similarity between a color histogram of each patch and color histograms of surrounding patches of the patch to indicate magnitude of change of color around the current patch, so as to determine whether the current patch and surrounding patches are subjected to blocking; and if the current patch and the surrounding patches are subjected to blocking, when an inter-patch relationship change coefficient BRC is calculated in a following particle filter algorithm, decreasing a weight of the current patch, thereby decreasing an effect of the blocking on the accuracy of the tracking method. The method provided by the present invention is capable of tracking a target region in a video with high efficiency, high speed and accuracy.
Description
Technical field
The present invention relates to a kind of novel video tracing method based on partition strategy, belong to technical field of computer vision.
Background technology
Nowadays, video tracking has become the large focus of one in computer vision research, and has been widely used in the aspects such as traffic administration, digital supervision and intelligent city.At present military science and technology and civilian scientific and technological in, video tracking has become a noticeable field, obtains heavy demand and application in numerous occasion, and therefore, video tracking has manifested huge market potential.The application of video tracking technology in civilian science and technology, is mainly distributed in intelligent monitoring aspect, such as, and the supervision of public arena and private context, people's flow monitoring in big assembly place, the security protection target following of important place; In addition, in intelligent transportation field, comprising: vehicle tracking monitoring, vehicle count, intelligent driving etc.: in computing machine and the communications field, comprising: man-machine interaction, data compression, virtual reality technology etc.; The tracking application of military field video object is more extensive and complicated, comprise: the TV navigation of robot guidance, aircraft, the warning inspection etc. of critical facility, video tracking technology and system are had higher requirement, various battlefield surroundings complicated and changeable and bad working environments must be met, stable and reliable trace information is provided.
Video tracking refers to by the analysis to picture frame, obtains the position of institute's tracking target in video sequence and to go forward side by side the method for row labels.But, in video tracking technology, such as block, to disturb and the series of problems of complex background etc. can have a strong impact on the accuracy of video tracking, in order to improve the accuracy of video tracking, researcher both domestic and external proposes and much in monitor video, resists the algorithm blocking and disturb.Traditional video tracking algorithm extracts the feature of target area and analyze, and in candidate region, find out the region of mating the most with target signature, as the result of following the tracks of.Nowadays conventional video tracing method has mean-shift method, Kalman filtering method, particle filter method scheduling algorithm.Wherein, particle filter algorithm due to its adapt to the condition of gaussian sum non-gaussian under target is followed the tracks of, the background of complexity is had simultaneously and follows the tracks of accuracy preferably, therefore, be widely used.
The cardinal principle of particle filter method is: first, extracts in initial pictures to the feature in the target area chosen, and is generally color, Texture eigenvalue.Particle is sowed at random afterwards near next frame objective area in image, each particle represents a candidate region, respectively feature extraction is carried out to the candidate region representated by these particles, and mate with the feature of special target area of taking out at the beginning respectively, the highest candidate region of selected characteristic similarity is as the target area in next frame.And this region is used for the target area searched in next frame, by that analogy until terminate.
In the algorithm of video tracking, the extraction of feature and coupling are the keys that whole algorithm accuracy ensures.Feature extraction only for target area is accurately reasonable, and characteristic matching process is afterwards just meaningful.Therefore, the extraction of target area feature requires as far as possible accurately and comprehensively to target to show and describe, and reduces the interference of other factors.And in traditional particle filter algorithm, normally extract the feature of the information such as color and texture as region of target area.Well-known color characteristic is substantially the most also the most significant feature as target, can very intuitively and be described target accurately.But meanwhile, color characteristic is easily blocked and the impact of the factor such as shape sudden change.Such as, target area in video tracking process is a pedestrian be dressed in white, when pedestrian passes by an electric pole, pedestrian's part can even all be blocked by electric pole, the colouring information of pedestrian will be lost, color histogram cannot have matched as feature, and therefore eclipse phenomena can make the accuracy of conventional video tracking reduce.
For the eclipse phenomena generally occurred in video, the method having some experts and scholar to propose piecemeal before this solves, several little blocks are divided into by target area, calculate the color histogram information in each block respectively, and carry out the coupling of feature respectively, and the similarity of characteristic matching is summed up, show that total similarity is for judging target area.But in video, once target is blocked by obstructions, so the colouring information of shelter then can replace the colouring information of original target by as feature extraction out, and then affects the feature accuracy of entirety of target area, and the method for piecemeal can not address this problem well.
Summary of the invention
For the deficiencies in the prior art, the invention discloses a kind of novel video tracing method based on partition strategy;
Target area is divided into several fritters by the present invention, considers to there is relevance between each fritter, utilize each fritter color histogram and around it fritter color histogram between similarity
Represent the severe degree that current block ambient color changes, judge whether current block and surrounding fritter eclipse phenomena occur with this.If current block and around fritter there occurs eclipse phenomena, between computing block during relationship change coefficient B RC by particle filter algorithm below, reduce the weight shared by this current block, thus reduce eclipse phenomena to the impact of tracking accuracy.Present invention achieves efficiently, quickly and accurately in video to the tracking of target area.
Technical scheme of the present invention is:
The novel video tracing method based on partition strategy, described video comprises M two field picture, m=1, and in m two field picture, mark the target area needing to follow the tracks of, concrete steps comprise:
(1) target area is divided into the identical fritter of N number of size; And calculate the similarity of the color of described each fritter and surrounding fritter;
(2) in (m+1) two field picture, by particle filter algorithm with target area same position near shed particle, obtain several candidate regions, each candidate region is divided into the identical candidate blocks of N number of size, and calculates the similarity of the color of each candidate blocks and surrounding candidate blocks;
(3) variation factor between the similarity of the corresponding candidate blocks that the similarity of each fritter that calculation procedure (1) is asked for is asked for step (2);
(4) the color relationship variation factor of each fritter and the corresponding candidate blocks of candidate region is asked for, i.e. the average of some variation factors asked for of step (3);
(5) ask for the similarity between each candidate region and target area, get candidate region corresponding to wherein maximal value as target area, m adds 1, if m=M, terminates, otherwise, return step (1).
Preferred according to the present invention, in step (1), concrete steps comprise:
A, target area is divided into the identical fritter of N number of size;
B, ask for the color histogram of each fritter that step (1) divides;
C, adopt the color histogram of Pasteur Bhattacharyya formulae discovery step b each fritter and around it fritter color histogram between similarity: through type (I), (II), (III), (IV) calculate fritter n color histogram and around it fritter color histogram between similarity, n={1,2 ... N}, described surrounding fritter comprises four fritters of the top adjacent with fritter n, below, left, right:
BC
u=Bhattacharyya(Hist
0,Hist
u)(Ⅰ)
BC
d=Bhattacharyya(Hist
0,Hist
d)(Ⅱ)
BC
l=Bhattacharyya(Hist
0,Hist
l)(Ⅲ)
BC
r=Bhattacharyya(Hist
0,Hist
r)(Ⅳ)
In formula (I), (II), (III), (IV), BC
urepresent fritter n color histogram and above it fritter color histogram between similarity, Hist
0represent the color histogram of fritter n, Hist
uthen represent the color histogram of the top fritter of fritter n; BC
drepresent fritter n color histogram and below it fritter color histogram between similarity, Hist
drepresent the color histogram of the below fritter of fritter n; BC
lrepresent the similarity between the color histogram of fritter n and the color histogram of its left fritter, Hist
lrepresent the color histogram of the left fritter of fritter n; BC
rrepresent the similarity between the color histogram of fritter n and the color histogram of its right fritter, Hist
rrepresent the color histogram of the right fritter of fritter n.
Preferred according to the present invention, in step (2), in (m+1) two field picture, by particle filter algorithm with target area same position near shed particle, obtain T candidate region, perform steps d-f to each candidate region t: wherein, 1≤t≤T, concrete steps comprise:
D, candidate region is divided into the identical candidate blocks of the identical N number of size of the tile position identical with the N number of size described in step a;
E, ask for the color histogram of each candidate blocks that steps d divides;
F, adopt Pasteur Bhattacharyya formulae discovery step e each candidate blocks color histogram and around it candidate blocks color histogram between similarity: the color histogram of through type (V), (VI), (VII), (VIII) calculated candidate block n ' and around it candidate blocks color histogram between similarity, n '={ 1,2 ... N}, described surrounding candidate blocks comprises four candidate blocks of the top adjacent with candidate blocks n ', below, left, right:
BC′
u=Bhattacharyya(Hist′
0,Hist′
u)(Ⅴ)
BC′
d=Bhattacharyya(Hist′
0,Hist′
d)(Ⅵ)
BC′
l=Bhattacharyya(Hist′
0,Hist′
l)(Ⅶ)
BC′
r=Bhattacharyya(Hist′
0,Hist′
r)(Ⅷ)
In formula (V), (VI), (VII), (VIII), BC '
urepresent candidate blocks n ' color histogram and above it candidate blocks color histogram between similarity, Hist '
0represent the color histogram of candidate blocks n ', Hist '
urepresent the color histogram of the top candidate blocks of candidate blocks n '; BC '
drepresent candidate blocks n ' color histogram and below it candidate blocks color histogram between similarity, Hist '
drepresent the color histogram of the below candidate blocks of candidate blocks n '; BC '
lrepresent the similarity between the color histogram of candidate blocks n ' and the color histogram of its left candidate blocks, Hist '
lrepresent the color histogram of the left candidate blocks of candidate blocks n '; BC '
rrepresent the similarity between the color histogram of candidate blocks n ' and the color histogram of its right candidate blocks, Hist '
rrepresent the color histogram of the right candidate blocks of candidate blocks n '.
Preferred according to the present invention, in step (3), concrete steps comprise:
Through type (Ⅸ), (Ⅹ), (Ⅺ), (Ⅻ) calculate BC
u, BC
d, BC
l, BC
rrespectively with BC '
u, BC '
d, BC '
l, BC '
rbetween variation factor;
In formula (Ⅸ), (Ⅹ), (Ⅺ), (Ⅻ), BRC
urepresent BC
uwith BC '
uvariation factor; BRC
drepresent BC
dwith BC '
dvariation factor; BRC
lrepresent BC
lwith BC '
lvariation factor; BRC
rrepresent BC
rwith BC '
rvariation factor.
Preferred according to the present invention, in step (4), concrete steps comprise:
Through type (Ⅹ III) calculates the color relationship variation factor of each fritter n of target area and the candidate blocks n ' of candidate region: namely ask for BRC
u, BRC
d, BRC
l, BRC
raverage BRC:
BRC=average(BRC
u,BRC
d,BRC
l,BRC
r)(ⅩⅢ)
In formula (Ⅹ III), BRC represents the color relationship variation factor of fritter n and candidate blocks n '.
Preferred according to the present invention, in step (5), concrete steps comprise:
G, ask for the similarity between candidate region t and target area in (m+1) two field picture, computing formula is such as formula shown in (XIV):
In formula (XIV),
represent the similarity between candidate region t and target area in (m+1) two field picture, N represents the quantity of the fritter be divided in candidate region;
H, ask for the similarity in (m+1) two field picture between each candidate region and target area by step g, that is:
get
in candidate region corresponding to maximal value as target area, m adds 1, if m=M, terminates, otherwise, return step (1).
Preferred according to the present invention, 1≤N≤200.
The advantage herein designed is, reduces the complexity of calculating under the prerequisite of result precision ensureing video tracking.
Beneficial effect of the present invention is:
1, the related pixel information of the present invention to every two field picture of video takes into full account, the feature of the new target area proposed, namely relationship change coefficient B RC between block, identifies target area efficiently and accurately, follows the tracks of, and substantially increases the accuracy of video tracking;
2, the present invention efficiently solve occur in video tracking process block, the problem such as interference, keep high identification and follow the tracks of accuracy rate;
3, the present invention has high-intelligentization, high accuracy, high innovative advantage.
Accompanying drawing explanation
Fig. 1 is BRC schematic diagram of the present invention.
Embodiment
Below in conjunction with Figure of description and embodiment, the present invention is further qualified, but is not limited thereto.
Embodiment
The novel video tracing method based on partition strategy, described video comprises M two field picture, m=1, and in m two field picture, mark the target area needing to follow the tracks of, concrete steps comprise:
(1) target area is divided into the identical fritter of N number of size; And calculate the similarity of the color of described each fritter and surrounding fritter; Concrete steps comprise:
A, target area is divided into the identical fritter of N number of size;
B, ask for the color histogram of each fritter that step a divides;
C, adopt the color histogram of Pasteur Bhattacharyya formulae discovery step b each fritter and around it fritter color histogram between similarity: through type (I), (II), (III), (IV) calculate fritter n color histogram and around it fritter color histogram between similarity, n={1,2 ... N}, described surrounding fritter comprises four fritters of the top adjacent with fritter n, below, left, right:
BC
u=Bhattacharyya(Hist
0,Hist
u)(Ⅰ)
BC
d=Bhattacharyya(Hist
0,Hist
d)(Ⅱ)
BC
l=Bhattacharyya(Hist
0,Hist
l)(Ⅲ)
BC
r=Bhattacharyya(Hist
0,Hist
r)(Ⅳ)
In formula (I), (II), (III), (IV), BC
urepresent fritter n color histogram and above it fritter color histogram between similarity, Hist
0represent the color histogram of fritter n, Hist
uthen represent the color histogram of the top fritter of fritter n; BC
drepresent fritter n color histogram and below it fritter color histogram between similarity, Hist
drepresent the color histogram of the below fritter of fritter n; BC
lrepresent the similarity between the color histogram of fritter n and the color histogram of its left fritter, Hist
lrepresent the color histogram of the left fritter of fritter n; BC
rrepresent the similarity between the color histogram of fritter n and the color histogram of its right fritter, Hist
rrepresent the color histogram of the right fritter of fritter n;
(2) in (m+1) two field picture, by particle filter algorithm with target area same position near shed particle, obtain several candidate regions, each candidate region is divided into the identical candidate blocks of N number of size, and calculates the similarity of the color of each candidate blocks and surrounding candidate blocks; Concrete steps comprise:
D, candidate region is divided into the identical candidate blocks of the identical N number of size of the tile position identical with the N number of size described in step a;
E, ask for the color histogram of each candidate blocks that steps d divides;
F, adopt Pasteur Bhattacharyya formulae discovery step e each candidate blocks color histogram and around it candidate blocks color histogram between similarity: the color histogram of through type (V), (VI), (VII), (VIII) calculated candidate block n ' and around it candidate blocks color histogram between similarity, n '={ 1,2 ... N}, described surrounding candidate blocks comprises four candidate blocks of the top adjacent with candidate blocks n ', below, left, right:
BC′
u=Bhattacharyya(Hist′
0,Hist′
u)(Ⅴ)
BC′
d=Bhattacharyya(Hist′
0,Hist′
d)(Ⅵ)
BC′
l=Bhattacharyya(Hist′
0,Hist′
l)(Ⅶ)
BC′
r=Bhattacharyya(Hist′
0,Hist′
r)(Ⅷ)
In formula (V), (VI), (VII), (VIII), BC '
urepresent candidate blocks n ' color histogram and above it candidate blocks color histogram between similarity, Hist '
0represent the color histogram of candidate blocks n ', Hist '
urepresent the color histogram of the top candidate blocks of candidate blocks n '; BC '
drepresent candidate blocks n ' color histogram and below it candidate blocks color histogram between similarity, Hist '
drepresent the color histogram of the below candidate blocks of candidate blocks n '; BC '
lrepresent the similarity between the color histogram of candidate blocks n ' and the color histogram of its left candidate blocks, Hist '
lrepresent the color histogram of the left candidate blocks of candidate blocks n '; BC '
rrepresent the similarity between the color histogram of candidate blocks n ' and the color histogram of its right candidate blocks, Hist '
rrepresent the color histogram of the right candidate blocks of candidate blocks n ';
(3) variation factor between the similarity of the corresponding candidate blocks that the similarity of each fritter that calculation procedure (1) is asked for is asked for step (2); Concrete steps comprise:
Through type (Ⅸ), (Ⅹ), (Ⅺ), (Ⅻ) calculate BC
u, BC
d, BC
l, BC
rrespectively with BC '
u, BC '
d, BC '
l, BC '
rbetween variation factor;
In formula (Ⅸ), (Ⅹ), (Ⅺ), (Ⅻ), BRC
urepresent BC
uwith BC '
uvariation factor; BRC
drepresent BC
dwith BC '
dvariation factor; BRC
lrepresent BC
lwith BC '
lvariation factor; BRC
rrepresent BC
rwith BC '
rvariation factor.
(4) the color relationship variation factor of each fritter and the corresponding candidate blocks of candidate region is asked for, i.e. the average of some variation factors asked for of step (3); Concrete steps comprise:
Through type (Ⅹ III) calculates the color relationship variation factor of each fritter n of target area and the candidate blocks n ' of candidate region: namely ask for BRC
u, BRC
d, BRC
l, BRC
raverage BRC:
BRC=average(BRC
u,BRC
d,BRC
l,BRC
r)(ⅩⅢ)
In formula (Ⅹ III), BRC represents the color relationship variation factor of fritter n and candidate blocks n '.
(5) ask for the similarity between each candidate region and target area, get candidate region corresponding to wherein maximal value as target area, m adds 1, if m=M, terminate, otherwise return step (1), concrete steps comprises:
G, ask for the similarity between candidate region t and target area in (m+1) two field picture, computing formula is such as formula shown in (XIV):
In formula (XIV),
represent the similarity between candidate region t and target area in (m+1) two field picture, N represents the quantity of the fritter be divided in candidate region;
H, ask for the similarity in (m+1) two field picture between each candidate region and target area by step g, that is:
get
in candidate region corresponding to maximal value as target area, m adds 1, if m=M, terminates, otherwise, return step (1).
Wherein, N=200.
Claims (7)
1. the novel video tracing method based on partition strategy, described video comprises M two field picture, m=1, and in m two field picture, mark the target area needing to follow the tracks of, it is characterized in that, concrete steps comprise:
(1) target area is divided into the identical fritter of N number of size; And calculate the similarity of the color of described each fritter and surrounding fritter;
(2) in (m+1) two field picture, by particle filter algorithm with target area same position near shed particle, obtain several candidate regions, each candidate region is divided into the identical candidate blocks of N number of size, and calculates the similarity of the color of each candidate blocks and surrounding candidate blocks;
(3) variation factor between the similarity of the corresponding candidate blocks that the similarity of each fritter that calculation procedure (1) is asked for is asked for step (2);
(4) the color relationship variation factor of each fritter and the corresponding candidate blocks of candidate region is asked for, i.e. the average of some variation factors asked for of step (3);
(5) ask for the similarity between each candidate region and target area, get candidate region corresponding to wherein maximal value as target area, m adds 1, if m=M, terminates, otherwise, return step (1).
2. a kind of novel video tracing method based on partition strategy according to claim 1, it is characterized in that, in step (1), concrete steps comprise:
A, target area is divided into the identical fritter of N number of size;
B, ask for the color histogram of each fritter that step (1) divides;
C, adopt the color histogram of Pasteur Bhattacharyya formulae discovery step b each fritter and around it fritter color histogram between similarity: through type (I), (II), (III), (IV) calculate fritter n color histogram and around it fritter color histogram between similarity, n={1,2 ... N}, described surrounding fritter comprises four fritters of the top adjacent with fritter n, below, left, right:
BC
u=Bhattacharyya(Hist
0,Hist
u)(Ⅰ)
BC
d=Bhattacharyya(Hist
0,Hist
d)(Ⅱ)
BC
l=Bhattacharyya(Hist
0,Hist
l)(Ⅲ)
BC
r=Bhattacharyya(Hist
0,Hist
r)(Ⅳ)
In formula (I), (II), (III), (IV), BC
urepresent fritter n color histogram and above it fritter color histogram between similarity, Hist
0represent the color histogram of fritter n, Hist
uthen represent the color histogram of the top fritter of fritter n; BC
drepresent fritter n color histogram and below it fritter color histogram between similarity, Hist
drepresent the color histogram of the below fritter of fritter n; BC
lrepresent the similarity between the color histogram of fritter n and the color histogram of its left fritter, Hist
lrepresent the color histogram of the left fritter of fritter n; BC
rrepresent the similarity between the color histogram of fritter n and the color histogram of its right fritter, Hist
rrepresent the color histogram of the right fritter of fritter n.
3. a kind of novel video tracing method based on partition strategy according to claim 2, it is characterized in that, in step (2), in (m+1) two field picture, by particle filter algorithm with target area same position near shed particle, obtain T candidate region, steps d-f performed to each candidate region t: wherein, 1≤t≤T, concrete steps comprise:
D, candidate region is divided into the identical candidate blocks of the identical N number of size of the tile position identical with the N number of size described in step a;
E, ask for the color histogram of each candidate blocks that steps d divides;
F, adopt Pasteur Bhattacharyya formulae discovery step e each candidate blocks color histogram and around it candidate blocks color histogram between similarity: the color histogram of through type (V), (VI), (VII), (VIII) calculated candidate block n ' and around it candidate blocks color histogram between similarity, n '={ 1,2 ... N}, described surrounding candidate blocks comprises four candidate blocks of the top adjacent with candidate blocks n ', below, left, right:
BC′
u=Bhattacharyya(Hist′
0,Hist′
u)(Ⅴ)
BC′
d=Bhattacharyya(Hist′
0,Hist′
d)(Ⅵ)
BC′
l=Bhattacharyya(Hist′
0,Hist′
l)(Ⅶ)
BC′
r=Bhattacharyya(Hist′
0,Hist′
r)(Ⅷ)
In formula (V), (VI), (VII), (VIII), BC '
urepresent candidate blocks n ' color histogram and above it candidate blocks color histogram between similarity, Hist '
0represent the color histogram of candidate blocks n ', Hist '
urepresent the color histogram of the top candidate blocks of candidate blocks n '; BC '
drepresent candidate blocks n ' color histogram and below it candidate blocks color histogram between similarity, Hist '
drepresent the color histogram of the below candidate blocks of candidate blocks n '; BC '
lrepresent the similarity between the color histogram of candidate blocks n ' and the color histogram of its left candidate blocks, Hist '
lrepresent the color histogram of the left candidate blocks of candidate blocks n '; BC '
rrepresent the similarity between the color histogram of candidate blocks n ' and the color histogram of its right candidate blocks, Hist '
rrepresent the color histogram of the right candidate blocks of candidate blocks n '.
4. a kind of novel video tracing method based on partition strategy according to claim 3, it is characterized in that, in step (3), concrete steps comprise:
Through type (Ⅸ), (Ⅹ), (Ⅺ), (Ⅻ) calculate BC
u, BC
d, BC
l, BC
rrespectively with BC '
u, BC '
d, BC '
l, BC '
rbetween variation factor;
In formula (Ⅸ), (Ⅹ), (Ⅺ), (Ⅻ), BRC
urepresent BC
uwith BC '
uvariation factor; BRC
drepresent BC
dwith BC '
dvariation factor; BRC
lrepresent BC
lwith BC '
lvariation factor; BRC
rrepresent BC
rwith BC '
rvariation factor.
5. a kind of novel video tracing method based on partition strategy according to claim 4, it is characterized in that, in step (4), concrete steps comprise:
Through type (Ⅹ III) calculates the color relationship variation factor of each fritter n of target area and the candidate blocks n ' of candidate region: namely ask for BRC
u, BRC
d, BRC
l, BRC
raverage BRC:
BRC=average(BRC
u,BRC
d,BRC
l,BRC
r)(ⅩⅢ)
In formula (Ⅹ III), BRC represents the color relationship variation factor of fritter n and candidate blocks n '.
6. a kind of novel video tracing method based on partition strategy according to claim 5, it is characterized in that, in step (5), concrete steps comprise:
G, ask for the similarity between candidate region t and target area in (m+1) two field picture, computing formula is such as formula shown in (XIV):
In formula (XIV),
represent the similarity between candidate region t and target area in (m+1) two field picture, N represents the quantity of the fritter be divided in candidate region;
H, ask for the similarity in (m+1) two field picture between each candidate region and target area by step g, that is:
get
in candidate region corresponding to maximal value as target area, m adds 1, if m=M, terminates, otherwise, return step (1).
7., according to the arbitrary described a kind of novel video tracing method based on partition strategy of claim 1-6, it is characterized in that, 1≤N≤200.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510471102.6A CN105139418A (en) | 2015-08-04 | 2015-08-04 | Novel video tracking method based on partitioning policy |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510471102.6A CN105139418A (en) | 2015-08-04 | 2015-08-04 | Novel video tracking method based on partitioning policy |
Publications (1)
Publication Number | Publication Date |
---|---|
CN105139418A true CN105139418A (en) | 2015-12-09 |
Family
ID=54724750
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510471102.6A Pending CN105139418A (en) | 2015-08-04 | 2015-08-04 | Novel video tracking method based on partitioning policy |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105139418A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106023260A (en) * | 2016-05-26 | 2016-10-12 | 无锡天脉聚源传媒科技有限公司 | Television video target tracking method and apparatus |
CN106023261A (en) * | 2016-06-01 | 2016-10-12 | 无锡天脉聚源传媒科技有限公司 | TV video target tracking method and TV video target tracking device |
CN107590838A (en) * | 2017-08-18 | 2018-01-16 | 陕西维视数字图像技术有限公司 | A kind of metal surface colour vision detecting system |
CN107742277A (en) * | 2017-10-11 | 2018-02-27 | 上海电力学院 | MPS analogy methods based on partition strategy |
CN109410245A (en) * | 2018-09-13 | 2019-03-01 | 北京米文动力科技有限公司 | A kind of video target tracking method and equipment |
CN110659566A (en) * | 2019-08-15 | 2020-01-07 | 重庆特斯联智慧科技股份有限公司 | Target tracking method and system in shielding state |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101339655A (en) * | 2008-08-11 | 2009-01-07 | 浙江大学 | Visual sense tracking method based on target characteristic and bayesian filtering |
CN103440645A (en) * | 2013-08-16 | 2013-12-11 | 东南大学 | Target tracking algorithm based on self-adaptive particle filter and sparse representation |
CN103646407A (en) * | 2013-12-26 | 2014-03-19 | 中国科学院自动化研究所 | Video target tracking method based on ingredient and distance relational graph |
-
2015
- 2015-08-04 CN CN201510471102.6A patent/CN105139418A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101339655A (en) * | 2008-08-11 | 2009-01-07 | 浙江大学 | Visual sense tracking method based on target characteristic and bayesian filtering |
CN103440645A (en) * | 2013-08-16 | 2013-12-11 | 东南大学 | Target tracking algorithm based on self-adaptive particle filter and sparse representation |
CN103646407A (en) * | 2013-12-26 | 2014-03-19 | 中国科学院自动化研究所 | Video target tracking method based on ingredient and distance relational graph |
Non-Patent Citations (2)
Title |
---|
BIN SUN ET AL.: "Robust and Adaptive Block Tracking Method based on particle filter", 《EAI ENDORSED TRANSACTIONS ON COGNITIVE COMMUNICATIONS》 * |
L CHUNXI ET AL.: "Robust Object Tracking with Adaptive Fusion of Color and Edge Strength Local Mean Features Based on Particle Filter", 《INTERNATIONAL FORUM ON INFORMATION TECHNOLOGY&APPLICATIONS》 * |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106023260A (en) * | 2016-05-26 | 2016-10-12 | 无锡天脉聚源传媒科技有限公司 | Television video target tracking method and apparatus |
CN106023261A (en) * | 2016-06-01 | 2016-10-12 | 无锡天脉聚源传媒科技有限公司 | TV video target tracking method and TV video target tracking device |
CN106023261B (en) * | 2016-06-01 | 2019-11-29 | 无锡天脉聚源传媒科技有限公司 | A kind of method and device of television video target following |
CN107590838A (en) * | 2017-08-18 | 2018-01-16 | 陕西维视数字图像技术有限公司 | A kind of metal surface colour vision detecting system |
CN107742277A (en) * | 2017-10-11 | 2018-02-27 | 上海电力学院 | MPS analogy methods based on partition strategy |
CN109410245A (en) * | 2018-09-13 | 2019-03-01 | 北京米文动力科技有限公司 | A kind of video target tracking method and equipment |
CN109410245B (en) * | 2018-09-13 | 2021-08-10 | 北京米文动力科技有限公司 | Video target tracking method and device |
CN110659566A (en) * | 2019-08-15 | 2020-01-07 | 重庆特斯联智慧科技股份有限公司 | Target tracking method and system in shielding state |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105139418A (en) | Novel video tracking method based on partitioning policy | |
CA3032487C (en) | Saliency-based method for extracting road target from night vision infrared image | |
Caraffi et al. | Off-road path and obstacle detection using decision networks and stereo vision | |
CN110533687B (en) | Multi-target three-dimensional track tracking method and device | |
Chen et al. | Boost 3-D object detection via point clouds segmentation and fused 3-D GIoU-L₁ loss | |
CN103281477A (en) | Multi-level characteristic data association-based multi-target visual tracking method | |
CN102243765A (en) | Multi-camera-based multi-objective positioning tracking method and system | |
CN103268480A (en) | System and method for visual tracking | |
Wang et al. | An overview of 3d object detection | |
Zhou et al. | Vehicle detection and disparity estimation using blended stereo images | |
Tang et al. | Multiple-kernel based vehicle tracking using 3D deformable model and camera self-calibration | |
Xu et al. | Dynamic obstacle detection based on panoramic vision in the moving state of agricultural machineries | |
CN111145211B (en) | Method for acquiring pixel height of head of upright pedestrian of monocular camera | |
Du et al. | Recognition of mobile robot navigation path based on K-means algorithm | |
CN105118071B (en) | A kind of video tracing method based on adaptive piecemeal | |
CN103578121A (en) | Motion detection method based on shared Gaussian model in disturbed motion environment | |
Shu et al. | Small moving object detection and tracking based on event signals | |
CN108469729A (en) | A kind of human body target identification and follower method based on RGB-D information | |
Zhang et al. | Bundle adjustment for monocular visual odometry based on detected traffic sign features | |
CN104573703A (en) | Method for quickly identifying power transmission line based on partial derivative distribution and boundary strategy | |
CN106530325A (en) | Multi-target visual detection and tracking method | |
Wang et al. | Dynamic object separation and removal in 3d point cloud map building | |
Xiang | A simultaneous object detection and tracking framework based on point cloud | |
Limin | Vehicle shape recovery and recognition using generic models | |
Lachachi et al. | LIDAR and Stereo-Camera fusion for reliable Road Extraction |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20151209 |
|
WD01 | Invention patent application deemed withdrawn after publication |