Image simulation based analysis for hyperspectral target detection

Lianru Gao; Bing Zhang; Wenjuan Zhang; Qiong Ran

doi:10.1117/12.832782

Translator Disclaimer

30 October 2009 Image simulation based analysis for hyperspectral target detection

Lianru Gao, Bing Zhang, Wenjuan Zhang, Qiong Ran

Author Affiliations +

Proceedings Volume 7498, MIPPR 2009: Remote Sensing and GIS Data Processing and Other Applications; 749829 (2009) https://doi.org/10.1117/12.832782
Event: Sixth International Symposium on Multispectral Image Processing and Pattern Recognition, 2009, Yichang, China

Abstract

Performance of the selected algorithm is crucial for hyperspectral target detection. However, it is sure that there is no algorithm that is appropriate for all hyperspectral data or applications. Result of detection can be affected by many factors, including spectral feature of targets and backgrounds, characteristic of spectrometer and pre-processing method. Thus to improve practicability of hyperspectral target detection, studies should not be focused solely on the algorithms, but also be extended to discussion of above affecting factors. In this paper, hyperspectral imagery measurements and data processing are studied. Six aspects including paints, spectral response, imaging spectrometer mode, image noise and dimension reduction are discussed. Typical abnormity and matching algorithm are applied on images simulated with different factors. The influences of these factors on target detection are analyzed. At last, this research provides a strategy for precision control of target detection in hyperspectral imagery. Researches in this paper can be of great help for algorithm development and algorithm selection in diversified hyperspectral data and applications.

Citation Download Citation

Lianru Gao, Bing Zhang, Wenjuan Zhang, and Qiong Ran "Image simulation based analysis for hyperspectral target detection", Proc. SPIE 7498, MIPPR 2009: Remote Sensing and GIS Data Processing and Other Applications, 749829 (30 October 2009); https://doi.org/10.1117/12.832782

ACCESS THE FULL ARTICLE