DocumentCode :
469011
Title :
A robust road segmentation method
Author :
Zhang, Li ; Zhou, Wen-hui ; Liu, Ji-lin
Author_Institution :
Zhejiang Univ., Hangzhou
Volume :
2
fYear :
2007
fDate :
2-4 Nov. 2007
Firstpage :
912
Lastpage :
917
Abstract :
A robust method of road segmentation for Autonomous Land Vehicle (ALV) navigation system is presented. The main contribution of the present road segmentation method consists of an effective improvement on the mean shift algorithm dedicated to road segmentation and an extension to the Bayesian method due to its suffering from incorrectly predicted edge and the non-generalization from the sampled pixels to the unsampled pixels. The improved mean shift algorithm transforms the road image to get better feature representation which highlights the intrinsic characteristics of road images. Road images are segmented by fusing with the results gotten by the improved mean shift algorithm and the extended Bayesian method, the scene variation information between adjacent frames, and the vehicle motion information. The method needn ´t assume a simplified road model and overcomes the shortcomings brought out by it. The experimental results show that the method has good performance and increases the segmentation accuracy to an extent.
Keywords :
image fusion; image representation; image segmentation; mobile robots; autonomous land vehicle; extended Bayesian method; feature representation; mean shift algorithm; navigation system; road images; robust road segmentation method; scene variation information; vehicle motion information; Bayesian methods; Clustering algorithms; Image edge detection; Image segmentation; Layout; Navigation; Pattern analysis; Roads; Robustness; Wavelet analysis; Bayesian method; Road segmentation; homography; information fusion; mean shift algorithm; visual navigation;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Wavelet Analysis and Pattern Recognition, 2007. ICWAPR '07. International Conference on
Conference_Location :
Beijing
Print_ISBN :
978-1-4244-1065-1
Electronic_ISBN :
978-1-4244-1066-8
Type :
conf
DOI :
10.1109/ICWAPR.2007.4420799
Filename :
4420799
Link To Document :
بازگشت