Rigel VUX－1無人機光達點雲資料處理及精度評估

Rigel VUX-1 UAS LiDAR POINT CLOUD PROCESSING AND DATA ASSESSMENT

王星為(Hsin-Wei Wang) ； 江晉霆(Chin-Ting Chiang) ； 彭德熙(Te-Hsi Peng) ； 鄭鈺雯(Yu-Wen Cheng) ； 余翠紋(Carol Yu) ； 陳立邦(Leo Chen) ； 謝有忠(Yu-Chung Hsieh) ； 費立沅(Li-Yuan Fei) ； 張國楨(Kuo-Jen Chang)

《中國土木水利工程學刊》 31卷1期 (2019/03) Pp. 29-39

https://doi.org/10.6652/JoCICHE.201903_31(1).0003

摘要

無人機系統的持續穩定發展及雷射測距光達掃瞄系統的迷你化，以無人機搭載高精度光達測量系統成為一新穎且能快速取得空間資訊之技術。唯就本技術，因暫時受限於載具、任務風險及成本等諸多方面的因素，於國內外等產學界仍未大量使用於相關測繪工作。本研究結合Pulse Aerospace之無人直升機Vapor 55，以及Riegl之VUX－1 UAV光達系統，進行無人載具及光達系統之系統整合、飛航任務測試，以及掃瞄資料蒐集及資料處理，同時並進行現地地面測量，以評估光達系統資料之解析度、精度及誤差分析。另一方面並進行無人機影像測繪資料、既有空載光達掃瞄等資料之比較分析工作。本研究之設計掃瞄任務，以單一航帶下方近地點之點雲密度約159點／平方公尺之飛航參數來設計。所掃瞄之初步成果，首先以軌跡解算軟體Applanix POSPac MMS進行無人機之航跡解算，掃瞄之點雲以Riegl點雲處理軟體PiPROCESS來進行點雲之航線平差、拼接，以及點雲分類等工作。針對點雲資料，傳統上空載光達點雲的處理，一般常用Terrasolid之Terrascan, Terramatch等軟體來進行，前人文獻並建議點雲分類相關參數。本研究之基於無人機光達資料，同時運用不同參數進行無人機點雲分類之參數敏感性分析，以求取較為適當的點雲分類參數。藉由自動化的分類將點雲歸類為地面點與非地面點，所萃取之地面點雲密度大於100點／平方公尺，進而用以建置空間解析度高於10公分之數值高程模型。另以e－GNSS、RTK（Real Time Kinematic）及全測站三角三邊測量等現地測量成果，將現地量測之數據與無人機影像、無人機光達，以及空載光達等不同來源之數值地型模型，進行資料之比對與較差分析。模型結果顯示無人機光達之模型高程精度誤差小於5公分。透過本研究呈現無人機光達空間資料技術及精度，並可應用於高解析度測繪工作上。

關鍵字

無人飛機系統 ； 光達 ； 點雲 ； 解析度 ； 精度

並列摘要

The LiDAR sensor mounted on a UAV becomes a new powerful tool for geomatic technology. This study we integrate autonomous unmanned helicopter Pulse Aerospace Vapor 55, carrying Riegl VUX-1 UAV LiDAR with Trimble Applanix AP20 for the surveying mission. Based on the drone and instrument capacity, adjusted by the terrain landform, the optimal drone mission planning and scanning parameters are thus assigned, thus capable to acquire dense point clouds by 159 points/m^2 in nadir direction for a single fly line. To access the dataset, several software packages are used, including: the Trimble Applanix POSPac Mobile Mapping Suite software, GNSS-Aided Inertial post-processing for georeferencing data collected from UAS LiDAR. The Riegl RiPROCESS designed for managing, processing, analyzing, and visualizing and data export for the data acquired based on Riegl Laser Scanners. And finally, access and evaluate the dataset by means of Riegl RiPROCESS software for managing, processing, especially for fly-line adjustment and classification the UAS LiDAR point clouds, so as to compare the UAS dataset with the airborne's. The study tries to evaluate the parameters for fully-automatic point cloud classification by Terrascan, which is used regularly in Taiwan. This paper analyzes the influence and efficiency of different parameters for point cloud classification, to separates the non-ground point from the ground point so as to construct the digital elevation model. Finally, the density of the ground point is higher than 100 points/m^2, thus the spatial resolution of digital elevation model (DEM) is about 10 by 10cm. Compared with the data point measured from site surveying by e-GNSS, RTK (real time kinematic GPS survey) and total station, ground control points and check points, the elevation errors is less than five centimeters; thus the high resolution and high precision digital terrain models (DTMs) are capable to construct. UAS LiDAR point cloud after instrument calibration and flight trajectory adjustment, the surveying data acquisition can achieved as centimetric precision. According to the results, the technology of UAS LiDAR is capable and suitable for high resolution geoinfomatic studies and data acquisition.

並列關鍵字

unmanned aircraft system (UAS) ； light detection and ranging (LiDAR) ； point cloud ； resolution ； precision