Population and settlement mapping is one of the most important tasks to do for the following purposes a) to assess the anthropogenic stress upon the environment, b) to assess the vulnerability of any community to environmental hazards and disasters, c) to estimate the degree and direction of urbanisation etc. Mixed, complex and confusing spectral characteristics and low spatial resolution are the limitation of Optical-Multispectral remote sensing with conventional classification techniques (Supervised and unsupervised classification) besides Hyperspectral, microwave/SAR (synthetic aperture radar) data provide promising result but often convey complicated analysis and cost-intensives. In this research an alternative methodology has been implemented where open access multispectral data has been used to extract the spatial distribution of settlement and afterward Census population data has been integrated with that spatial map to generate spatial population map. Bayesian network classifier integrated with unsupervised k means technique used to extract settlement later population data of census 2011 has been added according to area of the settlement cluster under a specific micro administrative unit (MAU). Current method has been compared with simple maximum likelihood classification (MLC) and produced better accuracy compared to MLC with overall classification accuracy of 97.83%.
