Techniques selection for classification of aerial images unmanned aerial vehicle derived more appropriate for environmental analys

Abstract

The research tested photographs classification techniques with high spatial resolution, from an unmanned aerial vehicle (UAV) fixed wing, focusing on the analysis of land use in cerrado environment. The study area is located in the Campo Verde, Mato Grosso. The aerial photography used has a spatial resolution of approximately 10 cm. We used supervised classification techniques and unsupervised classification. The supervised classification by region (segmentation) with degree of similarity and 20 area 200, showed the best result according to the cluster analysis (correlation of 94%) compared with the result of photo-interpretation map. The search result is important benefit in terms of automated thematic maps, environmental planners to act in monitoring, environmental monitoring and expertise, especially in large degraded areas. In addition, survey data can be used for habitat assessment protocols, which collaborate with the consolidation of the current Brazilian Forest Code.