應用地面光達掃描於林分狀況調查之探討

Abstract

光達（LiDAR，Light Dection And Ranging）是近代快速發展的一種三維探測與資料收集技術，經過了長期的技術演進，光達的型式與種類已呈現出多樣化得面貌；地面光達掃描在過去主要是用來偵測與記錄建築物、工業管線、工程評估等具有明確且完整三維結構的物體，應用於森林環境之調查與資料收集的發展還處於初期的階段;相較於人工結構，森林中環境之複雜程度相當難以被衡量。本研究嘗試以1公頃大樣區來完整呈現出一林分中之狀態，希望藉由地面光達掃描建立高精度與高完整性之資料，發展出較有效率之萃取方式來取得樣區林分中的各種介量，如立木位置、胸高直徑、樹高等之完整資訊，來量化該林分之狀態。 研究結果顯示，光達資料在林木株數、立木位置與胸高直徑值之偵測及萃取上有不錯之表現，經與現場量測之結果進行比較後，總株數誤差在1 %以內，立木位置之差異情形沒有統計上的顯著差異；平均之胸高直徑值誤差量則為1~3 cm，顯示了在適當的掃描站點分布以及萃取方法下，地面光達將有能力偵測到大部分之林木位置與胸高直徑；然而在垂直結構上，地面光達掃描所取得的點雲資料在全林分樹高之推估上因遮蔽之影響與限制而難以有良好的表現，除了加強掃描站點規劃外，也可配合其他介量或是單木萃取之方式來取得樹高迴歸式以補其不足，在未來應著重於演算法與萃取過程之精簡與改進，並嘗試以小樣區高密度掃描方式來提升資料在垂直空間上之品質，以獲得更加完整之調查結果。

LiDAR（Light Detection And Ranging）is a kind of powerful instrument that can obtain three dimension information of targeted objects or environmental state. With long term developing, LiDAR has many different forms, such as airborne LiDAR or ground-based LiDAR. Ground-based LiDAR has been mainly used to measure the construction in civil engineering or industrial buildings that content complete shape features; but it’s more complicated and varied in forest environment because of the combination of different kind of organism. In this study, for analyzing the stand traits, we try to obtain the stand information in 1 ha plot with ground-based LiDAR and to retrieve the forest parameters from point cloud data. The result shows that the data collected by ground-based LiDAR has a good performance in the total number of trees (average error is almost 1 %) and their DBH measurement (average error is 1~3 cm), but poor in tree height, volume and canopy construction detection. It should be improved to increase the quality of stand vertical construction data retrieved from point cloud data, but we notice that the environmental state plays a very important role in this issue. In the future studies, we should focus on developing better algorithms and retrieving processes, and design the proper investigation manner to gain the better and complete data with ground-based LiDAR.