Processamento multitemporal de imagens landsat para monitoramento de cobertura de solo e dinâmica da linha de costa na região de Mecúfi - Cabo Delgado

Abstract

Using the technique called Remote Sensing, the present work aims to analyze Landsat images for monitoring ground cover and shoreline dynamics in Mecúfi region. The methodology consisted of data acquisition and Digital Image Processing (PDI): The data acquisition was done on the US Geological Survey Online Platform (USGS), where 2 scenes (orbital images) of Landsat 5-TM and Landsat 8-OLI satellites were downloaded in a 17-year time series, dated to 08/14/2001 and 23/04/2018, both in orbit 164, point 69 and selected according to lower cloud cover. The PDI involved three steps namely: Preprocessing, Processing and Postprocessing. Preprocessing consisted of converting digital numbers into reflectance, clipping, and extracting masks, Processing consisted of shoreline outlines, RGB channel color compositions, supervised classification, spectral índex calculations, and calculations involving statistical parameters in the Excel 2007 spreadsheet and Minitab 18 statistical Software. The last post-processing step was an evaluation of the accuracy of the MAXVER classification. The results were presented as tables, graphs and maps of land use and coverage. The Mecufi region focuses on shoreline progressions or retrogradations, which were visualized on 1: 10,000 scale maps in ArcGis 10.5. The quality of the MAXVER mapping assessment was very good for overall accuracy around 62.7% and good for kappa Index with 55.3% in both Sensors. Minimum and maximum reflectance threshold values are crucial for solving target discrimination problems through Landsat image classifying algorithms.