A method for detecting and identifying thin and vague roads in aerial images is proposed. The method first smooths an image by using a Gaussian filter, then uses a fractional differential operator to enhance the image for sharpening roads, then applies a one-pass ridge edge detection algorithm to roughly detect the roads, and finally utilizes a number of post functions to accurately identify roads. For each detected point in an aerial road image, the new ridge detection algorithm detects if it is a candidate for the ridge edge points by searching through four different directions. After that, the extracted segments of lines/curves are smoothed and their gaps are linked according to preset thresholds of lengths and directions, and the noisy lines are removed based on the rules of the curve length and shape information. If roads are thick, the image can be shrunk to a road with a width less than six pixels, then the detection result for the course resolution image is mapped into the original image to accurately re-identify the road. In experiments, by comparison to traditional methods, the studied method can have better detection results for thin and vague roads, which are difficult to detect with traditional algorithms. c The Authors. Published by SPIE under a Creative Commons Attribution 3.0 Unported License. Distribution or reproduction of this work in whole or in part requires full attribution of the original publication, including its DOI.
CITATION STYLE
Yang, X., & Wang, W. (2014). Road identification in aerial images on fractional differential and one-pass ridge edge detection. Journal of Applied Remote Sensing, 8(1), 083597. https://doi.org/10.1117/1.jrs.8.083597
Mendeley helps you to discover research relevant for your work.