Desarrollo de un código de programación en lenguaje Python para el estudio de zonas forestales

Abstract

The aim of this thesis is to study forest zones through the development of a programming code in Python language capable of determining plant cover by image processing. To achieve this goal two methodologies were proposed, the first one was grayscale image processing and the second one was color image processing of a forest zone previously studied by Bambang Trisakti in his paper “Vegetation type classification and vegetation cover percentage estimation in urban green zone using pleiades imagery”. First, the results of the two methodologies were obtained. Then, these results were compared with the results of the previously study made by mentioned author. This comparison was made in order to define and validate the optimal methodology for the study of forest zones. Color image processing using HSV model was defined as the appropriate methodology for forest zones analysis because it presented the least error and its results were closer to the results obtained in the article used for the validation. Afterwards, the color methodology was applied in images of the Wildlife Refuge La Chiquita and a statistical analysis was performed in the image processing results to conclude that the developed methodology gave the expected results. Finally, this thesis showed that it is possible to study physical-thermal parameters of forest zones using image processing because the results of the image processing was related with the soil moisture in the study area.