車載光達點雲中直立圓桿之模塑

Modeling of Vertical Pole-Like Objects from Vehicle-Borne LiDAR Point Cloud

邱式鴻(Shih-Hong Chio)；吳志文(Chih-Wen Wu)

車載光達 ； 直立圓桿 ； 移動測繪 ； Mobile Mapping ； Vehicle-Borne LiDAR ； Vertical Pole-Like Object

國土測繪與空間資訊

2卷1期（2014 / 01 / 01）

23 - 41

繁體中文

隨著移動測繪系統（mobile mapping system，MMS）發展，車載光達系統可有效獲取詳細路廊（road corridor）的三維點雲資料。由於車載光達系統紀錄大量的點雲資料與複雜的路廊資訊，需經處理，才可模塑點雲中路廊的物件。其中，直立圓桿是路廊資訊中重要地物的基本元件，故本研究發展模塑車載點雲中直立圓桿的演算法。直立圓桿模塑的演算法必須面臨車載光達系統所蒐集的點雲無法完整涵蓋整個直立圓桿表面，以及所蒐集的點雲不僅包含直立圓桿上的點、亦可能包含附著於直立圓桿上的物件點（如廣告招牌等）等兩個重大的問題，意即這些點雲資料蒐集不完全且包含許多雜訊。因此，本研究所發展的演算法中先以物空間資訊將車載點雲中的地面點濾除並留下非地面點，其中地面點包含屬於路面與人行道上的點。接著，將非地面點透過八分樹體元結構化（octree-structured voxel space）後，並以其相鄰性加以群聚，進而組成非地面點之點群。由於直立圓桿上可能含有許多附著物（如廣告招牌等），因此本研究發展以RANSAC（RANdom SAmple Consensus）為基礎之演算法判斷經前述處理聚集後之非地面點群是否包含直立圓桿，並計算其圓面參數。實驗結果顯示在複雜的街景中，本研究所發展直立圓桿模塑之漏授率（Omission）為31.8%、誤授率（Commission）為60.5%；各直立圓桿求定之圓面參數與人工量測之結果比較在X坐標方向的RMSE為0.032 m，在Y坐標方向RMSE為0.046 m，而半徑的RMSE則為0.031 m。

With the development of mobile mapping system (MMS), vehicle-borne LiDAR system can obtain precise 3D point cloud of the detailed road corridor efficiently. Because the vehicle-borne LiDAR system records numerous points cloud and complicated information of road corridor, those point cloud can be used in reconstructing the objects in road corridor after data preprocessing. In the objects of road corridor, vertical pole-like objects is one of most important and basic objects. Therefore, this study focuses on modeling vertical pole-like objects from vehicle-borne LiDAR point cloud. In the vehicle-borne point cloud, the data might describe vertical pole-like objects incompletely and might be always with many irrelevant points from the attached objects. First of all, the ground points will be filtered through scene knowledge. Then, the non-ground points will be clustered through the octree-structured voxel space and connected-component labeling (CCL) algorithm. In the clustering LiDAR points, they cannot describe complete vertical pole-like objects and some points might belong to attached objects. Therefore, the automatic algorithm based on RANSAC (RANdom SAmple Consensus) is developed to extract and model vertical pole-like objects from those clustering LiDAR points in this study. The result shows the omission of vertical pole-like objects by the modeling approach developed by this study is 31.8% and the commission of vertical pole-like objects is 60.5% under the complicated street environment. Moreover, the circle parameters of vertical pole-like objects, i.e. the coordinate of circular center and the circular radius, are compared with those measured manually in this study. The RMSEs in X, Y coordinate components are 0.032 m and 0.046 m, respectively. The RMSE of circular radius is 0.031m.

1. Mumtaz, S. A., 2007, Extracting physical and environmental information for Irish roads using airborne laser scanning (ALS), MPhil Research Project, Dublin Institute of Technology. http://www.tnet.teagasc.ie/rerc/downloads/EORS/3rd%20session/salman%20ali.pdf.
2. 迅聯光電有限公司，2010，產品世界--移動載具型雷射掃描儀，迅聯光電有限公司，http://www.linkfast.com.tw/product_rieg_c.htm，取用日期：2010 年12 月28 日。
3. Bolles, R.,Fischler, M.(1981).A ransac-based approach to model fitting and its application to finding cylinders in range data.Proceedings of the 7th International Joint Conference on Artificial Intelligence,Vancouver:
4. Brenner, C.(2009).Extraction of features from mobile laser scanning data for future driver assistance systems.Advances in GIScience, Lecture Notes in Geoinformation and Cartography
5. Chaperon, T.,Goulette, F.(2001).Extracting cylinders in full 3D data using a random sampling method and the Gaussian image.Proceedings of the Vision Modeling and Visualization Conference,Germany:
6. Chen, Y.-Z.,Zhao, H.-J.,Shibasaki, R.(2007).A mobile system combining laser scanners and cameras for urban spatial objects extraction.Proceedings of the Sixth International Conference on Machine Learning and Cybernetics,Hong Kong:
7. Fischler, M.A.,Bolles, R.B.(1981).Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography.Communications of the ACM,24(6),381-395.
8. Kukko, A.,Andrei, C.-O.,Salminen, V.-M.,Kaartinen, H.,Chen, Y.,Rönnholm, P.,Hyyppä, H.,Hyyppä, J.,Chen, R.,Haggrén, H.,Kosonen, I.,Čapek, K.(2007).Road environment mapping system of the Finnish Geodetic Institute - FGI Roamer.International Archives of the Photogrammetry, Remote Sensing,XXXVI(3/W52),241-247.
9. Lehtomäki, M.,Jaakkola, A.,Hyyppä, Juha,Kukko, A.,Kaartinen, H.(2010).Detection of vertical pole-like objects in a road environment using vehicle-based laser scanning data.Remote Sensing,2,641-664.
10. Lukács, G.,Martin, R.,Marshall, D.(1998).Faithful least-squares fitting of spheres, cylinders, cones and tori for reliable segmentation.Proceedings of the 5th European Conference on Computer Vision,Freiburg: Germany:
11. Luo, D.,Wang, Y.(2008).Rapid extracting pillars by slicing point clouds.International Archives of the Photogrammetry, Remote Sensing,XXXVII(B3b),215-218.
12. Manandhar, D.,Shibasaki, R.(2001).Feature extraction from range data.Proceedings of the 22nd Asian Conference on Remote Sensing,Singapore:
13. Marshall, D.,Lukacs, G.,Martin, R.(2001).Robust segmentation of primitives from range data in the presence of geometric degeneracy.IEEE Transactions on Pattern Analysis and Machine Intelligence,23,304-314.
14. Petrie, G.,Toth, C.K.(2009).Terrestrial laser scanners: in topographic laser ranging and scanning, principles and processing.Boca Raton:CRC Press.
15. Rabbani, T.,van den Heuvel, F.(2005).Efficient hough transform for automatic detection of cylinders in point clouds.International Archives of the Photogrammetry, Remote Sensing,XXXVI(3/W19),60-65.
16. Schnabel, R.,Wahl, R.,Klein, R.(2007).Efficient ransac for point-cloud shape detection.Computer Graphics Forum,26,214-226.
17. Schwarz, K.P.,El-Sheimy, N.(2004).Mobile mapping systems - state of the art and future trends.International Archives of the Photogrammetry, Remote Sensing,XXXV(B5),759-768.
18. 王淼(2011)。博士論文(博士論文)。國立成功大學測量及空間資訊學系。
19. 李育華(2009)。碩士論文(碩士論文)。國立成功大學測量及空間資訊學系。