AVALIAÇÃO DE TÉCNICAS DE PROCESSAMENTO DE IMAGEM NA IDENTIFICAÇÃO DE IMPACTOS AMBIENTAIS EM LINHA DE TRANSMISSÃO DE ENERGIA ELÉTRICA

Autor: Claudionor Ribeiro Silva, Admilson da Penha Pacheco, Samuel Lacerda de Andrade
Rok vydání: 2016
Předmět:
Zdroj: Ciência e Natura. 38
ISSN: 2179-460X
0100-8307
Popis: The projects in the energy sector are basically composed of cables and metal structures. While providing benefits, installation and operation of these projects have significant environmental impacts along its path. The aim of this study was to evaluate image processing techniques to identify environmental impacts in energy powerline (path: Ibicoara-Brumado in Bahia / Brazil). Tests were performed on orthophotos with MaxVer classification and segmentation process object-oriented, with experiments integrating LiDAR. The results were analyzed on the basis of commission errors (user accuracy), omission errors (producer accuracy) and kappa index. According to Kappa index calculated, all tests showed ratings classified as excellent. The use of integrated LiDAR with orthophotos demonstrated an improvement in refining the results, with higher accuracy in classification and segmentation. The main evidence of environmental impacts is areas with exposed soil related to access roads, implementation of the infrastructure and use/occupation with agriculture. The area related to exposed soil was 16% demonstrating the need for prevention and conservation actions in LT projects.
Databáze: OpenAIRE