Road Boundary Extraction Based on Blue Color Segmentation of Digital Images

Abstract

Road edge extraction is a  basic task in   computer vision applications in last   decades , including autonomous navigation, intelligent transportation systems, and driver assistance systems. This paper presents a method for extracting road edges from digital images which  it has a lot of environment objects, like stones, trees, and road besides, others.  This approach based on image segmentation of input image called  blue color segmentation. The proposed approach begins by preprocessing filtering, converting the input image into an appropriate color space, followed by isolating the blue-colored regions that represent the road boundaries or markings. Thresholding techniques are then applied to generate a Gradient binary mask that highlights the road edges. Morphological operations and edge detection filters are subsequently used to refine the extracted edges and remove noise. The performance of the proposed method is evaluated on a set of digital road images under varying lighting and environmental conditions. Experimental results demonstrate that blue color segmentation provides an effective and computationally efficient means of extracting road edges and overlapped on input image, achieving satisfactory accuracy compared to traditional edge detection methods and others, like textures it has complicated calculations. This proposed approach is simple to implement and suitable towards  real-time applications.