Autonomous embedded navigation system: detection of tracks with monocular camera and optimization based on genetics algorithms

Abstract

This project is part of a larger one entitled "Autonomous Embedded Navigation System" developed in the Mobile Robotics Lab whose main objective is to develop a vehicle capable to move around in a semi-autonomous way. The student's project will make the identification of lanes of traffic on highways and in urban environments. For this, he will use images from a monocular camera attached to the vehicle for the implementation of algorithms post-processed in Matlab. The research will be based on three edge detection algorithms: Laplacian Operator, Canny Operator, and Sobel Operator. These algorithms will be tested and optimized with the support of a genetic algorithm for the determination of the most efficient. It also will be examined the efficiency of using Hough Transform for the identification of tracks after the detection of edges.(AU)