CN111862173A - On-line fitting and wearing method based on point cloud registration - Google Patents
On-line fitting and wearing method based on point cloud registration Download PDFInfo
- Publication number
- CN111862173A CN111862173A CN202010631686.XA CN202010631686A CN111862173A CN 111862173 A CN111862173 A CN 111862173A CN 202010631686 A CN202010631686 A CN 202010631686A CN 111862173 A CN111862173 A CN 111862173A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- registration
- point
- fitting
- line fitting
- 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
- 238000000034 method Methods 0.000 title claims abstract description 22
- 238000000605 extraction Methods 0.000 claims abstract description 7
- 230000000694 effects Effects 0.000 claims abstract description 5
- 238000001914 filtration Methods 0.000 claims abstract description 4
- 238000012800 visualization Methods 0.000 claims abstract description 4
- 239000011159 matrix material Substances 0.000 claims description 30
- 238000013519 translation Methods 0.000 claims description 26
- 230000009466 transformation Effects 0.000 claims description 10
- 238000010606 normalization Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 3
- 238000005259 measurement Methods 0.000 claims description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0641—Shopping interfaces
- G06Q30/0643—Graphical representation of items or shoppers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformation in the plane of the image
- G06T3/60—Rotation of a whole image or part thereof
-
- G06T5/70—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/50—Extraction of image or video features by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
-
- 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
- G06T2207/10028—Range image; Depth image; 3D point clouds
Abstract
The invention belongs to the technical field of three-dimensional point cloud, and particularly relates to an on-line fitting and fitting method based on point cloud registration, which comprises the following steps: s1: scanning to obtain a human body point cloud and a clothes point cloud, and reading the point clouds; s2: filtering and denoising the human body point cloud and the clothes point cloud in the S1; s3: carrying out feature extraction on the filtered and denoised human body point cloud and the clothing point cloud in the S2, and selecting proper parameters according to the actual conditions of the point clouds; s4: registering the human body point cloud and the clothes point cloud; s5: the point cloud after registration achieves the visualization effect, can realize magnification, reduction and rotation, and has stereoscopic impression. The invention can solve the problems that the prior on-line fitting and putting-on has no stereoscopic impression, has errors in fitting and putting-on size and the like, and can effectively carry out on-line fitting and putting-on.
Description
Technical Field
The invention relates to the technical field of three-dimensional point cloud, in particular to an on-line fitting and fitting method based on point cloud registration.
Background
Point cloud registration is a mathematical calculation process for converting a large-capacity three-dimensional space data point set in two or more coordinate systems into a unified coordinate system, actually, a transformation relation between the two coordinate systems is to be found, and the point cloud registration is widely applied to the fields of reverse engineering, computer vision, cultural relic digitization and the like.
At present, buying clothes on the internet is an indispensable part of daily life of modern young people, but the buying of clothes on the internet exposes problems, such as the buying of clothes is not proper, the buying of clothes is found to be in a style contrary to the self, and the like. Although the corresponding fitting room function is also promoted on the internet at present, the fitting and fitting effect is not obvious enough, and therefore, the on-line fitting and fitting method based on point cloud registration is provided.
Disclosure of Invention
The invention aims to solve the defects in the prior art and provides an on-line fitting and putting-on method based on point cloud registration.
In order to achieve the purpose, the invention adopts the following technical scheme: an on-line fitting and putting-on method based on point cloud registration comprises the following steps:
s1: scanning to obtain a human body point cloud and a clothes point cloud, and reading the point clouds;
s2: filtering and denoising the human body point cloud and the clothes point cloud in the S1;
s3: carrying out feature extraction on the filtered and denoised human body point cloud and the clothing point cloud in the S2, and selecting proper parameters according to the actual conditions of the point clouds;
s4: registering the human body point cloud and the clothes point cloud;
s5: the point cloud after registration achieves the visualization effect, can realize magnification, reduction and rotation, and has stereoscopic impression.
Preferably, in S1, the human body point cloud and the clothing point cloud are obtained by measuring with a three-dimensional laser scanner.
Preferably, in S3, the feature extracts a normal vector, a curvature, and the like of the point cloud, and the geometric description is made by using a point feature histogram through a spatial difference between a point and a nearby point, where the information provided by the point feature histogram has rotation invariance.
Preferably, in S4, the registration includes a coarse registration and a fine registration, and both the coarse registration and the fine registration are implemented by an algorithm.
Preferably, the coarse registration obtains an initial rotational-translational parameter from the features.
Preferably, the coarse registration algorithm comprises the steps of:
c1: selecting n points from the source point cloud B, wherein the distance of the selected points must be smaller than a given minimum threshold value in order to ensure that the selected points have different point characteristic histograms;
c2: and searching points which satisfy similar conditions with the source point cloud B by the target point cloud A, and keeping a one-to-one corresponding relation.
C3: and calculating a rotation matrix and a translation matrix of the corresponding point, and judging according to a Huber function.
Preferably, in said C3, the Huber function is
Where m is a given threshold value, liAnd repeating C1, C2 and C3 for the distance difference of the ith group after the corresponding point is transformed until the result is optimal, namely the error function takes the minimum value to obtain a translation matrix and a rotation matrix.
Preferably, the fine registration is performed according to different personal dressing habits of different customers, the corresponding distance threshold is set to meet the fitness required by the different customers, the iteration is performed circularly under the condition that the distance threshold is not met until the condition is met, and finally the rotation and translation parameters of the coordinate transformation of the human body point cloud and the clothes point cloud are obtained.
Preferably, the fine registration algorithm comprises the steps of:
d1: a point set ai is taken from the target point cloud A, and a corresponding point bi is found from the source point cloud B, so that the target point cloud A is obtained;
d2: calculating a rotation matrix and a translation matrix to make the objective function take the minimum value;
d3: performing rotation translation transformation on the target point cloud A, and updating to obtain a new point cloud data set A';
d4: calculating the distance between all corresponding points in the updated point cloud A' and the source point cloud B, and performing normalization processing to obtain
And (4) giving a threshold value, if the average distance d is smaller than the given threshold value, repeating the steps, and if not, considering convergence.
Preferably, the point clouds a and B have a point set of a ═ a1, a2, a3... an }, and B ═ B1, B2, b3... bm }, and after the rotational translation transformation, the points in the point cloud A, B correspond one to one, and a is ai＝R·bi+ T, where R is the rotation matrix, T is the translation matrix, and the rotation matrix R and the translation matrix T are such that the objective function is 
Take the minimum value, at which time R, T is the optimum parameter.
Compared with the prior art, the invention has the beneficial effects that: the invention can solve the problems that the prior on-line fitting and putting-on has no stereoscopic impression, has errors in fitting and putting-on size and the like, and can effectively carry out on-line fitting and putting-on.
Drawings
FIG. 1 is a flow chart of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Examples
Referring to fig. 1, the present invention provides a technical solution: an on-line fitting and putting-on method based on point cloud registration comprises the following steps:
s1: scanning to obtain a human body point cloud and a clothes point cloud, and reading the point clouds;
s2: filtering and denoising the human body point cloud and the clothes point cloud in the S1;
s3: carrying out feature extraction on the filtered and denoised human body point cloud and the clothes point cloud in the S2, and selecting proper parameters according to the actual condition of the point cloud in the feature extraction, so that the extraction efficiency is improved;
S4: registering the human body point cloud and the clothes point cloud;
s5: the point cloud after registration achieves the visualization effect, can realize magnification, reduction and rotation, and has stereoscopic impression.
And in the S1, the human body point cloud and the clothes point cloud are obtained through the measurement of the three-dimensional laser scanner, and the method is rapid, high in precision and high in resolution.
In the S3, the normal vector, curvature, and the like of the point cloud are generally extracted by feature extraction, but the number of geometric features around the point is large and the similarity is high, and global feature information of the point cloud cannot be obtained, so that a point feature histogram is used to make geometric description through the spatial difference between the point and the adjacent point, and information provided by the point feature histogram has rotational invariance, and is very robust for the point cloud.
In S4, the registration includes coarse registration and fine registration, and both the coarse registration and the fine registration are implemented by an algorithm, so as to improve the registration accuracy.
And the coarse registration obtains an initial rotation translation parameter according to the characteristic.
The coarse registration algorithm comprises the steps of:
c1: selecting n points from the source point cloud B, wherein the distance of the selected points must be smaller than a given minimum threshold value in order to ensure that the selected points have different point characteristic histograms;
c2: and searching points which satisfy similar conditions with the source point cloud B by the target point cloud A, and keeping a one-to-one corresponding relation.
C3: and calculating a rotation matrix and a translation matrix of the corresponding point, and judging according to a Huber function.
In said C3, the Huber function is
Where m is a given threshold value, liAnd repeating C1, C2 and C3 for the distance difference of the ith group after the corresponding point is transformed until the result is optimal, namely the error function takes the minimum value to obtain a translation matrix and a rotation matrix.
The precise registration is performed according to different personal dressing habits of different customers, corresponding distance thresholds are set to meet the requirements of different customers on fitness, for example, the garments such as swimwear need to be close-fitting, some T-shirts are loosely worn and put together and look good, the registered distance thresholds are set according to the fitness of bodies and the garments, iteration is performed circularly under the condition that the distance thresholds are not met until conditions are met, and finally the rotation translation parameters of the coordinate transformation of the human body point cloud and the garment point cloud are obtained.
The fine registration algorithm comprises the following steps:
d1: a point set ai is taken from the target point cloud A, and a corresponding point bi is found from the source point cloud B, so that the target point cloud A is obtained;
d2: calculating a rotation matrix and a translation matrix to make the objective function take the minimum value;
d3: performing rotation translation transformation on the target point cloud A, and updating to obtain a new point cloud data set A';
d4: calculating the distance between all corresponding points in the updated point cloud A' and the source point cloud B, and performing normalization processing to obtain 
And (4) giving a threshold value, if the average distance d is smaller than the given threshold value, repeating the steps, and if not, considering convergence.
The point clouds a and B, point sets a ═ a1, a2,a3... an }, B ═ B1, B2, b3... bm }, after the rotational translation transformation, the points in the point cloud A, B correspond one to one, and ai＝R·bi+ T, where R is the rotation matrix, T is the translation matrix, and the rotation matrix R and the translation matrix T are such that the objective function is
Take the minimum value, at which time R, T is the optimum parameter.
Compared with the prior art, the invention has the beneficial effects that: the invention can solve the problems that the prior on-line fitting and putting-on has no stereoscopic impression, has errors in fitting and putting-on size and the like, and can effectively carry out on-line fitting and putting-on.
It should be noted that the device structure and the accompanying drawings of the present invention mainly describe the principle of the present invention, and in the technology of the design principle, the arrangement of the power mechanism, the power supply system, the control system, and the like of the device is not completely described, but the details of the power mechanism, the power supply system, and the control system can be clearly known by those skilled in the art on the premise that the above inventive principle is understood.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be considered to be within the technical scope of the present invention, and the technical solutions and the inventive concepts thereof according to the present invention should be equivalent or changed within the scope of the present invention.
Claims (10)
1. An on-line fitting and wearing method based on point cloud registration is characterized in that: the method comprises the following steps:
s1: scanning to obtain a human body point cloud and a clothes point cloud, and reading the point clouds;
s2: filtering and denoising the human body point cloud and the clothes point cloud in the S1;
s3: carrying out feature extraction on the filtered and denoised human body point cloud and the clothing point cloud in the S2, and selecting proper parameters according to the actual conditions of the point clouds;
s4: registering the human body point cloud and the clothes point cloud;
s5: the point cloud after registration achieves the visualization effect, can realize magnification, reduction and rotation, and has stereoscopic impression.
2. The on-line fitting and putting-on method based on point cloud registration as claimed in claim 1, wherein: and in the step S1, a human body point cloud and a clothes point cloud are obtained through measurement of the three-dimensional laser scanner.
3. The on-line fitting and putting-on method based on point cloud registration as claimed in claim 1, wherein: and in the step S3, extracting the normal vector and curvature of the point cloud by using the features, using a point feature histogram to make geometric description through the space difference between the points and the adjacent points, wherein the information provided by the point feature histogram has rotation invariance.
4. The on-line fitting and putting-on method based on point cloud registration as claimed in claim 1, wherein: in S4, the registration includes a coarse registration and a fine registration, and both the coarse registration and the fine registration are implemented by an algorithm.
5. The point cloud registration-based on-line fitting and fitting method according to claim 4, wherein: and the coarse registration obtains an initial rotation translation parameter according to the characteristic.
6. The point cloud registration-based on-line fitting and fitting method according to claim 4, wherein: the coarse registration algorithm comprises the steps of:
c1: selecting n points from the source point cloud B, wherein the distance of the selected points is less than a given minimum threshold value;
c2: and searching points which satisfy similar conditions with the source point cloud B by the target point cloud A, and keeping a one-to-one corresponding relation.
C3: and calculating a rotation matrix and a translation matrix of the corresponding point, and judging according to a Huber function.
7. According to the claimsSolving 6 the on-line fitting and fitting method based on point cloud registration is characterized in that: in said C3, the Huber function is
Where m is a given threshold value, liAnd repeating C1, C2 and C3 for the distance difference of the ith group after the corresponding point is transformed until the result is optimal, namely the error function takes the minimum value to obtain a translation matrix and a rotation matrix.
8. The point cloud registration-based on-line fitting and fitting method according to claim 4, wherein: and the precise registration is carried out according to different personal dressing habits of different customers, corresponding distance thresholds are set to meet the fitness required by different customers, and the iterative operation is carried out under the condition that the distance thresholds are not met until the conditions are met, so that the rotation and translation parameters of the coordinate transformation of the human body point cloud and the clothes point cloud are finally obtained.
9. The point cloud registration-based on-line fitting and fitting method according to claim 4, wherein: the fine registration algorithm comprises the following steps:
d1: a point set ai is taken from the target point cloud A, and a corresponding point bi is found from the source point cloud B, so that the target point cloud A is obtained;
d2: calculating a rotation matrix and a translation matrix to make the objective function take the minimum value;
d3: performing rotation translation transformation on the target point cloud A, and updating to obtain a new point cloud data set A';
d4: calculating the distance between all corresponding points in the updated point cloud A' and the source point cloud B, and performing normalization processing to obtain
And (4) giving a threshold value, if the average distance d is smaller than the given threshold value, repeating the steps, and if not, considering convergence.
10. The point cloud registration-based on-line fitting and fitting building method of claim 9The method is characterized in that: the point clouds a and B have point sets of a ═ a1, a2, a3... an }, and B ═ B1, B2, b3... bm }, and after rotational translation transformation, the points in the point cloud A, B are in one-to-one correspondence, and a is a point set of B, B is B1, B2, b3... bm }, and the points in the point cloud A, B are in one-to-one correspondencei＝R·bi+ T, where R is the rotation matrix, T is the translation matrix, and the rotation matrix R and the translation matrix T are such that the objective function is
Take the minimum value, at which time R, T is the optimum parameter.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010631686.XA CN111862173A (en) | 2020-07-03 | 2020-07-03 | On-line fitting and wearing method based on point cloud registration |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010631686.XA CN111862173A (en) | 2020-07-03 | 2020-07-03 | On-line fitting and wearing method based on point cloud registration |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111862173A true CN111862173A (en) | 2020-10-30 |
Family
ID=73153044
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010631686.XA Pending CN111862173A (en) | 2020-07-03 | 2020-07-03 | On-line fitting and wearing method based on point cloud registration |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111862173A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113455864A (en) * | 2021-07-27 | 2021-10-01 | 深圳市简如法工程咨询有限公司 | Automatic and rapid three-dimensional formwork supporting device and method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103533449A (en) * | 2012-12-20 | 2014-01-22 | Tcl集团股份有限公司 | Method and system for realizing three-dimensional fitting based on intelligent three-dimensional television |
| CN106558095A (en) * | 2015-09-30 | 2017-04-05 | 捷荣科技集团有限公司 | A kind of wear the clothes methods of exhibiting and system based on anthropometric dummy |
| CN107578323A (en) * | 2017-10-10 | 2018-01-12 | 中国科学院合肥物质科学研究院 | The three-dimensional online virtual fitting system of real human body |
| CN108564605A (en) * | 2018-04-09 | 2018-09-21 | 大连理工大学 | A kind of three-dimensional measurement spots cloud optimization method for registering |
| CN110211163A (en) * | 2019-05-29 | 2019-09-06 | 西安财经学院 | A kind of point cloud matching algorithm based on EPFH feature |
| CN111145232A (en) * | 2019-12-17 | 2020-05-12 | 东南大学 | Three-dimensional point cloud automatic registration method based on characteristic information change degree |
-
2020
- 2020-07-03 CN CN202010631686.XA patent/CN111862173A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103533449A (en) * | 2012-12-20 | 2014-01-22 | Tcl集团股份有限公司 | Method and system for realizing three-dimensional fitting based on intelligent three-dimensional television |
| CN106558095A (en) * | 2015-09-30 | 2017-04-05 | 捷荣科技集团有限公司 | A kind of wear the clothes methods of exhibiting and system based on anthropometric dummy |
| CN107578323A (en) * | 2017-10-10 | 2018-01-12 | 中国科学院合肥物质科学研究院 | The three-dimensional online virtual fitting system of real human body |
| CN108564605A (en) * | 2018-04-09 | 2018-09-21 | 大连理工大学 | A kind of three-dimensional measurement spots cloud optimization method for registering |
| CN110211163A (en) * | 2019-05-29 | 2019-09-06 | 西安财经学院 | A kind of point cloud matching algorithm based on EPFH feature |
| CN111145232A (en) * | 2019-12-17 | 2020-05-12 | 东南大学 | Three-dimensional point cloud automatic registration method based on characteristic information change degree |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113455864A (en) * | 2021-07-27 | 2021-10-01 | 深圳市简如法工程咨询有限公司 | Automatic and rapid three-dimensional formwork supporting device and method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105447441B (en) | Face authentication method and device | |
| CN111724433B (en) | Crop phenotype parameter extraction method and system based on multi-view vision | |
| CN108764089B (en) | Human back spine midline recognition method based on depth camera | |
| CN110111375B (en) | Image matching gross error elimination method and device under Delaunay triangulation network constraint | |
| Olague et al. | Evolving head tracking routines with brain programming | |
| CN106023187A (en) | Image registration method based on SIFT feature and angle relative distance | |
| CN110009745B (en) | Method for extracting plane from point cloud according to plane element and model drive | |
| Han et al. | An improved corner detection algorithm based on harris | |
| CN111858997B (en) | Cross-domain matching-based clothing template generation method | |
| Gao et al. | Classification of 3D terracotta warrior fragments based on deep learning and template guidance | |
| CN111862315A (en) | Human body multi-size measuring method and system based on depth camera | |
| Liu et al. | Humangaussian: Text-driven 3d human generation with gaussian splatting | |
| CN110909778B (en) | Image semantic feature matching method based on geometric consistency | |
| CN111862173A (en) | On-line fitting and wearing method based on point cloud registration | |
| Wang et al. | Automatic recognition and 3D modeling of the neck-shoulder human shape based on 2D images | |
| Wang et al. | Tc-sfm: Robust track-community-based structure-from-motion | |
| CN110378291B (en) | Human body characteristic parameter prediction method based on semi-supervised learning | |
| Wang et al. | Research on fingerprint image based on improved log-gabor filter enhancement algorithm | |
| CN110647925A (en) | Rigid object identification method and device based on improved LINE-MOD template matching | |
| CN114399547B (en) | Monocular SLAM robust initialization method based on multiframe | |
| CN114187178B (en) | Porcelain fragment classification and splicing system and method for auxiliary cultural relic restoration | |
| CN113225300B (en) | Big data analysis method based on image | |
| Du et al. | Research on filtering and measurement algorithms based on human point cloud data | |
| CN114120140A (en) | Method for automatically extracting building height based on satellite image | |
| CN112036251A (en) | Method, system, medium and terminal for detecting user's dressing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |