Application of geoinformatics to classify the landcover from satellite imagery using NDVI - A case of peri-urban regions of Southern Chennai

Abstract

The peri-urban regions of the Indian cities are growing at a rapid rate, engulfing all the available lands in their vicinity. It is vital to map the changes that occur in the available green areas to plan for sustainable urban environments and ecosystems. Normalized Difference Vegetation Index (NDVI) is a graphical indicator that utilizes multispectral remote sensing techniques for the classification of land cover. The aim of the current study is to use the Geographic Information Systems (GIS) tools and its indices to classify the land cover in the areas of interest. For this study, the primary data collection includes the identification of the study area, collection of the toposheets from the Survey of India and satellite images from the United States Geological Survey (USGS) for the study year 2021. The methodology includes the acquisition of data, image processing, application of NDVI technique, accuracy assessment and the presentation of results. For this exercise, the land cover has been classified under three categories namely, water, land and vegetation. The analytical outcomes segregated 30.06 sq.km (3%) area as water bodies, 152.85 sq.km (15%) as land/barren area and 821.97 sq.km (82%) as areas covered with vegetation, out of the total study area of 1004.88 sq. km. The obtained results were validated through field visits and ground truthing method. The accuracy assessment produced an overall accuracy of 91% with a kappa coefficient value of 0.87. Since the peri-urban areas are under constant transformation, it is challenging and time-consuming to physically map the landcover, especially with the lack on definite boundaries. Hence, the current study recommends that NDVI can be used as an effective technique for the classification of Landsat imagery in the constantly-evolving peri-urban regions.