Design and evaluation of a path detection algorithm in road images using Hough transform

One of the best applications of automated navigation systems is the use of automatic steering systems in agricultural vehicles. In this research، a machine vision algorithm was designed، based on image acquisition from a camera mounted on front of acombine harvester cab fordetection of its forward path. After filtration and optimization of the image، Hough transform was utilized for extracting the path as a line. Then lateral and angular deviations from the path obtained as main inputs for the automatic guidance system of combine harvester. The algorithm was tested in 7 lateral and 7 angular deviation treatments with three replications. The tests were carried out under different conditions of day light. Comparison of the values extraction by the image processing algorithm with angles and deviations measured manually showed that the algorithm is able to detect the designed line with an accuracy of 98%.