Classification of remotely sensed images using clonal selection theory of Artificial Immune System

摘要

In general, this paper deals with Image Processing using Metaheuristics Optimization Algorithms (IP-MOA). We are focused on supervised classification of remotely sensed images using a clonal selection theory of an Artificial Immune System (AIS). We shall propose a comparative study between the maximum likelihood (MLLH) classifier which is statistical and probabilistic approach and artificial immune system (AIS) which is a bio-inspired approach and commonly named “metaheuristcs”. The most motivations to explore this new kind of approaches for data classification are also presented. MLLH and AIS are applied to classify a multispectral image acquired on June 2001 by ETM+ sensor of Landsat-7 satellite. This multi-band image covers a northeastern part of Algiers (Algeria). From obtained results, we concluded that AIS approach may present a promising metaheuristic classifier for data classification.