CN102708589A - Three-dimensional target multi-viewpoint view modeling method on basis of feature clustering - Google Patents
Three-dimensional target multi-viewpoint view modeling method on basis of feature clustering Download PDFInfo
- Publication number
- CN102708589A CN102708589A CN2012101503808A CN201210150380A CN102708589A CN 102708589 A CN102708589 A CN 102708589A CN 2012101503808 A CN2012101503808 A CN 2012101503808A CN 201210150380 A CN201210150380 A CN 201210150380A CN 102708589 A CN102708589 A CN 102708589A
- Authority
- CN
- China
- Prior art keywords
- target
- eta
- viewpoint
- cluster
- modeling method
- 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 invention relates to a three-dimensional target multi-viewpoint view modeling method on the basis of the feature clustering, which comprises four steps of: 1, acquiring an all attitude image of a test target; 2, extracting a target feature vector set <xi|i=1,2,...N> from an acquired all attitude image set and adopting a 7-dimensional invariant moment formed by integrating an affine invariant moment with a geometric invariant moment and further carrying out regularization to describe an appearance feature of the all attitude image; 3, determining an optimal cluster number Cop; and 4, clustering target feature vectors into the Cop type by adopting a k-means clustering method and using a small amount of finally obtained clustering centers as modeling results. Aiming at the defect that under different viewpoints, due to difference of target images, the single target view description cannot identify the target, the invention establishes the three-dimensional target multi-viewpoint view modeling method which has a small description number and little redundant information and can be used for well describing the target all-attitude feature vector set. The three-dimensional target multi-viewpoint view modeling method has high practical value and wide application prospect in the field of the pattern recognition.
Description
Technical field
The present invention relates to a kind of many viewpoints of objective view modeling method, belong to area of pattern recognition, be specifically related to aspects such as Target Recognition, Target Modeling and data yojan based on feature clustering.Be used for the modeling of many viewpoints of objective, be applicable to the problem that target image difference causes under the different points of view simple target view description can not recognition objective.
Background technology
Objective identification is important research direction in the computer vision field.It is often very difficult in practical application, to obtain Three-dimension Target information at present, and the image that the identification objective still mainly forms through recognition objective two-dimensional projection is accomplished.Target two-dimensional imaging (projection) process has caused partial information to be lost, and under different points of view, the appearance difference of complex target is obvious, makes that the objective identification difficulty of fast and stable is very big.
Through objective being carried out modeling, be one of means that address this problem to form comprehensive description to target shape.This just needs many viewpoints view modeling method of research objective.Idea is that the description to target image under a plurality of viewpoints is integrated the description as target intuitively.Yet for same target, its attitude number can't be exhaustive, also comprised a lot of redundant informations between different the description.Therefore, need in the redundant description of yojan, keep important description as far as possible.
The viewpoint spherical surface partitioning algorithm of classical many viewpoints of objective view modeling method based target geometry and visible relation, the geometry topological characteristic of main based target image, the modeling result that finally obtains is the colourful attitude characteristic view collection about target.But these class methods depend on the geometry and the characteristic of target to the structure of target signature view sets, generally can only like solid of revolution, quadric surface body etc., be difficult in the reality use to a certain type of certain objects of certain complexity.
Summary of the invention
The objective of the invention is to: a kind of many viewpoints of objective view modeling method based on feature clustering is provided; The simple target view description that it causes to target image difference under the different points of view can not recognition objective shortcoming, made up a kind of many viewpoints of objective view modeling method of describing that quantity is little, redundant information is few, can better describing the full posture feature vector set of target.
Its technical scheme is following:
A kind of many viewpoints of objective view modeling method of the present invention based on feature clustering, it may further comprise the steps:
Step 1: the full attitude image that obtains test target.Target is placed viewpoint ball center, and the targeted attitude variation is equivalent to the difference object observing of video camera on viewpoint spherical surface.Uniform sampling on viewpoint spherical surface, obtaining quantity is the full attitude image set of the objective of N.
Step 2: from the full attitude image set that obtains, extract target feature vector set { x
i| i==1,2 ... N}.Employing combines the 7 dimension invariant moments that form after affine invariant moments and geometric invariant moment and the further regularization and describes the full attitude image of target resemblance.
Step 3: confirm optimum cluster numbers C
OpThrough analyzing cluster total error quadratic sum J
eAnd the relation curve (J between the cluster numbers C
e-c curve) and J
e-c second derivative curve zero crossing is confirmed optimum cluster numbers C
Op
Step 4: adopt k-mean cluster method that target feature vector is gathered and be C
OpType, with a small amount of cluster centre that finally obtains as modeling result.
Wherein, the described full attitude image that obtains test target of step 1, concrete implementation procedure is following:
At first establishment need be carried out the target three-dimensional model of many viewpoints view modeling.Then object module is placed the centre of sphere of imaginary viewpoint ball, be equivalent to the difference object observing of video camera on viewpoint spherical surface, therefore obtain video camera and observe the process that the process of viewpoint is the division viewpoint spherical surface because targeted attitude changes.Adopt the mode of similar division earth longitude and latitude, evenly divide viewpoint spherical surface with same intervals, and at the division points place of correspondence to target imaging, can obtain the full attitude image set of target.
Wherein, the described affine invariant moments of step 2 is obtained by (1) formula, wherein μ
PqThe p+q rank central moment of presentation video.
Being used for carrying out preceding four components of comprehensive geometric invariant moment with above-mentioned affine invariant moments is obtained by (2) formula
For solving the difference of the order of magnitude between above-mentioned invariant moments, need further above-mentioned invariant moments to be carried out the regularization computing
I in the formula
kThe combination invariant moments of representing not regularization,
Expression regularization combination invariant moments, Sign representes Heaviside function, S representes the summation of square factor exponent number in each square component.Finally obtain describing 7 dimension regularization invariant moments of target shape information.
Wherein, the described definite optimum cluster numbers C of step 3
Op, concrete implementation procedure is following:
Adopt k-mean cluster method that the full posture feature vector set of above-mentioned target is carried out cluster, making classification count c progressively increases, c=2, and c=3 ..., write down the total error quadratic sum J of each cluster
e, final according to obtaining J
e-c curve further calculates J
e-c second derivative curve under the situation of considering the certain numerical value fluctuation, is selected J
eOn-c second derivative the curve first approach zero point and relatively the mid point between flat zone as the cluster numbers c of optimum
OpEstimated value.
The advantage that the present invention has is: through feature clustering, represent the distribution of the full posture feature vector of target in feature space with the minority cluster centre, realized the data yojan, the method highly versatile can conveniently be confirmed optimum cluster numbers.
Description of drawings
Fig. 1 is a processing flow chart of the present invention;
The viewpoint ball synoptic diagram of Fig. 2 for relating among the present invention;
Fig. 3 is J among the present invention
e-c curve synoptic diagram;
Fig. 4 confirms optimum cluster numbers C for the present invention
OpSynoptic diagram.
Embodiment
The present invention is directed to many viewpoints of objective view modeling problem, proposed a kind of many viewpoints of objective view modeling method based on Support Vector data description.
See Fig. 1, a kind of many viewpoints of objective view modeling method of the present invention based on feature clustering, it may further comprise the steps:
Step 1: the full attitude image that obtains test target.Target is placed viewpoint ball center, and the targeted attitude variation is equivalent to the difference object observing of video camera on viewpoint spherical surface.Uniform sampling on viewpoint spherical surface, obtaining quantity is the full attitude image set of the objective of N.
Step 2: from the full attitude image set that obtains, extract target feature vector set { x
i| i=1,2 ... N}.Employing combines the 7 dimension invariant moments that form after affine invariant moments and geometric invariant moment and the further regularization and describes the full attitude image of target resemblance.
Step 3: confirm optimum cluster numbers C
OpThrough analyzing cluster total error quadratic sum J
eAnd the relation curve (J between the cluster numbers C
e-c curve) and J
e-c second derivative curve zero crossing is confirmed optimum cluster numbers C
Op
Step 4: adopt k-mean cluster method that target feature vector is gathered and be C
OpType, with a small amount of cluster centre that finally obtains as modeling result.
At first, obtain the full attitude image of target and need create the target three-dimensional model of treating modeling.See Fig. 2, object module is placed the centre of sphere of viewpoint ball.The viewpoint ball is the imaginary unit's ball that is defined in the three dimensional euclidean space; Can to regard target as be the process of starting point to the different directions projection with the centre of sphere to the imaging process of target under the different points of view; The corresponding projecting direction vector of every bit P P sees target along the rectangular projection of projecting direction P corresponding to P point on viewpoint spherical surface on the viewpoint spherical surface.Therefore obtain video camera and observe the process that the process of viewpoint is the division viewpoint spherical surface.Adopt the mode of similar division earth longitude and latitude, evenly divide viewpoint spherical surface with same intervals, and at the division points place of correspondence to target imaging, can obtain element number is the full attitude image set of target of N.
The attitude image set is extracted target signature, obtain to describe the proper vector set { x of the full attitude appearance information of this target
i| i=1,2 ... N}.The 7 dimension invariant moments that adopted comprehensive affine invariant moments and geometric invariant moment to form are described target shape.Affine invariant moments is obtained by (1) formula, wherein μ
PqThe p+q rank central moment of presentation video.
Being used for carrying out preceding four components of comprehensive geometric invariant moment with above-mentioned affine invariant moments is obtained by (2) formula
For solving the difference of the order of magnitude between above-mentioned invariant moments, need further above-mentioned invariant moments to be carried out the regularization computing
Wherein Ik representes the combination invariant moments of not regularization;
expression regularization combination invariant moments; Sign representes Heaviside function, and S representes the summation of square factor exponent number in each square component.For example for second component Ф of geometric invariant moment
2, S=(2+0) * 2=4, wherein (2+0) expression square factor η
20Exponent number with, * 2 the expression to square factor η
20Power operation.
7 dimension regularization invariant moments of target shape information have finally been obtained describing.
The full posture feature vector set of above-mentioned target is carried out cluster, with a small amount of cluster centre that finally obtains as modeling result.Adopted k-mean cluster method.The basis of k-average is the error sum of squares criterion.If N
iBe i cluster Γ
iIn number of samples, m
iIt is the average of these samples.Γ
iIn each sample y and average m
iBetween error sum of squares after to all types addition be:
J
eMeasured with c cluster centre m
1, m
2..., m
cThe total square-error that is produced when representing c sample subclass.To different cluster J
eValue be different, make J
eMinimum cluster is the optimal result under the error sum of squares criterion.
Adopt following method to confirm optimum cluster numbers c
Op: see shown in Figure 3ly, making classification count c progressively increases, to c=2, and c=3 ... use the k-mean algorithm respectively, error sum of squares will reduce along with the increase of c in theory.Then there is certain critical point OP, when cluster numbers is c
OpWhen individual, J
eC is increased to c from 1 along with cluster numbers
OpAnd reduce J when c increases again rapidly
eThough reduce to some extent, the speed that reduces is slowed down.See shown in Figure 4, according to calculating J
e-c curve further calculates J
e-c second derivative curve under the situation of considering the certain numerical value fluctuation, is selected J
eOn-c second derivative the curve first approach zero point and relatively the mid point between flat zone as the cluster numbers c of optimum
OpEstimated value.
Finally, adopting the k-means clustering algorithm, is c with the full posture feature vector set of target cluster
OpIndividual classification, with the cluster centre of each sub-category as many viewpoints of objective view modeling result.
Claims (4)
1. many viewpoints of objective view modeling method based on feature clustering, it is characterized in that: it may further comprise the steps:
Step 1: the full attitude image that obtains test target; Target is placed viewpoint ball center, and the targeted attitude variation is equivalent to the difference object observing of video camera on viewpoint spherical surface, uniform sampling on viewpoint spherical surface, and obtaining quantity is the full attitude image set of the objective of N;
Step 2: from the full attitude image set that obtains, extract target feature vector set { x
i| i=1,2 ... N}, employing combines the 7 dimension invariant moments that form after affine invariant moments and geometric invariant moment and the further regularization and describes the full attitude image of target resemblance;
Step 3: confirm optimum cluster numbers C
OpThrough analyzing cluster total error quadratic sum J
eAnd the relation curve between the cluster numbers C is J
e-c curve and J
e-c second derivative curve zero crossing is confirmed optimum cluster numbers C
Op
Step 4: adopt k-mean cluster method that target feature vector is gathered and be C
OpType, with a small amount of cluster centre that finally obtains as modeling result.
2. a kind of many viewpoints of objective view modeling method according to claim 1 based on feature clustering; It is characterized in that: the described full attitude image that obtains test target of step 1; Concrete implementation procedure is following: at first establishment need be carried out the target three-dimensional model of many viewpoints view modeling; Then object module is placed the centre of sphere of imaginary viewpoint ball; Be equivalent to the difference object observing of video camera on viewpoint spherical surface because targeted attitude changes, therefore obtain video camera and observe the process that the process of viewpoint is the division viewpoint spherical surface; Adopt the mode of similar division earth longitude and latitude, evenly divide viewpoint spherical surface with same intervals, and at the division points place of correspondence to target imaging, promptly obtain the full attitude image set of target.
3. a kind of many viewpoints of objective view modeling method based on feature clustering according to claim 1, it is characterized in that: the described affine invariant moments of step 2 is obtained by (1) formula, wherein μ
PqThe p+q rank central moment of presentation video;
Being used for carrying out preceding four components of comprehensive geometric invariant moment with above-mentioned affine invariant moments is obtained by (2) formula
For solving the difference of the order of magnitude between above-mentioned invariant moments, need further above-mentioned invariant moments to be carried out the regularization computing
In the formula, I
kThe combination invariant moments of representing not regularization,
Expression regularization combination invariant moments, Sign representes Heaviside function, S representes the summation of square factor exponent number in each square component, finally obtains describing 7 dimension regularization invariant moments of target shape information.
4. a kind of many viewpoints of objective view modeling method based on feature clustering according to claim 1 is characterized in that: the described definite optimum cluster numbers C of step 3
Op, concrete implementation procedure is following:
Adopt k-mean cluster method that the full posture feature vector set of above-mentioned target is carried out cluster, making classification count c progressively increases, c=2, and c=3 ..., write down the total error quadratic sum J of each cluster
e, final according to obtaining J
e-c curve further calculates J
e-c second derivative curve under the situation of considering the certain numerical value fluctuation, is selected J
eOn-c second derivative the curve first approach zero point and relatively the mid point between flat zone as the cluster numbers c of optimum
OpEstimated value.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210150380.8A CN102708589B (en) | 2012-05-15 | 2012-05-15 | Three-dimensional target multi-viewpoint view modeling method on basis of feature clustering |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210150380.8A CN102708589B (en) | 2012-05-15 | 2012-05-15 | Three-dimensional target multi-viewpoint view modeling method on basis of feature clustering |
Publications (2)
Publication Number | Publication Date |
---|---|
CN102708589A true CN102708589A (en) | 2012-10-03 |
CN102708589B CN102708589B (en) | 2014-10-29 |
Family
ID=46901305
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201210150380.8A Active CN102708589B (en) | 2012-05-15 | 2012-05-15 | Three-dimensional target multi-viewpoint view modeling method on basis of feature clustering |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102708589B (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103295025A (en) * | 2013-05-03 | 2013-09-11 | 南京大学 | Automatic selecting method of three-dimensional model optimal view |
CN104866872A (en) * | 2015-06-03 | 2015-08-26 | 哈尔滨工业大学 | Multi-feature parameter battery equivalent grouping method |
CN103870845B (en) * | 2014-04-08 | 2017-02-15 | 重庆理工大学 | Novel K value optimization method in point cloud clustering denoising process |
CN106407974A (en) * | 2015-07-28 | 2017-02-15 | 通用汽车环球科技运作有限责任公司 | Method for object localization and pose estimation for an object of interest |
CN107490356A (en) * | 2017-08-21 | 2017-12-19 | 上海航天控制技术研究所 | A kind of noncooperative target rotary shaft and rotation angle measuring method |
CN108647730A (en) * | 2018-05-14 | 2018-10-12 | 中国科学院计算技术研究所 | A kind of data partition method and system based on historical behavior co-occurrence |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100332210A1 (en) * | 2009-06-25 | 2010-12-30 | University Of Tennessee Research Foundation | Method and apparatus for predicting object properties and events using similarity-based information retrieval and modeling |
CN102024036A (en) * | 2010-11-29 | 2011-04-20 | 清华大学 | Three-dimensional object retrieval method and device based on hypergraphs |
-
2012
- 2012-05-15 CN CN201210150380.8A patent/CN102708589B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100332210A1 (en) * | 2009-06-25 | 2010-12-30 | University Of Tennessee Research Foundation | Method and apparatus for predicting object properties and events using similarity-based information retrieval and modeling |
CN102024036A (en) * | 2010-11-29 | 2011-04-20 | 清华大学 | Three-dimensional object retrieval method and device based on hypergraphs |
Non-Patent Citations (2)
Title |
---|
XUAN SONG等: "An Online Approach: Learning-Semantic-Scene-by-Tracking and Tracking-by-Learning-Semantic-Scene", 《2010 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION(CVPR)》 * |
翁文杰: "三维目标多尺度建模与混合神经网络识别方法", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103295025A (en) * | 2013-05-03 | 2013-09-11 | 南京大学 | Automatic selecting method of three-dimensional model optimal view |
CN103295025B (en) * | 2013-05-03 | 2016-06-15 | 南京大学 | A kind of automatic selecting method of three-dimensional model optimal view |
CN103870845B (en) * | 2014-04-08 | 2017-02-15 | 重庆理工大学 | Novel K value optimization method in point cloud clustering denoising process |
CN104866872A (en) * | 2015-06-03 | 2015-08-26 | 哈尔滨工业大学 | Multi-feature parameter battery equivalent grouping method |
CN104866872B (en) * | 2015-06-03 | 2018-06-05 | 哈尔滨工业大学 | The battery equivalent group technology of more characteristic parameters |
CN106407974A (en) * | 2015-07-28 | 2017-02-15 | 通用汽车环球科技运作有限责任公司 | Method for object localization and pose estimation for an object of interest |
CN106407974B (en) * | 2015-07-28 | 2019-10-15 | 通用汽车环球科技运作有限责任公司 | The method of target positioning and Attitude estimation for interesting target |
CN107490356A (en) * | 2017-08-21 | 2017-12-19 | 上海航天控制技术研究所 | A kind of noncooperative target rotary shaft and rotation angle measuring method |
CN107490356B (en) * | 2017-08-21 | 2020-04-07 | 上海航天控制技术研究所 | Non-cooperative target rotating shaft and rotation angle measuring method |
CN108647730A (en) * | 2018-05-14 | 2018-10-12 | 中国科学院计算技术研究所 | A kind of data partition method and system based on historical behavior co-occurrence |
CN108647730B (en) * | 2018-05-14 | 2020-11-24 | 中国科学院计算技术研究所 | Data partitioning method and system based on historical behavior co-occurrence |
Also Published As
Publication number | Publication date |
---|---|
CN102708589B (en) | 2014-10-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109753885B (en) | Target detection method and device and pedestrian detection method and system | |
Zhi et al. | LightNet: A Lightweight 3D Convolutional Neural Network for Real-Time 3D Object Recognition. | |
CN111160214B (en) | 3D target detection method based on data fusion | |
Zheng et al. | Non-local scan consolidation for 3D urban scenes | |
CN106716450B (en) | Image-based feature detection using edge vectors | |
Huang et al. | A coarse-to-fine algorithm for matching and registration in 3D cross-source point clouds | |
Gao et al. | View-based 3D object retrieval: challenges and approaches | |
CN102708589B (en) | Three-dimensional target multi-viewpoint view modeling method on basis of feature clustering | |
CN105608421B (en) | A kind of recognition methods of human action and device | |
Solanki et al. | Automatic Identification of Temples in Digital Images through Scale Invariant Feature Transform | |
CN102024036B (en) | Three-dimensional object retrieval method and device based on hypergraphs | |
CN102750704B (en) | Step-by-step video camera self-calibration method | |
CN101976461A (en) | Novel outdoor augmented reality label-free tracking registration algorithm | |
CN104063702A (en) | Three-dimensional gait recognition based on shielding recovery and partial similarity matching | |
CN109035327B (en) | Panoramic camera attitude estimation method based on deep learning | |
CN103902989B (en) | Human action video frequency identifying method based on Non-negative Matrix Factorization | |
CN103295025A (en) | Automatic selecting method of three-dimensional model optimal view | |
CN107944459A (en) | A kind of RGB D object identification methods | |
CN111754618B (en) | Object-oriented live-action three-dimensional model multi-level interpretation method and system | |
CN105488541A (en) | Natural feature point identification method based on machine learning in augmented reality system | |
Yun et al. | Supervoxel-based saliency detection for large-scale colored 3D point clouds | |
CN103839066A (en) | Feature extraction method from biological vision | |
Chiu et al. | See the difference: Direct pre-image reconstruction and pose estimation by differentiating hog | |
Kong et al. | Point cloud generation from multiple angles of voxel grids | |
CN109284692A (en) | Merge the face identification method of EM algorithm and probability two dimension CCA |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant |