Share Email Print

Optical Engineering

Target recognition for ladar range image using slice image
Author(s): Wenze Xia; Shaokun Han; Liang Wang
Format Member Price Non-Member Price
PDF $20.00 $25.00

Paper Abstract

A shape descriptor and a complete shape-based recognition system using slice images as geometric feature descriptor for ladar range images are introduced. A slice image is a two-dimensional image generated by three-dimensional Hough transform and the corresponding mathematical transformation. The system consists of two processes, the model library construction and recognition. In the model library construction process, a series of range images are obtained after the model object is sampled at preset attitude angles. Then, all the range images are converted into slice images. The number of slice images is reduced by clustering analysis and finding a representation to reduce the size of the model library. In the recognition process, the slice image of the scene is compared with the slice image in the model library. The recognition results depend on the comparison. Simulated ladar range images are used to analyze the recognition and misjudgment rates, and comparison between the slice image representation method and moment invariants representation method is performed. The experimental results show that whether in conditions without noise or with ladar noise, the system has a high recognition rate and low misjudgment rate. The comparison experiment demonstrates that the slice image has better representation ability than moment invariants.

Paper Details

Date Published: 17 December 2015
PDF: 12 pages
Opt. Eng. 54(12) 123107 doi: 10.1117/1.OE.54.12.123107
Published in: Optical Engineering Volume 54, Issue 12
Show Author Affiliations
Wenze Xia, Beijing Institute of Technology (China)
Shaokun Han, Beijing Institute of Technology (China)
Liang Wang, Beijing Institute of Technology (China)

© SPIE. Terms of Use
Back to Top