Automation in meliorated Forest Change Detection based on Landsat Satellites in the Upper Mutha basin, Pune District

Automation is a major key component in geospatial technologies. It provides ease while dealing with large-scale studies, and enhances yield precision in decision-making processes. Landsat satellites have proven the Earth’s surface land monitoring strength. Forest Change Detection (FCD) is one of the monitoring elements with proficient precision. Normalized Difference Vegetation Index (NDVI) trad approach-based FCD has been calculated from Landsat – 5 Thematic Mapper (TM) 1992, Landsat – 7 Enhanced Thematic Mapper Plus (ETM+) 2002, and Landsat – 8 Optical Land Imager & Thermal Infrared Sensors (OLI & TIRS) 2013 respectively. Overall changes in the studied area indicate positive changes at 12.93% & 62.64%; no change at about 43.22% & 33.01%, and negative changes at 43.85% & 04.35% with 70.81 overall accuracies. NDVI-based FCD has limitations like non-eliminated reflectance errors leading to false-positive results therefore, multi-band Tasseled Cap Coefficient Transformation (TCCT)has been used for meliorated FCD. The new approach eliminated known errors and include the application of a Feature Manipulation Engine (FME) to automate FCD through the combination of Artificial Intelligence (AI) and Machin Learning (ML). About 49.03% & 55.05% of estimated land indicate stable vegetation, 35.61% & 24.78% positive change, and 15.37% & 20.17% negative change with a more prominent 88.20% overall accuracy. The second approach has proven meliorated FCD proficiency and effective applicability for sustainable forest management.