Evaluating Machine Learning Algorithms for Landuse Dynamics

The study analyses various machine learning algorithms for land use and land cover classification and change detection of prakasam district, Andhra Pradesh. The study makes use of multi- temporal Remote Sensing imagery for the years 2017 and 2023 from Landsat - 8 sensor of Prakasam district. Three different machine learning algorithms were used to detect changes in land cover features over time based on classes like vegetation cover, water bodies and human settlements. Later, the work aims to present a complete understanding of change detection methodology used within this research, from preprocessing high-quality Landsat 8 satellite images for better interpretation and identification of spectral values, up to machine learning algorithms in terms land use/land cover classification. The findings reveal the positive transitions in areas of land use/cover between 2017 and 2023, especially cultivated lands as well as water bodies to less extent with a few urban parts. This research will be useful for the remote sensing community as an example of how to detect changes in land cover using ML methods with a high degree of accuracy, overtaking the traditional methods.