利用光達技術可詳實的紀錄地表及地物之資訊，但點雲資料所含的空間資訊是隱含在點雲分布的情形內，若要從點雲資料中獲得直接描述地物的資訊，則必須經過特徵萃取工作，將點雲資料隱性資訊轉換成顯性的資訊。因此本研究之目的在於藉由分析點雲資料在空間之分佈性質，將光達點雲特徵作一歸納整理及給予明確之定義，並發展可行之特徵萃取演算理論，期望能達到某種程度之自動化。

本文依照萃取方式將點雲特徵分為擬合特徵、交會特徵、及邊界特徵三大類，並依照此分類發展可行之演算理論。擬合特徵萃取理論是利用最小二乘平差理論將特徵模型與點雲資料擬合，使得所有點雲資料到特徵模型的距離平方和為最小，並給定一搜尋範圍，藉由逐漸縮小搜尋範圍，剔除偏離特徵之點資料。交會特徵萃取理論是藉由空間之幾何交會公式，計算擬合特徵交會時所產生之交會特徵，並給定約制條件判斷交會特徵之合理性。邊界特徵之萃取理論是利用擬合特徵所給予之資訊，找出屬於邊界之點雲資料，再由這些邊界點雲資料計算出邊界特徵。

本研究所使用之點雲資料包括以Optech ALTM 3070 空載光達系統在台中朝馬交流道附近地區所得之空載掃瞄資料，以及Optech ILRIS-3D雷射掃瞄儀所得之地面掃瞄資料。實驗首先針對不同類型之地物，進行擬合特徵驗證工作，其次測試所萃取的邊界線是否準確，並探討不同掃描狀況對於邊界特徵萃取成果是否造成邊界內縮或外擴之現象發生。

LIDAR technique can acquire densely distributed points on the surfaces of scanned objects, called point cloud. The spatial information of objects is implied in point cloud data. Extracting features from point cloud is the first step to transfer the implicit information to explicit information. Therefore, the major purpose of this research is to categorize and define features may be extracted from point cloud. Based on the definitions, algorithms are developed to extract meaningful features from LIDAR data.

Observable features in point cloud data are categorized into types of fitting features, intersection features, and boundary features. The definition of fitting feature is to fit a feature model with point cloud by means of Least-Squares estimation, so that to determine the parameters of the feature model. Features of intersection are determined by computing the intersection of fitting features extracted previously. Boundary features are the boundary lines of a group of points distributed on a fitting feature.

The test data applied in this research include a set of airborne laser scanning data covering Taizhong County and a set of ground laser scanning data. The experiments are divided into two parts. First, tests are conducted to verify the correctness of the definitions and methods of extracting features. Second, the accuracy of extracted boundary features are checked to estimated the effect of footprint size.

賴志恆，『雷射掃瞄點雲資料八分樹結構化之研究』，國立成功大學測量工程研究所碩士論文，2003。

賴志凱，『地面雷射掃瞄儀的精度分析與檢定』，國立成功大學測量工程研究所碩士論文，2004。

Alharthy, A. and J. Bethel, 2002, "Building Extraction and Reconstruction from LIDAR Data", Proceedings of ACSM-ASPRS.

Haala, N. and C. Brenner, 1999, "Extraction of Building and Trees in Urban Environments", ISPRS Journal of Photogrammetry & Remote Sensing 54: 130-137.

Hu, Y. and V. Tao, 2002, "Bald DEM Generation and Building Extraction Using Range and Reflectance LIDAR Data", Proceedings of ASPRS 2002 Conference(CD-ROM).

Hough, P.V.C.,1962, "Methods and Means for Recognising Complex Patterns.U.S.patent", No.3069654.

Lin, S. H. and J. J. Jaw, 2004, "Structuralization of LIDAR Point Cloud", Asian Conference on Remote Sensing, Chiang Mai, Thailand.

Maas, H. G.,1999, "Fast Determination of Parametric House Models from Dense Airborne Laser Scanner data", ISPRS Workshop on Mobile Mapping Technology.
Maas, H. G. and G. Vosselman, 1999, "Two Algorithms for Extracting Building Models from Raw Laser Altimetry Data", ISPRS Journal of Photogrammetry & Remote Sensing 54: 153-163.

Masaharu, H. and H. Hasegawa, 2000, " Three-Dimensional City Modeling from Laser Scanner Data by Extracting Building Polygons Using Region Segmentation Method", IAPRS, vol.XXXIII, Part B3,pp.556-562, Amsterdam.

Rottensteiner, F. and Ch. Briese, 2002,"A New Method for Building Extraction In Urban Areas from High-Resolution LIDAR Data",ISPRS,vol.XXXIII,pp.295-301,Graz,Austria.

Sampath, A. and J. Shan, 2004, "Urban Modeling Based on Segmentation and Regularization of Airborne LIDAR Point Clouds", ISPRS Congress, Istanbul.

Vosselman, G., 2000, "Slope-based filtering of laser altimetry data",Int. Arch. Of Photogrammetry and Remote Sensing, Vol. 33, Part B3/2, pp.935- 942.

Vosselman, G. and S. Dijkman, 2001, "3D Building Model Reconstruction from Point Clouds and Ground Plans", International Archives of Photogrammetry and Remote Sensing, Annapolis, Maryland.

Woo, H., E. Kang, et al., 2002, "A New Segmentation method for Point Cloud Data", International Journal of Machine & Manufacture 42: 167-178.
Wang, Z. and T. Schenk, 2000, "Building Extraction and Reconstruction from LIDAR Data", IAPRS, vol.XXXIII, Part B3,pp.958-964, Amsterdam.