Title :
Fine building model based on Terrestrial Laser Scanning and Support Vector Machine image segmentation
Author :
Jian, Li ; Youchuan, Wan ; Shiwei, Shao ; Lin, Xiang ; Peng, Tong
Author_Institution :
Sch. of Remote Sensing & Inf. Eng., Wuhan Univ., Wuhan, China
Abstract :
It is true that Terrestrial Laser Scanning (TLS) can quickly acquire the detailed and accurate three-dimensional data of ground objects, thus it is widely used in digital city, the protection of historic buildings and the deformation monitor of bridges, etc. However, the huge data of the point clouds would create huge TIN (Triangulated Irregular Network) models. It is difficult to use these data on digital city because there are too many buildings and objects not needed and it is also a waste of time to create this model. As the TLS can easily achieve detailed and accurate recording of surfaces, in addition to which, the digital camera mounted on the TLS could also acquire rich textures besides the measurements of edges and corners, it is easy to extract the outlines of objects by combining point clouds with the according images which are segmented by SVM(Support Vector Machine). However TIN models created on those parts are difficult to get outlines. The fine models can be accurately and quickly created using outlines whereas the structure is simple or using TIN modeling whereas the structure is difficult. Thus it is a fine model which is small in size and simple enough to operate. This paper describes the steps of the fine modeling of Tianjin Concert Hall from the VZ400 laser scanner.
Keywords :
data acquisition; image segmentation; image texture; optical scanners; support vector machines; Tianjin Concert Hall; VZ400 laser scanner; corner measurements; digital camera; digital city; edge measurements; historic building protection; image segmentation; point clouds; support vector machine; terrestrial laser scanning; textures; triangulated irregular network; Buildings; Clouds; Data models; Image segmentation; Laser modes; Solid modeling; Three dimensional displays; Fine 3D Model; Point Cloud; Support Vector Machine (SVM); Terrestrial Laser Scanning (TLS);
Conference_Titel :
Geoscience and Remote Sensing (IITA-GRS), 2010 Second IITA International Conference on
Conference_Location :
Qingdao
Print_ISBN :
978-1-4244-8514-7
DOI :
10.1109/IITA-GRS.2010.5602359