CN103279952A - Target tracking method and device - Google Patents
Target tracking method and device Download PDFInfo
- Publication number
- CN103279952A CN103279952A CN2013101831741A CN201310183174A CN103279952A CN 103279952 A CN103279952 A CN 103279952A CN 2013101831741 A CN2013101831741 A CN 2013101831741A CN 201310183174 A CN201310183174 A CN 201310183174A CN 103279952 A CN103279952 A CN 103279952A
- Authority
- CN
- China
- Prior art keywords
- matched
- image
- reference picture
- unique point
- point
- 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
Images
Landscapes
- Image Analysis (AREA)
Abstract
The embodiment of the invention discloses a target tracking method, which comprises the steps of using a presetting algorithm to extract the corner feature of a reference image to obtain the first feature point set of the reference image, removing abnormal points which exceed the range of images to be matched from the first feature point set to obtain a second feature point set, matching the second feature point set with the images to be matched and finishing the target tracking of the images to be matched according to matching results. The embodiment of the invention additionally discloses a target tracking device. By using the target tracking method and the target tracking device, the extracted corner feature has good identifiability and stability, the matching speed is greatly improved, the feature point set of the reference image is optimized, the corners which exceed the range are removed, the problem that the matching cannot be realized due to a reason that the feature points exceed the range in the prior art is solved and the matching reliability is improved.
Description
Technical field
The present invention relates to image processing field, relate in particular to a kind of method for tracking target and device.
Background technology
Target detection technique is also referred to as the feature extraction of image usually, is the concept of computer vision and Digital Image Processing, and whether the pixel that it refers to use Computing to determine each image is the image information leaching process of characteristics of image.The target following technology is determined area-of-interest exactly in video image or consecutive image sequence (Region Of Interest ROI) is also referred to as the position of template (Template), and the template of each two field picture is mapped.
Existing tracking technique scheme can be described below:
(1) determine a template in the image sequence of gathering, be commonly referred to reference picture, recording in the reference picture needs tracked target;
(2) in order to improve the efficient of computing, each pixel in the reference picture is not done conversion, equidistantly pixel is evenly extracted but be separated by in the reference picture, this process can be called as equidistant sampling, forms the primitive character point set;
(3) primitive character point set and realm information thereof are done computing, obtain new feature point set, determine matching area according to new feature point set at image to be matched;
(4) half-tone information between calculating matching area and the reference template, SSD (Sum of Squared Differences for example, squared difference and) method, utilize the method for minimum error, pass through iteration, make and to mate between matching area and the reference picture, even have identical image information in matching area and the reference picture, namely need tracked target.
(5) tracking that target is continuous by the template matches between picture frame and the frame, has finally been realized in image sequence repeating step (3)-(4) to collecting.
But there is following shortcoming in prior art:
(1) it is fashionable to obtain the feature point set of reference picture, employing be the method that the pixel in the reference picture is equidistantly sampled.The unique point that obtains like this is random big, and the image information that comprises usually is less, good token image feature, and reliability, stability are not high, make track algorithm not possess good robustness.
(2) obtain reference picture after, just with it as standard form, no longer make any change.But in the tracker of reality, motion along with image capture devices such as video cameras, the reference picture that may determine can partly shift out the image acquisition scope of video camera, can not imaging, thus the subregion of reference picture is not in the middle of the image of successive image sequence.Yet existing scheme adopts initial reference picture, can make the gradation of image information algorithm that calculates between reference picture and the matching area to restrain, and can not obtain correct result, thereby causes following the tracks of failure.
Summary of the invention
Embodiment of the invention technical matters to be solved is, a kind of method for tracking target is provided.Can solve the deficiency that characteristics of image stability is not high and the super scope of reference picture causes following the tracks of failure in the prior art.
In order to solve the problems of the technologies described above, first aspect present invention provides a kind of method for tracking target, comprising:
First unique point set that angle point feature that algorithm extracts reference picture obtains described reference picture is preset in employing, and described reference picture is used for the target of current image to be matched is followed the tracks of;
Reject the abnormity point that exceeds described image range to be matched in described first unique point set, obtain the set of second unique point;
Described second unique point set is carried out images match with described image to be matched, finish the target following of described image to be matched according to matching result.
In first kind of possible implementation, the described algorithm that presets comprises: Harris Corner Detection Algorithm, FAST accelerate to cut apart the small nut value similar area Corner Detection Algorithm of detected characteristics Corner Detection Algorithm, KLT Corner Detection Algorithm or SUSAN.
In conjunction with first aspect or first kind of possible implementation, in second kind of possible implementation, described employing is preset before the step of first feature point set that angle point feature that algorithm extracts reference picture obtains described reference picture, also comprises:
Obtain the image sequence of camera acquisition, with this image sequence as image to be matched;
And from first two field picture of described image sequence, determine the reference picture of this image to be matched.
In conjunction with second kind of first aspect possible implementation, in the third possible implementation, exceed the abnormity point of described image range to be matched in described first unique point set of described rejecting, the step that obtains the set of second unique point comprises:
Obtain each characteristic point coordinates in described first unique point set;
Judge whether to exceed the scope of described image to be matched according to each characteristic point coordinates in the set of described unique point, if yes, determine that then this unique point is abnormity point;
With all abnormity point elimination in described first unique point set, obtain the set of second unique point.
The third possible implementation in conjunction with first aspect, in the 4th kind of possible implementation, described described second unique point is gathered with described image to be matched carried out images match, and the step of finishing the target following of described image to be matched according to matching result comprises:
In described image to be matched, choose a subgraph to be matched with the identical size of described reference picture;
Adopt the described algorithm that presets to put forward the feature of described subgraph to be matched and obtain the 3rd unique point set;
Calculate the distance between described second unique point set and the set of described the 3rd unique point;
Judge that whether described distance is less than preset threshold value, if yes, the match is successful then to determine subgraph to be matched in described reference picture and the described image to be matched, if deny, then re-executes and choose a step with the subgraph to be matched of the identical size of described reference picture in described image to be matched.
The 4th kind of possible implementation in conjunction with first aspect, in the 5th kind of possible implementation, described the set of described second unique point is carried out images match with described image to be matched, finishes according to matching result after the step of target following of described image to be matched, also comprise:
Delete described reference picture and will with the described reference template subgraph to be matched that the match is successful as new reference picture.
Correspondingly, second aspect present invention also provides a kind of target tracker, comprising:
Characteristic extracting module is used for adopting the angle point feature that presets algorithm extraction reference picture to obtain first unique point set of described reference picture, and described reference picture is used for the target of current image to be matched is followed the tracks of;
The characteristic optimization module is used for rejecting the abnormity point that described first unique point set exceeds described image range to be matched, obtains the set of second unique point;
The images match module is used for described second unique point set is carried out images match with described image to be matched, finishes the target following of described image to be matched according to matching result.
In first kind of possible implementation, described characteristic extracting module be used for to adopt and to preset first unique point set that angle point feature that algorithm extracts reference picture obtains described reference picture, and the described algorithm that presets comprises that Harris Corner Detection Algorithm, FAST accelerate to cut apart the small nut value similar area Corner Detection Algorithm of detected characteristics Corner Detection Algorithm, KLT Corner Detection Algorithm or SUSAN.
In conjunction with second aspect and first kind of possible implementation, in second kind of possible implementation, also comprise:
The template determination module is used for obtaining the image sequence of camera acquisition, with this image sequence as image to be matched; And from first two field picture of described image sequence, determine the reference picture of this image to be matched.
In conjunction with second kind of possible implementation, in the third possible implementation, described characteristic optimization module comprises:
Acquiring unit is used for obtaining described each characteristic point coordinates of first unique point set;
Judging unit is used for gathering the scope that each characteristic point coordinates judges whether to exceed described image to be matched according to described unique point, if yes, determines that then this unique point is abnormity point;
Reject the unit, be used for all abnormity point elimination of described first unique point set are obtained the set of second unique point.
In conjunction with the third possible implementation, in the 4th kind of possible implementation, described images match module comprises:
Choose the unit, be used for choosing a subgraph to be matched with the identical size of described reference picture at described image to be matched;
Extraction unit be used for to adopt and describedly to preset the feature that algorithm extracts described subgraph to be matched and obtain the set of the 3rd unique point;
Computing unit is used for calculating the distance between described second unique point set and the set of described the 3rd unique point;
In conjunction with the 4th kind of possible implementation, in the 5th kind of possible implementation, also comprise:
The template renewal module, be used for the described reference picture of deletion and will with the described reference template subgraph to be matched that the match is successful as new reference picture.
Implement the embodiment of the invention, have following beneficial effect:
Obtain the unique point set of this reference picture by the angle point feature of extracting reference picture, reject the abnormity point that exceeds image range to be matched in the unique point set, the unique point set of rejecting abnormity point is carried out images match with image to be matched, to realize target following.Adopt the angle point feature to carry out the figure coupling and can reduce the data volume that participates in calculating when keeping the image graphics key character, the angle point feature has good identification and stability, improves the speed of coupling greatly.Set is optimized this method to the unique point in the reference picture simultaneously, rejects the angle point of super scope, has solved the deficiency that the super scope of unique point can't be mated in the prior art, has improved the reliability of coupling.
Description of drawings
In order to be illustrated more clearly in the embodiment of the invention or technical scheme of the prior art, to do to introduce simply to the accompanying drawing of required use in embodiment or the description of the Prior Art below, apparently, accompanying drawing in describing below only is some embodiments of the present invention, for those of ordinary skills, under the prerequisite of not paying creative work, can also obtain other accompanying drawing according to these accompanying drawings.
Fig. 1 is the schematic flow sheet of a kind of method for tracking target of the embodiment of the invention;
Fig. 2 is another schematic flow sheet of a kind of method for tracking target of the embodiment of the invention;
Fig. 3 is the structural representation of a kind of target tracker of the embodiment of the invention;
Fig. 4 is another structural representation of a kind of target tracker of the embodiment of the invention;
Fig. 5 is the structural representation of characteristic optimization module among Fig. 4;
Fig. 6 is the structural representation of images match module among Fig. 4;
Fig. 7 is the another structural representation of a kind of target tracker of the embodiment of the invention.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the invention, the technical scheme in the embodiment of the invention is clearly and completely described, obviously, described embodiment only is the present invention's part embodiment, rather than whole embodiment.Based on the embodiment among the present invention, those of ordinary skills belong to the scope of protection of the invention not making the every other embodiment that obtains under the creative work prerequisite.
Referring to Fig. 1, be the schematic flow sheet of a kind of method for tracking target of the embodiment of the invention, this method comprises:
Concrete, angle point does not have the explicit mathematical definition, generally believes that angle point is the maximum point that two dimensional image brightness changes curvature on violent point or the image border curvature.In the various features of image, angle point is a kind of stable, invariable rotary and the validity feature that can overcome gray inversion.The algorithm that target tracker extracts the angle point feature in the reference picture can be Harris Corner Detection Algorithm, FAST Corner Detection Algorithm, KLT Corner Detection Algorithm or SUSAN Corner Detection Algorithm, also can be other algorithms, the present invention is not restricted, the angle point feature that target tracker extracts reference picture obtains the set of first unique point, and this first feature point set is combined into a bianry image.Reference picture is for record needs the image of tracked target, adopts this reference picture and image to be matched to mate to realize tracking to target.
Exceed the abnormity point of described image range to be matched in step 102, described first unique point set of rejecting, obtain the set of second unique point.
Concrete, when the shape of the target in the image to be matched is imperfect, for example target is blocked or target outside image to be matched the time, there is the angle point that does not belong within the image range to be matched in first unique point set of then from reference picture, extracting, become abnormity point, if the calculating these abnormity point participation images match can cause mating result of calculation and can't converge to correct result, i.e. reference picture and images match to be matched failure.Generation for fear of this situation, target tracker obtains the coordinate of each angle point in the set of first unique point, judge according to the coordinate of angle point and the coordinate relation of image to be matched whether angle point exceeds image range to be matched, if yes, then obtaining this angle point is abnormity point, reject all abnormity point according to above-mentioned method, obtain the set of second unique point.
Concrete, in image to be matched, choose a subgraph to be matched with the identical size of reference picture, the angle point feature that adopts the angle point extraction algorithm identical with step 101 to extract this subgraph to be matched obtains the set of the 3rd unique point, coupling based on the angle point feature has dual mode, a kind of is to describe second unique point set of the reference picture that extracts and the 3rd unique point set of subgraph to be matched with predetermined mode, the coupling of image just is converted into the coupling of two vector sets.Another kind is that the unique point set of extracting is represented with bianry image, bianry image is carried out range conversion, with the similarity of two unique point set of distance measuring similarity.Can certainly adopt additive method, the present invention is not restricted.
Implement embodiments of the invention, obtain the unique point set of this reference picture by the angle point feature of extracting reference picture, reject the abnormity point that exceeds image range to be matched in the unique point set, the unique point set of rejecting abnormity point is carried out images match with image to be matched, to realize target following.Adopt the angle point feature to carry out the figure coupling and can reduce the data volume that participates in calculating when keeping the image graphics key character, the angle point feature has good identification and stability, improves the speed of coupling greatly.Set is optimized this method to the unique point in the reference picture simultaneously, rejects the angle point of super scope, has solved the deficiency that the super scope of unique point can't be mated in the prior art, has improved the reliability of coupling.
Referring to Fig. 2, be another embodiment synoptic diagram of a kind of method for tracking target of the embodiment of the invention, this method comprises: step 201, obtain the image sequence of camera acquisition, with this image sequence as image to be matched; And from first two field picture of described image sequence, determine the reference picture of this image to be matched.
Concrete, the form of the image sequence of camera acquisition can be BMP(Bitmap, bitmap, be called for short BMP), TIFF(Tagged Image File Format, Tagged Image File (TIF) Format, be called for short TIFF) or JPEG(Joint Photographic Experts Group, joint image expert group, be called for short JPEG) wait picture format, also but MPEG(Moving Pictures Experts Group/Motin Pictures Experts Group, dynamic image expert group, be called for short MPEG) or AVI(Audio Video Interleaved, the Audio Video Interleaved form is called for short AVI) etc. video format, target tracker with this image sequence as image to be matched, and definite reference picture in first two field picture from image sequence, recording in the reference picture needs tracked target, and the size of hypothetical reference image is M*N(M, and N is the number of pixel), the size of image to be matched is m*n(m, n is the number of pixel), M＜m is then arranged, N＜n.
Step 202, first unique point that adopts the angle point feature that presets in the algorithm extraction reference picture to obtain described reference picture are gathered, and described reference picture is used for the target of current image to be matched is followed the tracks of.
Concrete, the algorithm that target tracker extracts the angle point feature in the reference picture can be Harris Corner Detection Algorithm, FAST Corner Detection Algorithm, KLT Corner Detection Algorithm or SUSAN Corner Detection Algorithm, also can be other algorithms, the present invention is not restricted, the angle point feature that target tracker extracts reference picture obtains the set of first unique point, and this first feature point set is combined into a bianry image.Reference picture is for record needs the image of tracked target, adopts this reference picture and image to be matched to mate to realize tracking to target.
Step 203, reject the abnormity point that exceeds described image range to be matched in described first unique point set, gather to second unique point.
Concrete, target tracker obtains the coordinate of each angle point in the set of first unique point, obtain the coordinate of each pixel in the image to be matched simultaneously, judge according to the relation of both coordinates whether angle point exceeds the scope of image to be matched, if yes, determine that then this angle point is abnormity point, reject all abnormity point in the set of first unique point according to the method described above, obtain the set of second unique point.
Step 204, in described image to be matched, choose a subgraph to be matched with the identical size of described image.
Concrete, with the example in the step 201, the big or small M*N of hypothetical reference image, image to be matched is m*n, satisfy and to concern M＜m, N＜n, target tracker choose the subgraph to be matched that a size is M*N in image to be matched, the search strategy of subgraph to be matched generally adopts method or the genetic algorithm of traversal, and the present invention is not restricted.
Step 205, adopt and describedly to preset the angle point feature that algorithm extracts described subgraph to be matched and obtain the set of the 3rd unique point.
Concrete, the angle point feature that target tracker adopts the algorithm identical with step 202 to extract subgraph to be matched obtains the set of the 3rd unique point.
Distance between step 206, described second unique point set of calculating and the set of described the 3rd unique point.
Concrete, target tracker presets mapping algorithm to first unique point set employing to carry out obtaining bianry image A after the range conversion, set adopts the identical mapping algorithm that presets to carry out obtaining bianry image B after the range conversion to the 3rd unique point, and that this presets mapping algorithm can be 3-4DT.Distance between bianry image A and the bianry image B is calculated and can be adopted Euclidean distance or non-Euclidean distance, this sentences the Hausdorff distance and is example, the Hausdorff distance is a kind of two minimax (Max-Min) distances on the point set that are defined in, for example calculate above-mentioned bianry image A and the Hausdorff distance between the bianry image B, suppose A={a1, ap}, B={b1 ... bq}, then the Hausdorff distance definition between this two somes set be H (A, B)=max (h (A, B), h (B, A)) (1)
h(A,B)=max（a∈A）min（b∈B）‖a-b‖ (2)
h(B,A)=max（b∈B）min（a∈A）‖b-a‖ (3)
‖ ‖ be between point set A and B point set apart from normal form (as: L2 or Euclidean distance).
Here, formula (1) is called two-way Hausdorff distance, is the citation form of Hausdorff distance; H in the formula (2) (A, B) and h (B A) is called from A and gathers B set and gather the unidirectional Hausdorff distance of A set from B.Be h (A, B) in fact at first to each the some ai among the point set A to gathering sorting apart from ‖ ai-bj ‖ between the mid point bj apart from the nearest B of this ai, get maximal value in this distance then as h (A, value B).H (B A) in like manner can get, known by formula (1), two-way Hausdorff distance H (A, B) be one-way distance h (A, B) and h (it has measured maximum between two point sets degree that do not match for B, A) the greater among both.
Step 207, judge that whether described distance is less than preset threshold value.
Concrete, whether the distance that target tracker determining step 206 calculates is less than preset threshold value, and if yes, execution in step 206 is if not, then return execution in step 204.
Step 208, the match is successful to determine subgraph to be matched in described reference picture and the described image to be matched.
Concrete, target tracker determines to comprise in the described image to be matched the target of needs tracking, reference picture and images match to be matched are successful, and trace into the target in the subgraph to be matched.
Step 209, the deletion described reference picture and will with the described reference picture subgraph to be matched that the match is successful as new reference picture.
Concrete, as new reference picture, when next frame image to be matched is carried out target following, adopt new reference picture to carry out images match the subgraph to be matched that the match is successful in the image to be matched.By to the dynamically updating of reference picture, can adjust shape and the size of target in the reference picture, be that images match is more accurate.
Implement embodiments of the invention, obtain the unique point set of this reference picture by the angle point feature of extracting reference picture, reject the abnormity point that exceeds image range to be matched in the unique point set, the unique point set of rejecting abnormity point is carried out images match with image to be matched, to realize target following.Adopt the angle point feature to carry out the figure coupling and can reduce the data volume that participates in calculating when keeping the image graphics key character, the angle point feature has good identification and stability, improves the speed of coupling greatly.Set is optimized this method to the unique point in the reference picture simultaneously, rejects the angle point of super scope, has solved the deficiency that the super scope of unique point can't be mated in the prior art, has improved the reliability of coupling.
Referring to Fig. 3, be the structural representation of a kind of target tracker of the embodiment of the invention, hereinafter to be referred as device 1, this device 1 comprises:
Characteristic extracting module 11 is used for adopting the angle point feature that presets algorithm extraction reference picture to obtain first unique point set of described reference picture, and described reference picture is used for the target of current image to be matched is followed the tracks of.
Concrete, the algorithm that characteristic extracting module 11 is extracted the angle point feature in the reference picture can be Harris Corner Detection Algorithm, FAST Corner Detection Algorithm, KLT Corner Detection Algorithm or SUSAN Corner Detection Algorithm, also can be other algorithms, the present invention is not restricted, the angle point feature that target tracker extracts reference picture obtains the set of first unique point, and this first feature point set is combined into a bianry image.Reference picture is for record needs the image of tracked target, adopts this reference picture and image to be matched to mate to realize tracking to target.
Concrete, when the shape of the target in the image to be matched is imperfect, for example target is blocked or target outside image to be matched the time, there is the angle point that does not belong within the image range to be matched in first unique point set of then from reference picture, extracting, become abnormity point, if the calculating these abnormity point participation images match can cause mating result of calculation and can't converge to correct result, i.e. reference picture and images match to be matched failure.Generation for fear of this situation, characteristic optimization module 12 is obtained the coordinate of each angle point in the set of first unique point, judge according to the coordinate of angle point and the coordinate relation of image to be matched whether angle point exceeds image range to be matched, if yes, then obtaining this angle point is abnormity point, reject all abnormity point according to above-mentioned method, obtain the set of second unique point.
Images match module 13 is used for described second unique point set is carried out images match with described image to be matched, finishes the target following of described image to be matched according to matching result.
Concrete, images match module 13 is chosen a subgraph to be matched with the identical size of reference picture in image to be matched, adopt with step characteristic extracting module 11 in the identical angle point extraction algorithm angle point feature of extracting this subgraph to be matched obtain the 3rd unique point and gather, coupling based on the angle point feature has dual mode, a kind of is to describe second unique point set of the reference picture that extracts and the 3rd unique point set of subgraph to be matched with predetermined mode, the coupling of image just is converted into the coupling of two vector sets.Another kind is that the unique point set of extracting is represented with bianry image, bianry image is carried out range conversion, with the similarity of two unique point set of distance measuring similarity.Can certainly adopt additive method, the present invention is not restricted.
Implement embodiments of the invention, obtain the unique point set of this reference picture by the angle point feature of extracting reference picture, reject the abnormity point that exceeds image range to be matched in the unique point set, the unique point set of rejecting abnormity point is carried out images match with image to be matched, to realize target following.Adopt the angle point feature to carry out the figure coupling and can reduce the data volume that participates in calculating when keeping the image graphics key character, the angle point feature has good identification and stability, improves the speed of coupling greatly.Set is optimized this method to the unique point in the reference picture simultaneously, rejects the angle point of super scope, has solved the deficiency that the super scope of unique point can't be mated in the prior art, has improved the reliability of coupling.
Further, referring to Fig. 4-Fig. 6, be another structural representation of a kind of target tracker of the embodiment of the invention, this device 1 also comprises except comprising above-mentioned characteristic extracting module 11, characteristic optimization module 12 and images match module 13:
Concrete, the form that template determination module 14 obtains the image sequence of camera acquisition can be picture formats such as BMP, TIFF or JPEG, also video formats such as MPEG or AVI, template determination module 14 with this image sequence as image to be matched, and definite reference picture in first two field picture from image sequence, recording in the reference picture needs tracked target, the size of hypothetical reference image is M*N(M, N is the number of pixel), the size of image to be matched is m*n(m, n is the number of pixel), M＜m is then arranged, N＜n.
Concrete, template renewal module 15 as new reference picture, when next frame image to be matched is carried out target following, adopts new reference picture to carry out images match the subgraph to be matched that the match is successful in the image to be matched.By to the dynamically updating of reference picture, can adjust shape and the size of target in the reference picture, be that images match is more accurate.
Wherein, characteristic optimization module 12 comprises:
Acquiring unit 121 is used for obtaining described each characteristic point coordinates of first unique point set;
Judging unit 122 is used for gathering the scope that each characteristic point coordinates judges whether to exceed described image to be matched according to described unique point, if yes, determines that then this unique point is abnormity point;
Images match module 13 comprises:
Choose unit 131, be used for choosing a subgraph to be matched with the identical size of described reference picture at described image to be matched;
Implement embodiments of the invention, obtain the unique point set of this reference picture by the angle point feature of extracting reference picture, reject the abnormity point that exceeds image range to be matched in the unique point set, the unique point set of rejecting abnormity point is carried out images match with image to be matched, to realize target following.Adopt the angle point feature to carry out the figure coupling and can reduce the data volume that participates in calculating when keeping the image graphics key character, the angle point feature has good identification and stability, improves the speed of coupling greatly.Set is optimized this method to the unique point in the reference picture simultaneously, rejects the angle point of super scope, has solved the deficiency that the super scope of unique point can't be mated in the prior art, has improved the reliability of coupling.
Referring to Fig. 7, another structural representation for a kind of target tracker of the embodiment of the invention, this target tracker 1 comprises processor 61, storer 62, input media 63 and output unit 64, the quantity of the processor 61 in the target tracker 1 can be one or more, and Fig. 7 is example with a processor.In the some embodiments of the present invention, processor 61, storer 62, input media 63 and output unit 64 can be connected by bus or other modes, are connected to example with bus among Fig. 7.
Wherein, storage batch processing code in the storer 62, and processor 61 is used for calling storer 62 stored program code, operation below being used for carrying out:
First unique point set that angle point feature that algorithm extracts reference picture obtains described reference picture is preset in employing, and described reference picture is used for the target of current image to be matched is followed the tracks of;
Reject the abnormity point that exceeds described image range to be matched in described first unique point set, obtain the set of second unique point;
Described second unique point set is carried out images match with described image to be matched, finish the target following of described image to be matched according to matching result.
Further, in some embodiments of the invention, processor 61 adopts and comprises that the angle point feature that presets algorithm extraction reference picture of Harris Corner Detection Algorithm, FAST Corner Detection Algorithm, KLT Corner Detection Algorithm or SUSAN Corner Detection Algorithm obtains first unique point set of described reference picture.
Preferably, in some embodiments of the invention, processor 61 also is used for carrying out:
Obtain the image sequence of camera acquisition, with this image sequence as image to be matched;
And from first two field picture of described image sequence, determine the reference picture of this image to be matched.
Preferably, in some embodiments of the invention, processor 61 is carried out the abnormity point that exceeds described image range to be matched in described first unique point set of described rejecting, and the step that obtains the set of second unique point comprises:
Obtain each characteristic point coordinates in described first unique point set;
Judge whether to exceed the scope of described image to be matched according to each characteristic point coordinates in the set of described unique point, if yes, determine that then this unique point is abnormity point;
With all abnormity point elimination in described first unique point set, obtain the set of second unique point.
Preferably, in some embodiments of the invention, processor 61 is carried out described described second unique point is gathered with described image to be matched and is carried out images match, and the step of finishing the target following of described image to be matched according to matching result comprises:
In described image to be matched, choose a subgraph to be matched with the identical size of described reference picture;
Adopt the described algorithm that presets to put forward the feature of described subgraph to be matched and obtain the 3rd unique point set;
Calculate the distance between described second unique point set and the set of described the 3rd unique point;
Judge that whether described distance is less than preset threshold value, if yes, the match is successful then to determine subgraph to be matched in described reference picture and the described image to be matched, if deny, then re-executes and choose a step with the subgraph to be matched of the identical size of described reference picture in described image to be matched.
Preferably, in some embodiments of the invention, processor 61 also be used for to carry out the described reference picture of deletion and will with the described reference template subgraph to be matched that the match is successful as new reference picture.
Implement embodiments of the invention, obtain the unique point set of this reference picture by the angle point feature of extracting reference picture, reject the abnormity point that exceeds image range to be matched in the unique point set, the unique point set of rejecting abnormity point is carried out images match with image to be matched, to realize target following.Adopt the angle point feature to carry out the figure coupling and can reduce the data volume that participates in calculating when keeping the image graphics key character, the angle point feature has good identification and stability, improves the speed of coupling greatly.Set is optimized this method to the unique point in the reference picture simultaneously, rejects the angle point of super scope, has solved the deficiency that the super scope of unique point can't be mated in the prior art, has improved the reliability of coupling.
One of ordinary skill in the art will appreciate that all or part of flow process that realizes in above-described embodiment method, be to instruct relevant hardware to finish by computer program, described program can be stored in the computer read/write memory medium, this program can comprise the flow process as the embodiment of above-mentioned each side method when carrying out.Wherein, described storage medium can be magnetic disc, CD, read-only storage memory body (Read-Only Memory, ROM) or at random store memory body (Random Access Memory, RAM) etc.
Above disclosed only is a kind of preferred embodiment of the present invention, certainly can not limit the present invention's interest field with this, one of ordinary skill in the art will appreciate that all or part of flow process that realizes above-described embodiment, and according to the equivalent variations that claim of the present invention is done, still belong to the scope that invention is contained.
Claims (12)
1. a method for tracking target is characterized in that, comprising:
First unique point set that angle point feature that algorithm extracts reference picture obtains described reference picture is preset in employing, and described reference picture is used for the target of current image to be matched is followed the tracks of;
Reject the abnormity point that exceeds described image range to be matched in described first unique point set, obtain the set of second unique point;
Described second unique point set is carried out images match with described image to be matched, finish the target following of described image to be matched according to matching result.
2. the method for claim 1 is characterized in that, the described algorithm that presets comprises: Harris Corner Detection Algorithm, FAST accelerate to cut apart the small nut value similar area Corner Detection Algorithm of detected characteristics Corner Detection Algorithm, KLT Corner Detection Algorithm or SUSAN.
3. method as claimed in claim 1 or 2 is characterized in that, described employing is preset before the step of first feature point set that angle point feature that algorithm extracts reference picture obtains described reference picture, also comprises:
Obtain the image sequence of camera acquisition, with this image sequence as image to be matched;
And from first two field picture of described image sequence, determine the reference picture of this image to be matched.
4. method as claimed in claim 3 is characterized in that, exceeds the abnormity point of described image range to be matched in described first unique point set of described rejecting, and the step that obtains the set of second unique point comprises:
Obtain each characteristic point coordinates in described first unique point set;
Judge whether to exceed the scope of described image to be matched according to each characteristic point coordinates in the set of described unique point, if yes, determine that then this unique point is abnormity point;
With all abnormity point elimination in described first unique point set, obtain the set of second unique point.
5. method as claimed in claim 4 is characterized in that, described described second unique point is gathered with described image to be matched carried out images match, and the step of finishing the target following of described image to be matched according to matching result comprises:
In described image to be matched, choose a subgraph to be matched with the identical size of described reference picture;
Adopt the described algorithm that presets to put forward the feature of described subgraph to be matched and obtain the 3rd unique point set;
Calculate the distance between described second unique point set and the set of described the 3rd unique point;
Judge that whether described distance is less than preset threshold value, if yes, the match is successful then to determine subgraph to be matched in described reference picture and the described image to be matched, if deny, then re-executes and choose a step with the subgraph to be matched of the identical size of described reference picture in described image to be matched.
6. method as claimed in claim 5 is characterized in that, described the set of described second unique point is carried out images match with described image to be matched, finishes according to matching result after the step of target following of described image to be matched, also comprises:
Delete described reference picture and will with the described reference template subgraph to be matched that the match is successful as new reference picture.
7. a target tracker is characterized in that, comprising:
Characteristic extracting module is used for adopting the angle point feature that presets algorithm extraction reference picture to obtain first unique point set of described reference picture, and described reference picture is used for the target of current image to be matched is followed the tracks of;
The characteristic optimization module is used for rejecting the abnormity point that described first unique point set exceeds described image range to be matched, obtains the set of second unique point;
The images match module is used for described second unique point set is carried out images match with described image to be matched, finishes the target following of described image to be matched according to matching result.
8. device as claimed in claim 7, it is characterized in that, described characteristic extracting module be used for to adopt and to preset first unique point set that angle point feature that algorithm extracts reference picture obtains described reference picture, and the described algorithm that presets comprises that Harris Corner Detection Algorithm, FAST accelerate to cut apart the small nut value similar area Corner Detection Algorithm of detected characteristics Corner Detection Algorithm, KLT Corner Detection Algorithm or SUSAN.
9. as claim 7 or 8 described devices, it is characterized in that, also comprise:
The template determination module is used for obtaining the image sequence of camera acquisition, with this image sequence as image to be matched; And from first two field picture of described image sequence, determine the reference picture of this image to be matched.
10. device as claimed in claim 9 is characterized in that, described characteristic optimization module comprises:
Acquiring unit is used for obtaining described each characteristic point coordinates of first unique point set;
Judging unit is used for gathering the scope that each characteristic point coordinates judges whether to exceed described image to be matched according to described unique point, if yes, determines that then this unique point is abnormity point;
Reject the unit, be used for all abnormity point elimination of described first unique point set are obtained the set of second unique point.
11. device as claimed in claim 10 is characterized in that, described images match module comprises:
Choose the unit, be used for choosing a subgraph to be matched with the identical size of described reference picture at described image to be matched;
Extraction unit be used for to adopt and describedly to preset the feature that algorithm extracts described subgraph to be matched and obtain the set of the 3rd unique point;
Computing unit is used for calculating the distance between described second unique point set and the set of described the 3rd unique point;
Matching unit, be used for judging that whether described distance is less than preset threshold value, if yes, then the match is successful for the subgraph to be matched in definite described reference picture and the described image to be matched, if not, then do not re-execute and in described image to be matched, choose a step with the subgraph to be matched of the identical size of described reference picture.
12. device as claimed in claim 11 is characterized in that, also comprises:
The template renewal module, be used for the described reference picture of deletion and will with the described reference template subgraph to be matched that the match is successful as new reference picture.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310183174.1A CN103279952B (en) | 2013-05-17 | 2013-05-17 | A kind of method for tracking target and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310183174.1A CN103279952B (en) | 2013-05-17 | 2013-05-17 | A kind of method for tracking target and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103279952A true CN103279952A (en) | 2013-09-04 |
| CN103279952B CN103279952B (en) | 2017-10-17 |
Family
ID=49062459
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310183174.1A Active CN103279952B (en) | 2013-05-17 | 2013-05-17 | A kind of method for tracking target and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103279952B (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104715470A (en) * | 2013-12-13 | 2015-06-17 | 南京理工大学 | Klt corner point detecting device and method |
| CN105469427A (en) * | 2015-11-26 | 2016-04-06 | 河海大学 | Target tracking method applied to videos |
| CN106250863A (en) * | 2016-08-09 | 2016-12-21 | 北京旷视科技有限公司 | object tracking method and device |
| CN106960179A (en) * | 2017-02-24 | 2017-07-18 | 北京交通大学 | Rail line Environmental security intelligent monitoring method and device |
| WO2017167159A1 (en) * | 2016-03-29 | 2017-10-05 | 中兴通讯股份有限公司 | Image positioning method and device |
| CN107273801A (en) * | 2017-05-15 | 2017-10-20 | 南京邮电大学 | A kind of method of video multi-target tracing detection abnormity point |
| CN107564033A (en) * | 2017-07-26 | 2018-01-09 | 北京臻迪科技股份有限公司 | A kind of tracking of submarine target, underwater installation and wearable device |
| CN108021921A (en) * | 2017-11-23 | 2018-05-11 | 塔普翊海（上海）智能科技有限公司 | Image characteristic point extraction system and its application |
| CN109685830A (en) * | 2018-12-20 | 2019-04-26 | 浙江大华技术股份有限公司 | Method for tracking target, device and equipment and computer storage medium |
| CN109919971A (en) * | 2017-12-13 | 2019-06-21 | 北京金山云网络技术有限公司 | Image processing method, device, electronic equipment and computer readable storage medium |
| CN110148178A (en) * | 2018-06-19 | 2019-08-20 | 腾讯科技（深圳）有限公司 | Camera localization method, device, terminal and storage medium |
| CN113409373A (en) * | 2021-06-25 | 2021-09-17 | 浙江商汤科技开发有限公司 | Image processing method, related terminal, device and storage medium |
| CN114926508A (en) * | 2022-07-21 | 2022-08-19 | 深圳市海清视讯科技有限公司 | Method, device, equipment and storage medium for determining visual field boundary |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109567600B (en) * | 2018-12-05 | 2020-12-01 | 江西书源科技有限公司 | Automatic accessory identification method for household water purifier |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101226592A (en) * | 2008-02-21 | 2008-07-23 | 上海交通大学 | Method for tracing object based on component |
| CN101399969A (en) * | 2007-09-28 | 2009-04-01 | 三星电子株式会社 | System, device and method for moving target detection and tracking based on moving camera |
| JP2010002976A (en) * | 2008-06-18 | 2010-01-07 | Secom Co Ltd | Image monitoring device |
| CN101840507A (en) * | 2010-04-09 | 2010-09-22 | 江苏东大金智建筑智能化系统工程有限公司 | Target tracking method based on character feature invariant and graph theory clustering |
| CN102750691A (en) * | 2012-05-29 | 2012-10-24 | 重庆大学 | Corner pair-based image registration method for Cauchy-Schwarz (CS) divergence matching |
-
2013
- 2013-05-17 CN CN201310183174.1A patent/CN103279952B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101399969A (en) * | 2007-09-28 | 2009-04-01 | 三星电子株式会社 | System, device and method for moving target detection and tracking based on moving camera |
| CN101226592A (en) * | 2008-02-21 | 2008-07-23 | 上海交通大学 | Method for tracing object based on component |
| JP2010002976A (en) * | 2008-06-18 | 2010-01-07 | Secom Co Ltd | Image monitoring device |
| CN101840507A (en) * | 2010-04-09 | 2010-09-22 | 江苏东大金智建筑智能化系统工程有限公司 | Target tracking method based on character feature invariant and graph theory clustering |
| CN102750691A (en) * | 2012-05-29 | 2012-10-24 | 重庆大学 | Corner pair-based image registration method for Cauchy-Schwarz (CS) divergence matching |
Non-Patent Citations (3)
| Title |
|---|
| 冯增光，张炯，宁纪锋，颜永丰: "基于角点检测的实时目标跟踪方法", 《计算机工程与设计》 * |
| 汪颖进: "目标跟踪过程中的遮挡问题研究", 《中国优秀博硕士学位论文全文数据库 (硕士) 信息科技辑》 * |
| 罗刚，张云峰: "应用角点匹配实现目标跟踪", 《中国光学与应用光学》 * |
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104715470B (en) * | 2013-12-13 | 2017-09-22 | 南京理工大学 | A kind of klt Corner Detections device and method |
| CN104715470A (en) * | 2013-12-13 | 2015-06-17 | 南京理工大学 | Klt corner point detecting device and method |
| CN105469427A (en) * | 2015-11-26 | 2016-04-06 | 河海大学 | Target tracking method applied to videos |
| CN105469427B (en) * | 2015-11-26 | 2018-06-19 | 河海大学 | One kind is for method for tracking target in video |
| WO2017167159A1 (en) * | 2016-03-29 | 2017-10-05 | 中兴通讯股份有限公司 | Image positioning method and device |
| CN106250863B (en) * | 2016-08-09 | 2019-07-26 | 北京旷视科技有限公司 | Object tracking method and device |
| CN106250863A (en) * | 2016-08-09 | 2016-12-21 | 北京旷视科技有限公司 | object tracking method and device |
| CN106960179A (en) * | 2017-02-24 | 2017-07-18 | 北京交通大学 | Rail line Environmental security intelligent monitoring method and device |
| CN107273801A (en) * | 2017-05-15 | 2017-10-20 | 南京邮电大学 | A kind of method of video multi-target tracing detection abnormity point |
| CN107273801B (en) * | 2017-05-15 | 2021-11-30 | 南京邮电大学 | Method for detecting abnormal points by video multi-target tracking |
| CN107564033A (en) * | 2017-07-26 | 2018-01-09 | 北京臻迪科技股份有限公司 | A kind of tracking of submarine target, underwater installation and wearable device |
| CN108021921A (en) * | 2017-11-23 | 2018-05-11 | 塔普翊海（上海）智能科技有限公司 | Image characteristic point extraction system and its application |
| CN109919971B (en) * | 2017-12-13 | 2021-07-20 | 北京金山云网络技术有限公司 | Image processing method, image processing device, electronic equipment and computer readable storage medium |
| CN109919971A (en) * | 2017-12-13 | 2019-06-21 | 北京金山云网络技术有限公司 | Image processing method, device, electronic equipment and computer readable storage medium |
| CN110148178A (en) * | 2018-06-19 | 2019-08-20 | 腾讯科技（深圳）有限公司 | Camera localization method, device, terminal and storage medium |
| US11210810B2 (en) | 2018-06-19 | 2021-12-28 | Tencent Technology (Shenzhen) Company Limited | Camera localization method and apparatus, terminal, and storage medium |
| CN110148178B (en) * | 2018-06-19 | 2022-02-22 | 腾讯科技（深圳）有限公司 | Camera positioning method, device, terminal and storage medium |
| CN109685830B (en) * | 2018-12-20 | 2021-06-15 | 浙江大华技术股份有限公司 | Target tracking method, device and equipment and computer storage medium |
| CN109685830A (en) * | 2018-12-20 | 2019-04-26 | 浙江大华技术股份有限公司 | Method for tracking target, device and equipment and computer storage medium |
| CN113409373A (en) * | 2021-06-25 | 2021-09-17 | 浙江商汤科技开发有限公司 | Image processing method, related terminal, device and storage medium |
| CN114926508A (en) * | 2022-07-21 | 2022-08-19 | 深圳市海清视讯科技有限公司 | Method, device, equipment and storage medium for determining visual field boundary |
| CN114926508B (en) * | 2022-07-21 | 2022-11-25 | 深圳市海清视讯科技有限公司 | Visual field boundary determining method, device, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103279952B (en) | 2017-10-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103279952A (en) | Target tracking method and device | |
| US11657514B2 (en) | Image processing apparatus, image processing method, and storage medium | |
| EP3168810A1 (en) | Image generating method and apparatus | |
| KR101524548B1 (en) | Apparatus and method for alignment of images | |
| CN110363179B (en) | Map acquisition method, map acquisition device, electronic equipment and storage medium | |
| JP7159384B2 (en) | Image processing device, image processing method, and program | |
| CN114782499A (en) | Image static area extraction method and device based on optical flow and view geometric constraint | |
| CN110969164A (en) | Low-illumination imaging license plate recognition method and device based on deep learning end-to-end | |
| CN108229281B (en) | Neural network generation method, face detection device and electronic equipment | |
| CN112614167A (en) | Rock slice image alignment method combining single-polarization and orthogonal-polarization images | |
| CN111161348B (en) | Object pose estimation method, device and equipment based on monocular camera | |
| CN108447084B (en) | Stereo matching compensation method based on ORB characteristics | |
| CN106778822B (en) | Image straight line detection method based on funnel transformation | |
| US9392146B2 (en) | Apparatus and method for extracting object | |
| CN113095385A (en) | Multimode image matching method based on global and local feature description | |
| CN113228105A (en) | Image processing method and device and electronic equipment | |
| CN113763468B (en) | Positioning method, device, system and storage medium | |
| WO2022174603A1 (en) | Pose prediction method, pose prediction apparatus, and robot | |
| CN111144489B (en) | Matching pair filtering method and device, electronic equipment and storage medium | |
| CN114025089A (en) | Video image acquisition jitter processing method and system | |
| Clark et al. | Perspective correction for improved visual registration using natural features. | |
| Deshmukh et al. | Embedded vision system for atmospheric turbulence mitigation | |
| Zhang et al. | A stereo matching algorithm based on multiresolution and epipolar constraint | |
| Luo et al. | Uav Large Oblique Image Geo-Localization Using Satellite Images in the Dense Buildings Area | |
| Ikehata et al. | Confidence-based refinement of corrupted depth maps |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into 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: 20191218 Address after: No.1, floor 3, No.319, zhanggongshan Road, Yuhui District, Bengbu City, Anhui Province Patentee after: Bengbu guijiu Intellectual Property Service Co., Ltd Address before: 518129 Bantian HUAWEI headquarters office building, Longgang District, Guangdong, Shenzhen Patentee before: Huawei Technologies Co., Ltd. |
|
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20201021 Address after: C 013, C 015, C 016, C 020, C 021, C 022, 3 / F, e-commerce Industrial Park, Nantong home textile city, Jinchuan Avenue, Chuanjiang Town, Tongzhou District, Nantong City, Jiangsu Province 226000 Patentee after: Ruide Yinfang (Nantong) Information Technology Co., Ltd Address before: No.1, floor 3, No.319, zhanggongshan Road, Yuhui District, Bengbu City, Anhui Province Patentee before: Bengbu guijiu Intellectual Property Service Co.,Ltd. |
|
| TR01 | Transfer of patent right |