|Table of Contents|

Improved Mean Shift Algorithm and Its Application


Research Field:
Publishing date:


Improved Mean Shift Algorithm and Its Application
ZHANG LiTAN Xin-lianZHOU Wen-huiLIU Ji-lin
1.School of Information Science and Engineering,Zhejiang University,Hangzhou 310027,China;2.School of Information Engineering,Zhengzhou University,Henan 450001,China
road segmentation mean shift algorithm image binarization shadow detection visual navigation
A road segmentation application based on the improved mean shift algorithm is presented.The improved mean shift algorithm projects the road image into a two-dimensional feature plane which takes the red and blue color component as coordinate axes.The whole road segmentation approach combines the improved mean shift algorithm,the scene variation information and the vehicle motion information between adjacent frames.It doesn’t need a simplified road model formed by many unsuitable assumptions and overcomes the shortcomings brought out by them when roads have branches.Compared with the segmentation results obtained by the raw mean shift and other methods,the proposed approach increases the segmentation accuracy and gives road portion for safe driving as much as possible.


[1] Boykov Y Y, JollyM P. In terac tive graph cuts for optim al boundary& reg ion segm entation o f objects in ND im ages [ A]. Proceedings o f the Inte rnational Conference on Computer V ision [ C] . Canada: Vancouve r, 2001. 105- 111.
[2] FukunagaK, H ostetle rL O. The estim ation of the g rad ient o f a density function, w ith applications in pattern recognition [ J]. IEEE Trans Inform Theory, 1975, IT - 21 ( 1): 32- 40.
[3] M eer P, Geo rgescu B. Edge detection w ith embedded confidence [ J]. IEEE Transactions on Patte rn Ana ly sis andM ach ine Inte lligence, 2001, 23: 1 351- 1 365.
[4] Com aniciu D, M eer P. M ean sh ift: a robust approach tow ard feature space analysis [ J]. IEEE Transactions on Pattern Ana ly sis andM achine Inte lligence, 2002, 24 ( 5): 603- 619.
[5] Johan S. V ision based autonom ous road fo llow ing fo r a w hee led ou tdoor robot [ Z]. Sweden: KTH Num er ica l Analysis and Computer Sc ience, 2005.
[6] 纪天明, 贺跃, 于同, 等. 智能车辆导航系统中的实时 道路检测[ J]. 计算机应用, 2005, 25( 12): 228- 230.
[7] 彭宁嵩, 杨杰, 刘志, 等. M ean-Sh ift跟踪算法中核 函数窗宽的自动选取[ J]. 软件学报, 2005, 16( 9): 1 542- 1 550.
[8] Tu rkM A, M orgentha lerD G, K eith D G, et a.l V ITS - A v is ion system fo r autonomous land vehicle nav igation [ J]. IEEE T ransactions on Pattern Analysis and M ach ine In tellig ence, 1988, 10 ( 3): 342- 361.
[9] A ishy Am er, Er ic Dubo is. Image segm entation by robust b inarization and fast m orpho log ica l edge de tection [ A]. Pro cV ision Interface [ C] . Montrea,l C anada: IEEE Press, 2000. 357- 364.


Last Update: 2012-12-05