以點雲建制的三維模型已經是普遍且廣泛地被應用,不論是使用地面光達掃描或是無人載具(Unmanned Aerial Vehicle, UAV)拍攝影像生成點雲。地面光達是利用雷射光束進行掃描式測距之系統,其優點著重於自動化、大範圍的量測,此一系統可以提供每秒數千到數萬點的三維座標資訊,且其三維座標精度可達公分或公厘等級,可快速獲取高精度的三維座標資訊,但受限於儀器架設位置,對於掃描物件的資料缺口無明顯解決方式;然而, UAV所拍攝的影像,通常會因鄰近物阻擋,造成主要構件的側邊點雲資料產生缺口。在三維模型重建的過程中很多研究投入的是演算法的開發和過程的優化,但仍有許多問題尚未得到解決,其最常見的是使用不完整的點雲資料重建三維模型。本研究評估地面光達點雲的精度和UAV點雲精度,而在比較精度的過程中,輔助使用傳統測量的方式來確認兩者間控制點的精度;藉由座標轉換,對同一結構物,將地面光達與UAV攝影所獲取的點雲資料,使用相同座標點來做接合,以此作為彌補以上述兩種點雲方式所獲得的三維模型之缺漏。 關鍵詞:點雲、UAV、點雲融合
3D model can be restored by employing the point clouds collected by using terrestrial lidar or unmanned aerial vehicle (UAV). However, the point clouds collected by a terrestrial lidar may lose some information because the laser beams emitted from the terrestrial lidar can be blocked. The point clouds collected by UAV may be incomplete because of not enough information available to satisfy the geometric constraint conditions. This paper tries to fuse two kinds of point clouds to form a complete point cloud such that the 3D model of the interested objects can be formed. We propose an approach based on the common coordinate system such that two kinds of point clouds can be fused together. The combined accuracy of the fused model is evaluated. Keyword: Point Cloud Fusion, Coordinate System