Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
HIERARCHICAL GAUSSIAN MIXTURE MODEL-BASED FAST AND ROBUST ROBOT THREE-DIMENSIONAL RECONSTRUCTION METHOD
Document Type and Number:
WIPO Patent Application WO/2022/095302
Kind Code:
A1
Abstract:
Disclosed is a hierarchical Gaussian mixture model-based fast and robust robot three-dimensional reconstruction method. The method includes the following steps: a robot obtains point cloud data of a measurement object, a GPU performs accelerated generation of a hierarchical Gaussian mixture model and a test set, a registration network is constructed and updated, the registration network is globally optimized, a reconstructed Gaussian mixture model is updated, the above steps are repeated until the robot completes measurement at all measurement points, a three-dimensional point cloud model of the measurement object is reconstructed, and a reconstruction result is analyzed and evaluated. The present method accelerates generation of the hierarchical Gaussian mixture model from point cloud data by means of GPU parallel computation, and is also able to handle noise and measurement uncertainty, speed and efficiency of three-dimensional reconstruction are improved, joint registration errors are reduced by means of reconstructing the registration network, updating the registration network, and globally optimizing the registration network, and three-dimensional reconstruction precision is guaranteed. The present invention is highly automated, has a fast reconstruction speed, is robust, and is particularly suitable for dense point cloud three-dimensional reconstruction of a large measurement object in an industrial scenario.

Inventors:
WANG YAONAN (CN)
TANG YONGPENG (CN)
MAO JIANXU (CN)
ZHU QING (CN)
ZHANG HUI (CN)
ZHOU XIANEN (CN)
JIANG YIMING (CN)
WU ZIJIE (CN)
NIE JINGMOU (CN)
Application Number:
PCT/CN2021/075625
Publication Date:
May 12, 2022
Filing Date:
February 05, 2021
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
UNIV HUNAN (CN)
International Classes:
G06T17/00
Foreign References:
CN112308961A2021-02-02
CN108921935A2018-11-30
CN109949349A2019-06-28
CN109754020A2019-05-14
CN105787895A2016-07-20
US20190088004A12019-03-21
Other References:
LEE, SEONG-WHAN ; LI, STAN Z: " Advances in biometrics : international conference, ICB 2007, Seoul, Korea, August 27 - 29, 2007 ; proceedings", vol. 11219 Chap.43, 7 October 2018, SPRINGER , Berlin, Heidelberg , ISBN: 3540745491, article ECKART BENJAMIN; KIM KIHWAN; KAUTZ JAN: "HGMR: Hierarchical Gaussian Mixtures for Adaptive 3D Registration", pages: 730 - 746, XP047489045, 032548, DOI: 10.1007/978-3-030-01267-0_43
ECKART BEN; KIM KIHWAN; TROCCOLI ALEJANDRO; KELLY ALONZO; KAUTZ JAN: "Accelerated Generative Models for 3D Point Cloud Data", 2016 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), IEEE, 27 June 2016 (2016-06-27), pages 5497 - 5505, XP033021746, DOI: 10.1109/CVPR.2016.593
SUN GUANGLING, TANG XIANG-LONG: "A Semi-Supervised Learning Algorithm Based on a Hierarchical GMM", JOURNAL OF COMPUTER RESEARCH AND DEVELOPMENT, KEXUE CHUBANSHE, BEIJING, CN, vol. 41, no. 1, 31 January 2004 (2004-01-31), CN , pages 156 - 161, XP055927473, ISSN: 1000-1239
Attorney, Agent or Firm:
CHANGSHA HUHANG PATENT AGENCY (SPECIAL GENERAL PARTNERSHIP) (CN)
Download PDF: