應用手持式光達儀於市地地籍現況測量

現況測量於地籍測量作業主要利用測量之現況成果與圖解數化地籍圖經過套疊分析，解決地籍圖之圖地不符問題。現行地籍相關之現況測量多以全測站儀、衛星定位儀施測，但此方法耗費大量人力、時間成本，且無法測得大範圍現況資料；而光達儀器能夠在短時間獲取大量之三維點雲資料，具有高精度、高效率、可視覺化等優點，藉此辨識地物特徵並進而於內業執行現況測量，然而地面光達儀器於市地狹窄、蜿蜒巷弄之現況測量遭遇許多環境限制。本研究使用之手持式光達儀，為一種移動式測繪系統，其採用SLAM演算法進行定位，不需依賴GNSS資料，故可應用於都市地區中狹窄蜿蜒之巷弄環境。本研究首先藉由基於平面特徵之率定方式計算手持式光達儀之乘常數與加常數兩個測距系統誤差參數，以提升手持式光達點雲精度；接著使用手持式光達儀於選定之試驗區進行現況資料蒐集，經點雲過濾與測距系統誤差參數改正後，利用VBS-RTK測得之地面控制點將現況點雲以剛體六參數轉換至TWD97地籍坐標系，並以人工數化方式產製現況成果圖。成果顯示手持式光達點雲數化之現況位置約97%與地測現況點誤差小於15公分，已足夠使用於圖解數化區之現況測量；經率定計算之測距系統誤差改正後，在不考慮改正前後誤差差值小於5毫米的資料下，約73%數化現況位置與地測現況點誤差相較改正前降低，顯示經測距系統誤差改正可進一步降低大部分手持式光達點雲之誤差；最後針對人工數化之現況線段與RANSAC自動化擬合之現況線段進行比較分析，成果表明利用RANSAC可以由現況點雲中去除牆面附著物並擬合牆面位置，RANSAC擬合線段與人工數化線段垂直誤差約介於1至2公分，顯示兩種產製方式皆能獲得可靠現況位置，得以用於圖解數化區之現況線段產製。

關鍵字

地籍測量；現況測量；手持式光達儀；率定

並列摘要

The results of detail survey for cadastral survey are used for the overlap analysis with the digitized graphic cadastral maps to solve the problem of inconsistence between cadastral maps and current land situation. The methods of detail survey are mostly employed the total stations and satellite positioning instruments, but these methods are labor-intensive and time-consuming. Meanwhile, these instruments cannot collect a wide range of land detail data. This study applied the handheld LiDAR scanner, which uses SLAM algorithm for positioning without GNSS data, to collect 3D point cloud with an efficient way for detail survey in urban environments with narrow and winding streets. A plane-based calibration method for the handheld LiDAR scanner was proposed to calibrate the systematic error parameters to improve the accuracy of the handheld LiDAR point cloud. Moreover, urban detail data in the test area was collected by the handheld LiDAR scanner. After point cloud filtering and ranging systematic error correction, the handheld LiDAR point cloud were transformed to the TWD97 cadastral coordinate system using ground control points. The land detail map was artificially digitized from the handheld LiDAR point cloud. The results show that about 97% of the digitized detail positions have errors with less than 15 cm compared to the check positions surveyed by a total station. From the results, it demonstrated the feasibility of using handheld LiDAR scanner to perform urban detail survey in digitized graphic areas. Without considering the error differences less than 5 mm between before and after ranging error correction, about 73% of the digitized detail points whose error is reduced after the ranging error correction, which presents the ranging error parameters can further reduce the errors of most handheld LiDAR point cloud. Finally, the artificially digitized detail lines were compared with detail lines extracted by RANSAC, the results show that the difference is about 1-2 cm, indicating that both production methods can obtain a reliable position of the boundary line and can be used for the production of the detail lines in digitized graphic areas.