Satellite-Based Land Cover Classification in the Itajaí River Basin: Methods and Analysis

Abstract. The cities within the Itajaí River Basin in Brazil have been experiencing continuous flooding in recent years, causing significant material damage to the population and public services. This study aims to present land use and classification techniques based on image analysis obtained from the CBERS-4A satellite, using the WFI sensor. The image was cropped to include only the Itajaí River basin. Four classification methods were employed: Maximum Likelihood (Maxver), Minimum Distance, Spectral Angle, and Random Forest. Classified maps were generated for each algorithm, and the results of each land cover class area were analyzed. Validation was performed using sample areas of each class in the original image, presenting the area of each class obtained by each algorithm and the confidence level for each.