A novel descriptor optimization method for multispectral images

Abstract

This paper presents an optimized descriptor method for multispectral images. The method proposed is based on LGHD (Log-Gabor Histogram Descriptor)[1]. Initially, all feature points are detected from Long wave Infrared and Visible spectrum images, and descripted by LGHD, then PCA (Principal Component Analysis) is used to reduce the dimension of the two different descriptors, finally the optimized descriptors are used to match the points. Experimental results show that proposed approach achieves a better matching performance than LGHD.