Curved Lane Detection using Improved Hough Transform and CLAHE in a Multi-Channel ROI

Abstract

With the increase in advent of intelligent vehicle technology, LDWS technique is actively developing to provide assistance to drivers in generating warning signal in case of lane departure. In this paper, the input image is divided into four regions of interest. Then the CLAHE is applied on the given ROI for better image contrast and removal of noise using gradient filter. Improved Hough Transform is applied on the given image for the detection of curved and straight lanes. After detecting the lanes, they are colored with green color so that they are clearly visible. The proposed algorithm is analysed based on the parameters like Accuracy, BER, Fmeasure, and Sensitivity. The analysis shows that this algorithm provides efficient and effective performance compared to the existing algorithm.