The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Download
Publications Copernicus
Download
Citation
Articles | Volume XLII-2/W13
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-2/W13, 1127–1133, 2019
https://doi.org/10.5194/isprs-archives-XLII-2-W13-1127-2019
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-2/W13, 1127–1133, 2019
https://doi.org/10.5194/isprs-archives-XLII-2-W13-1127-2019

  05 Jun 2019

05 Jun 2019

TRAJECTORY-BASED VISUALIZATION OF MMS POINT CLOUDS

G. Takahashi1,2 and H. Masuda1 G. Takahashi and H. Masuda
  • 1Dept. of Mechanical Engineering and Intelligent Systems, The University of Electro-Communications, 1-5-1 Chofugaoka, Chofu, Tokyo, Japan
  • 2Dept. of R&D, KOKUSAI KOGYO CO., LTD., 2-24-1 Harumi-cho, Fuchu-shi, Tokyo, 183-0057, Japan

Keywords: Mobile Mapping System, point clouds, visualization, data construction, labelling, road facilities

Abstract. MMSs allow us to obtain detailed 3D information around roads. Especially, LiDAR point clouds can be used for map generation and infrastructure management. For practical uses, however, it is necessary to add labels to a part of the points since various objects can be included in the point clouds. Existing automatic classification methods are not completely error-free, and may incorrectly classify objects. Therefore, even though automatic methods are applied to the point clouds, operators have to verify the labels. While operators classify the point clouds manually, selecting 3D points tasks in 3D views are difficult. In this paper, we propose a new point-cloud image based on the trajectories of MMSs. We call our point-cloud image trajectory-based point-cloud image. Although the image is distorted because it is generated based on rotation angles of laser scanners, we confirmed that most objects can be recognized from point-cloud images by checking main road facilities. We evaluated how efficient the annotation can be done using our method, and the results show that operators could add annotations to point-cloud images more efficiently.