全波形光達(Full Waveform LIDAR)是一種新型態的主動式遙測設備。與傳統光達儀器相較，全波形光達不僅可量測雷射光束回波的距離以計算目標物之座標，產生三維的點雲資料，且可以記錄完整的雷射脈衝回波形態。全波形光達的波形資訊除了可以提高點雲的密度外，透過對回波波形的模擬與分析，可進一步用以反演雷射光束所涵蓋三度空間範圍內各目標物件的其他物理特性。因此，全波形光達系統可同時提供目標區域豐富的幾何和反射波形資訊，有很大的潛力可以進行更複雜和先進的應用。現行光達資料的處理方法皆是針對以測距及座標為主的離散點雲資料所設計，並無法完全適用於全波形光達的資料處理和分析，也無法充分發揮全波形光達資料的優點。因此，有必要研發創新性的資料處理和分析方法，以更有效地利用全波形光達遙測進行前瞻性的應用。本計畫將以三年為期，針對全波形光達資料的特性，開發有效的演算法與流程，充分探索此種新型態的全波形光達資料，以真正發揮其優點，進行有效的分析和應用。研究重點除了全波形光達資料處理及特徵萃取之演算法和流程的開發外，也將針對植被和都市特徵萃取、目標辨識、三維重建、植被覆蓋坡地地形重建和特性反演等進行實際之應用。本計畫的預期成果除了對全波形光達資料的基本特性有深入的了解外，也將就波形模擬、分析、特徵萃取等研發有效的系統化資料處理分析方法，並進行相關的整合性應用。 Full Waveform LIDAR is an emerging active remote sensing technology. In comparison with conventional laser scanning systems, Full Waveform LIDAR not only can generate point clouds with higher density, it can also record the complete waveform of the backscattered signal echo. Therefore, in addition to range information, Full Waveform LIDAR data also provide a high potential for retrieving physical properties of objects and thus may be used for more sophisticated applications. However, existing LIDAR data processing and analysis are all designed for dealing with range measurements and are not adequate for processing and analyzing waveform information. Consequently, it is necessary to develop innovative data processing and analysis methodology in order to take full advantages of Full Waveform LIDAR remote sensing for sophisticated applications. This project proposes to undertake the advanced research and development of systematic data processing and feature extraction algorithms specifically designed for analyzing Full Waveform LIDAR data. Important topics of the research include: understanding of basic Full Waveform LIDAR data characteristics, waveform modeling and fitting, derivative waveform analysis and volumetric feature extraction. In addition, this research will also integrate developed data processing and analysis algorithms and procedures to perform related applications in vegetated and urban areas. These applications include object reconstruction, target identification, segmentation and classification, and vegetationcoved slope terrain reconstruction. 研究期間：10008 ~ 10107
