PROJECT TITLE :
Iterative Hyperspectral Image Classification Using Spectral–Spatial Relational Features
This paper describes the principles and implementation of an algorithm for the classification of hyperspectral remote sensing pictures. The proposed approach is novel and will be included within the category of the spectral–spatial classification algorithms. The components of novelty of the algorithm are as follows: one) the implementation of 2 classifiers that work iteratively, every one exploiting the choice of the other to improve the coaching section, and a couple of) the utilization of relational features based mostly on the current labeling and on the spatial structure of the image. The two classifiers are fed with the spectral features and with the spatial options, respectively. The spatial features are engineered using the relative abundance of each category in a neighborhood of the pixel (homogeneity index), where the neighborhood is properly defined. An vital contribution to the success of the strategy is the adoption of a multiclass classifier, the multinomial logistic regression, and a proper use of the posterior probabilities to infer the class labeling and build the relational knowledge. The results of the 2 classifiers are eventually combined by means of an ensemble call. The algorithm has been successfully tested on three customary hyperspectral pictures taken from the Airborne Visible–Infrared Imaging Spectrometer and ROSIS airborne sensors and compared with classification algorithms recently proposed within the literature.
Did you like this research project?
To get this research project Guidelines, Training and Code... Click Here