Analysis of Forest Fragmentation in Areas of the Atlantic Forest Biome from Various Sensor Systems

Abstract

As human population growth has intensified, deforestation of native forest areas has also intensified. One of the most degraded Brazilian biomes is the Atlantic Forest, which after intense man action has become highly fragmented. The present study aimed to analyze the real influence of the spatial resolution of orbital images in the automatic detection of forest fragments in the sub-basin of the secret stream that is located within the Atlantic Forest biome. Thus, the RapidEye / REIS, Sentinel-2A, Landsat8 /OLI and Modis / TERRA sensors were used, with spatial resolutions of 5 m, 10 m, 30 m and 250 m, respectively. The images were processed mapping the land use and cover by supervised classification using the Bhattacharya algorithm 99.9%, according to seven thematic classes: Water, Exposed Soil, Urban Area, Native Forest, Forest Planted, Field and Agriculture. After the isolated study of the Native Forest class was carried out. From this, a great variation was observed when comparing the images of the RapidEye / REIS and Modis /TERRA sensor systems, showing a 96% decrease in the number of fragments found. Considering the results presented in this study, it can be concluded that the Modis/TERRA sensor system does not present sufficient spatial resolution to properly image the study area.

Keywords

• Remote Sensing
• Forest Fragmentation;
• orbital images