International Journal of Applied Mathematics and Machine LearningAuthor's: Jie Zhao, Yanchun Zhang and Yongmei Qi
Pages: [93] - [122]
Received Date: September 15, 2014
Submitted by: Shuaiqi Liu.
The three-dimensional reconstruction technology is an important tool to analyze and study external shape and space status of special objects. The two-dimensional image series of an object, obtained by optical image sensor, can be used to perform three-dimensional reconstruction. But there exist cases that some objects are covered by some media or located inside some media. Hence, we can only analyze the distance sequences from one-dimensional sensors to obtain three-dimensional information of the target. However, because of little information and weak correlation with one-dimensional data, it is very difficult to restore the complete three-dimensional information. A novel method is proposed to perform three-dimensional reconstruction by mining one-dimensional range data sequences. This method can be widely used in most of the one-dimensional sensor systems. Finally, simulation experiments on two typical radar systems show the feasibility and efficiency of our proposed method.
3D reconstruction, range data sequence, euclid update, bundle adjustment.
