Recent advances in sensor technology have led to an increased availability of hyperspectral remote sensing images with high spectral and spatial resolutions. These images are composed by hundreds of contiguous spectral channels, covering a wide spectral range of frequencies, in which each pixel contains a highly detailed representation of the reflectance of the materials present on the ground, and a better characterization in terms of geometrical detail. The burst of informative content conveyed in the hyperspectral images permits an improved characterization of different land coverages. In spite of that, it increases significantly the complexity of the analysis, introducing a series of challenges that need to be addressed, such as the computational complexity and resources required. This dissertation aims at defining novel strategies for the analysis and classification of hyperspectral remote sensing images, placing the focal point on the investigation and optimization techniques for the extraction and integration of spectral and spatial information. In the first part of the thesis, a thorough study on the analysis of the spectral information contained in the hyperspectral images is presented. Aiming at overcoming the issues and limitations that affect the analysis of hyperspectral image classification, the following contribution is made:
1. Deeply investigate the behavior and performance, in terms of classification accuracy, of the most widely used deep learning algorithms in the remote sensing field, under different experimental set-ups.
2. Developed novel strategies to limit the Hughes phenomenon for hyperspectral image classification by exploiting deep learning based architectures such as stacked auto encoder, Deep belief net and PCANet.
3. Designed an innovative technique for spatial information extraction by using boundary adjustment based criteria, while addressing the information redundancy and noise issues.
4. Defined a methodology that integrates both spectral and spatial information within a classification scheme.