Detection of pavement areas in aerial images obtaneid from unmanned aerial vehicles (UAV) using Gabor filter

Abstract

This paper presents the detection of pavement areas in aerial images captured by commercial unmanned aerial vehicle, using the Gabor filter of the MATLAB software for segmentation. The Gabor filter performs the segmentation according to the texture. First, flights were made to capture the images and the or thomosaic map was formed to show the total area of study; the flight height was 45.6 meters. Some images that contained pavement and other objects were selected afterwards to make a comparison of the segmentation between a manual method to calculate the pavement area and the image processing through the Gabor filter, resulting in the analysis of five different images with an average error of 3.16% between both methods.