Automatic Identification and Extraction of Yellow-Ipe Individuals in Images Obtained by UAV

Abstract

Digital image processing (DIP) represents a set of techniques that allow not only to perform the treatment of images, but also to facilitate the extraction, analysis and interpretation of information contained in digital images. Along with the emergence of Unmanned Aerial Vehicles (UAVs), significant gains were observed in terms of spatial and temporal resolution. Images acquired by UAV can be used in several activities, including those focused on environmental conservation, such as the monitoring of tree species protected by law or in extinction, which provides greater control of changes in conservation areas. Thus, this study proposes the application of DIP techniques in images acquired by UAV in order to detect, extract and spatialize individuals related to species of yellow-Ipe. The methodology includes the development of routines, using the Python language, and uses the maximum likelihood classifier and mathematical morphology techniques. As results, we obtained the geographic coordinates of the individuals detected in the images. The results were coherent with reality and showed that the algorithm can be applied for environmental purposes. The proposed approach simplifies the organization of the work manner in solving environmental problems involving the localization of forest species for their management and/or control.