MAPPING LAND COVER VIA SATELLITE IMAGES

MAPPING LAND COVER VIA SATELLITE IMAGES

Gowthami V1,  Vinay Patel G.L2

[1] Student, Department of MCA, BIET, Davangere

[2] Assistant Professor, Department of MCA, BIET, Davangere

ABSTRACT

The current study examines how image fusion impacts land-cover categorization accuracy using a range of machine learning algorithms. According to our research, picture fusion outperforms single-source imagery alone in categorization. This study uses multi-temporal and multi-spectral Landsat satellite pictures to investigate the urban growth in the Bangalore region. The study, which spans 39 years, examines alterations in the land cover in a 2182 km² area utilizing data from remote sensing. India's rapid population growth and urbanization have significantly impacted resource availability, making land use changes need to be monitored. The evolution of Bangalore's land cover is monitored using Landsat's multi-temporal and multi-sensor technology, which includes MST, TTM, and ETM+ pictures from 1973, 1992, 1999, 2002, 2005, 2008, and 2011. By applying the maximum likelihood classification method, temporal alterations were found and four types of land cover were revealed: desert terrain, vegetation, water, and built-up areas. The findings show a significant rise in urban areas, from 6% in 1973 to 25% in 2011, with 43% of the observed changes being attributed to urbanization. This study emphasizes the significance of ongoing remote sensing to monitor resource management and urban growth, as well as  the notable shift of arid terrain to densely forested areas, impacting zones of plants and water.

Keyword: Land-cover classification, Machine Learning.