透過您的圖書館登入
IP:18.117.183.150
  • 學位論文

發展多視角立體影像擷取設備以重建三維模型之研究

A Study in the Development of a Multi-View Stereo-Pair Image Acquisition Device for Reconstruction of 3D Models

指導教授 : 陳冠宇

摘要


在立體視覺三維重建領域中,使用二台攝影機擷取左右影像,直接模擬人類雙眼處理景物的方法,從二個視點觀察一個景物,由二台攝影機擺設相同水平的左右位置,本文採用多視角立體視覺架構,通過影像擷取、攝影機標定、立體校正、影像立體匹配、視差深度計算、物體表面重建三維模型。近年來隸屬稠密性匹配方法的半全域塊匹配演算法有著快速匹配和稠密性匹配點的優勢,在本文中運用及探討。本文將針對立體視覺重建三維模型研究的立體匹配方法和鏡頭的校正過程做基礎理論的探討,再以實驗及分析方式,針對實驗影像以半全域塊匹配演算法方式輸出視差圖,最後再由視差圖產生三維點雲,並由該方法所產生的點雲進行紋理貼圖達到重建三維模型的目的。

並列摘要


In the field of stereoscopic three-dimensional reconstruction, about using two cameras to capture images directly simulate the human eye treatment scene, we observed a scene from two viewpoints, the two cameras display the same level of the left and right positions, we use multi-angle stereoscopic architecture, through image capture, camera calibration, stereo calibration, image matching, parallax depth calculation, the surface reconstruction of 3D models. Recently, the semi-global block matching (SGBM) algorithm which has the advantage of speed and dense matching has discussed and be used widely. In this study, the correction process stereo matching method for stereo vision reconstructed model and lens do discuss the basic theory, and then experimentally and analytically, for experimental video to half full gamut mapping algorithms to output the disparity map, and finally by the disparity map 3D point cloud produced by the method of point cloud generated by texture mapping to achieve the purpose of reconstruction of three-dimensional model.

參考文獻


[5] 周晏德,運用獨立成份分析於移動目標追蹤之研究,碩士論文,中原大學機械工程學系,2011。
[23] 錢鉦津,改良式立體視覺演算法應用於機器人視覺系統之研究,博士論文,中原大學機械工程學系,2013。
[24] 陳彥良,即時立體視覺物體追蹤系統,碩士論文,中原大學機械工程學系,2003。
[25] 王荷凱,立體視覺應用於多自主行動機器人無線感測網路訂位技術之研究,,碩士論文,中原大學機械工程學系,2015。
[26] K. Y. Chen, C. C. Chien, W. L. Chang, and C. C. Hsieh, “Improving the accuracy of depth estimation using a modified stereo vision model in binocular vision,” The 10th International Symposium on Measurement Technology and Intelligent Instruments, Daejeon, Korea, 2011.

延伸閱讀