Object-Based Classification to Evaluate LULC Changes and Socio-Economic Mobility with Google Earth Engine: A Case Study of Rajarhat-New Town Agglomeration, Kolkata, India

Abstract

Rajarhat-New Town lies north of the East Calcutta Wetlands, a Ramsar site, and a natural sewage treatment plant for the Kolkata metropolitan area. However, the rapid growth of residential, commercial, and industrial developments has induced land use land cover (LULC) change that dramatically affects the inhabitants and the environment. Therefore, to identify development pathways, this study aims to measure land-use adjustments and land cover impacts on the socio-economic mobility of Rajarhat-Newtown. The earth observation data from 1991 to 2021 has been analysed to quantify the pattern of LULC changes using an Object-Oriented (OO) classification approach, integrated with the flexible cloud framework in the Google Earth Engine (GEE) platform to analyse the LULC. The final classification approach is adopted by combining the machine learning (ML) algorithms of Support Vector Machine (SVM) and Random Forest (RF). Principal Components Analysis (PCA) has been applied to the significant "Grey-Level Co-occurrence Matrix" (GLCM) indices to synthesise the textual data needed for the OO categorisation within a single band. From the results, negative changes have been identified for almost all the features concerning the base year of 1991. Furthermore, a time series analysis has been undertaken to monitor the spatial LULC changes between 1991 and 2021 at 10-year intervals over the region. Aside from that, special attention is being put on converting agricultural land to built-up areas, supporting the socio-economic transition of the study area's population. Based on the changing LULC pattern from 1991 to 2021, a prediction of LULC for 2031 has been executed for the study area. Conclusively, identifying LULC changes and their pathways has beneficial and detrimental impacts on the region's society, economy, and environmental sustainability.