Lane detection is a key technique for intelligent vehicle driving. Aiming at the detection performance of existing lane detection algorithms, based on least square fitting, we propose a lane detection algorithm for structural road. The lane videos are gotten by the monocular camera installed in the car. Image preprocessing is applied to improve image contrast and then the image is segmented by improved Otsu. At last, the current lanes are extracted and equations are rebuilt by the least square fitting. The experiment results show that the proposed method has better accuracy and robustness compared to existing lane detection algorithms.
Lane Detection Based on Least Square Fitting
Advanced Materials Research ; 403-408 ; 4068-4072
2011-11-29
5 pages
Article (Journal)
Electronic Resource
English
A least-square-distance curve-fitting technique
TIBKAT | 1971
|Lane Geometry Reconstruction: Least Square Curve Fit Versus Kalman Filter
British Library Conference Proceedings | 1995
|Direct and specific least-square fitting of hyperbolae and ellipses
British Library Online Contents | 2004
|