SWARM INTELLIGENCE IN REMOTE SENSING

Abstract

In recent years remote sensing is used for the knowledge elicitation of earth’s surface and atmosphere on a very global scale. Land cover mapping is a method for acquiring the geo-spatial information from satellite data. The land cover problem is solved by image classification of the satellite image. There are soft computing techniques like fuzzy sets and rough sets for remote sensing image classification. In this thesis we focus on the optimised approach of image classification of satellite multi spectral images. This thesis deals with the land cover mapping by using swarm computing techniques. Here we have classified the Satellite images by Swarm Intelligence approach. We are using improved Ant Miner algorithm i.e, cAnt Miner and the Hybrid PSO-ACO algorithm for the rule-set generation, used for image classification. The Swarm Intelligence inspired techniques produce comparable results with the classical approaches Minimum distance classification, Maximum Likelihood method. There are...