汇报题目:参加International Optics and Photonics Conference国际会议回国报告
汇报时间:2019年12月31日(星期二)14:30
汇报地点:交大兴庆校区 逸夫科学馆263会议室
汇报人:段琼琼
会议名称:International Optics and Photonics Conference-2019
会议时间:2019年12月9日至10日
会议地点:Heathrow, London, UK
会议简介:OLCOPC(International Optics and Photonics Conference-2019)该会议是一场关于光学与成像领域的科学盛会。它为科研人员、科学家、博士、教授、领域关键决策者、行业专家等相通物理行业领域人士提供了一种快捷高效的讨论合作平台,并分享前沿研究工作。解决研究过程中遇到的相关行业难题。为金沙威尼斯欢乐娱人城金沙威尼斯欢乐娱人城认定高水平国际会议。
参会论文信息:
Title: Research on Stereo Matching Method for Array Distribution Features Based on Plane Projection Transformation and Local Dynamic Plane Approximation
Author: Qiongqiong Duan, Zhao Wang, Junhui Huang, Jianmin Gao
Abstract:
Stereo matching is the key part in stereo vision research. At present, the typical matching methods are mainly based on gray correlation or feature of objects. The gray correlation matching method utilizes the pixel gray-scale statistical features of the target object to identify and match them. The feature matching method utilizes the similarity of the feature matching primitives in the images to identify and match them. However, the typical stereo matching methods are some limited by the scene with large distortion, insignificant and array distribution features. Taking the Tube Sheet Holes(TSH) matching as an example, this paper proposes a stereo matching method for array distribution feature based on plane projection transformation and local dynamic plane. Firstly, the center matching of the TSH in the left and right images is completed with the matching method based plane projective transformation and polar constraint rapidly. And then the edge matching and 3-D reconstruction of the TSH is completed with the method based on the local dynamic plane and the camera model precisely. Finally, the redundant dimensions of the TSH is selected by the redundant size evaluation method proposed in this paper. Experimental results show that the proposed method can effectively match the array features in complex scenes, and finally accurately calculate dimensions of the TSH. Moreover, compared with the typical matching methods, the method proposed in this paper has proved to be more effective and precisely.