<b>Semi automatic extraction of local roads base on genetic algorithms</b> - doi: 10.4025/actascitechnol.v32i2.4763

Abstract

The aim objective of this paper is to present a semi-automatic extraction of local roads. The research is divided in two different phases. In the first one, a method to determine road width is proposed; in the second one, a fitness function for genetic algorithms is proposed. The referred function makes possible the detection of candidate segments for local roads axis. Three digital images are used in conducting the experiments. For the study, the validation of the obtained results was accomplished based on a reference image. The reference image was created through manual vectoring, and an accuracy evaluation was computed based on the following indexes: completeness, correctness and RMS. The obtained results point out the proposed methodology as promising.