以三維直線特徵進行地面光達點雲套合

Registration of Ground-based LiDAR Point Clouds by Using 3D Line Features

莊子毅(Tzu-Yi Chuang)

指導教授 ： 趙鍵哲

國立臺灣大學/工學院/土木工程學研究所/碩士(2006年)

https://doi.org/10.6342/NTU.2006.03178

摘要

本研究論述多測站地面光達點雲資料，藉由三維直線特徵萃取、三維直線特徵匹配以及以三維直線特徵為主的空間相似轉換進行套合工作。研究中介紹人工選取、自動直線特徵偵測及面擬合交會三維直線特徵三種萃取模式，工作中比較分析各模式適用場景並聯合三種萃取模式於地面光達點雲資料套合作業；三維直線特徵匹配乃建構於空間角度與距離的幾何約制條件及空間相似轉換的近似模式演算法來進行共軛特徵之篩選及確認；以三維直線特徵為基礎之空間相似轉換具有近似模式、嚴密模式及聯合模式，聯合模式可免於提供起始近似值之作業困擾並達到嚴密求解轉換參數之功效。本研究以地面光達多測站之點雲資料以前述方法進行套合試驗，證實以三維直線特徵進行光達點雲資料套合之可行性。

關鍵字

光達點雲 ； 套合 ； 三維直線特徵 ； 三維直線特徵匹配 ； 空間相似轉換 ； 幾何約制

並列摘要

In this paper, the author shows how to extract the 3D line features from ground-based LiDAR point clouds in three different ways and develops the algorithms in which 3D line feature matching for registering overlapping scenes can be established by using geometric constraints, like spatial angles and distances, in an automatic fashion. 3D spatial similarity transformations by employing 3D line features are proposed for both its rigorous and approximate models in this study. The rigorous model involves providing approximation of parameters when initializing the computation. The approximate model, on the other hand, requires no approximation for a less accurate solution, however. The two models are combined, called hybrid model, in an attempt to free the necessity of approximation and acquire a highly accurate solution. The experiments for registering overlapping point cloud data sets collected by a ground-based LiDAR system demonstrates the satisfactory data registration result based on the proposed scheme.

並列關鍵字

LiDAR Point Cloud ； Registration ； 3D Line Features ； 3D Line Matching ； Similarity Transformation ； Geometric Constraints

參考文獻

黃國彥，2005。結合三維網格及迭代式霍夫轉換進行光達點雲資料結構化處理，國立台灣大學土木研究所碩士論文。

Akca, D., 2003. Full Automatic Registration of Laser Scanner Point Clouds, Optical 3D Measurement Techniques VI, pp.330-337.

Besl, P.J., and N.D. McKay, 1992. A method for registration of 3D shape, IEEE Trans. Pattern Analysis and Machine Intelligence, Vol. 14, No. 2, pp.239-256.

Boyer, K.L., and, A.C. Kak, 1988. Structural Stereopsis for 3D Vision, IEEE Trans. Pattern Analysis and Machine Intelligence, Vol.10, no.2, pp.144-166.

Canny, J., 1986. A Computational Approach to Edge Detection, IEEE Trans. Pattern Analysis and Machine Intelligence, Vol. PAMI-8, No.6, 679-698.