Development of Multispectral Post-Classification Change Detection Technique

Coastal area that are the comfortable shelter for the growing human population are undergoing biographic and demographic decline. Increasing population density and economic activities are badly threatening the coastal ecosystem. A study of the impact of such civilization on the coastal ecosystem is studied by different researchers. Diversified image-based change detection methods, surveillance systems, sensor technology, bio-diagnosis and research work have taken place for decades. This paper is presenting a development of multispectral post-classification change detection. We hope that it will try to ease the task of researchers that are involved in the field of change detection using computational approaches. According to the comparative analysis, Random Forest (RF) method greatly boosted accuracy by 96.73% in terms of identifying Land-Use Land Cover, which is the highest gain of any method when compared to others.