Target Recognition Algorithm for Fused Hyperspectral Image by Using Combined Spectra

Abstract

A spectrum recognition method of fused hyperspectral image is proposed based on the normalized spectral information divergence and normalized spectral angle mapper. First, the hyperspectral image and registered high-resolution image are fused. Subsequently, the spectrum of the targets in the original hyperspectral image is abstracted and regarded as the reference spectrum information. Finally, the spectral information from the fused hyperspectral image is used to recognize the targets by using the proposed algorithm. Some numerical simulations are made to test the validity and capability of the proposed combined spectrum recognition algorithm.

Keywords:

• hyperspectral image
• spectral variability
• target recognition

Acknowledgments

The authors would like to thank Fei Zhang and Mingzhe Li for assistance with the experiment and Chengwei Yang for valuable discussion. The authors are indebted to the four reviewers for their helpful comments.

Notes

Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/lstl.