Change Detection of Nallamala Forest, India using Sentinel-1 SAR Data, Supervised Classification and ANOVA

Temporal effect (Change Detection) is measured using series of consecutive multi-temporal data sets at same location with the help of various methods in the past. Data sets of spectral, multi-spectral, hyper-spectral, thermal and microwave bands are observed. Various methods were applied to estimate Land Use Land Cover from multi-temporal datasets to identify the Change Detection (CD) in the study area. The adapted methods include spatial classification in RS and GIS, python codes for supervised and unsupervised classification etc. In this study, CD of Nallamala forest, India (sharing between States of Telangana and Andhra Pradesh) is done using Sentinel-1 SAR data in Sentinel Applications Platform (SNAP). Supervised classification techniques such as Random Forest, KD Tree and KNN, Maximum Likelihood and Minimum Distance methods are applied on summer season data over time period from 2020 to 2024 considering six classifications to find CD. Three classifications each are considered under forest cover and non-forest cover with in Nallamala forest. Single factor ANOVA tests are done for checking statistical suitability of methods. Results prove that ‘Only Forest Area’ within Nallamala forest remains unchanged in considered duration. Random Forest and KDT & KNN methods performed well with least mean difference and least F-statistic in all hypothesis tests.