Fusion Methods for Hyperspectral Image and LIDAR Data at Pixel-Level

Abstract

Hyperspectral image data and LIDAR data have found to be complimentary modailities in case of remotely sensed images, which can be fused if both are geo-referenced. Hyperspectral images provide the spectral response of each object in the area and can be used to identify the material composition of the image which can be used for the object classification. LIDAR data provides the elevation and geometrical information of the objects in the scene. Pixel-level fusion ensures no loss of information because there is no dimensionality reduction. This paper assesses the different methods of pixel fusion like wavelet transform, IHS transform and linear pixel fusion.