AUTOMATED LAND USE LAND COVER CLASSIFICATION OF SANDUR TALUK, BELLARY DISTRICT, KARNATAKA, INDIA.

The Land Use Land Cover(LULC) assessment forms the foundation for change detection studies. LULC can be used as an effective and efficient approach in not only monitoring land transformation, natural resources, natural disasters but also in land suitability applications and urban planning and development. This paper contributes in classifying and analyzing multi spectral images with a machine learning approach for the purpose of studying land transformation and change detection. Supervised classification algorithms involve making use of apriori knowledge of the data in order to extract statistical parameters to aid the process of classification. Some of the most popular and efficient supervised classification techniques are Maximum Likelihood Classifier(MLC) and Support Vector Machine(SVM). This paper presents LULC delineation using supervised classification algorithm using the Maximum Likelihood Classifier for the study area Sandur, Bellary District, Karnataka, India. Using QGIS the study area was delineated into six classes viz. water body, agricultural area, forest, built up area, mining area and barren land. The training process was conducted using Google Earth Pro. The overall accuracy of classification for 2001, 2011 and 2021 obtained was 87.5%, 86.3% and 87.32% respectively. The Kappa Coefficient values obtained were 85.01%, 83.43% and 84.8% respectively.