一种基于纹理和颜色目标跟踪方法.pdf

28 1 2011 1 : 1006- 9348( 2011) 01- 0273- 04 李 巍, 赵英凯,钱厚亮 (, 210009) : , ( Camshift) H, H (V ), , , HSV H, ( LBP), LBP LBP, LBP , Camshift, , , , , , : ; ; ; :TP391. 4 : CamshiftTrackingAlgorithm Based on LBP Feature LIW ei, ZH O Y ing- ka,i Q I N Hou- liang ( utomation and ElectricalEngineering college, Nanjing University ofTechnology, N anjing Jiangsu 210009, China) ABSTRACT: TraditionalCamshift is a tracking algorithm based onH channel and can be disturbed by the illum ina- tion easily. The paper proposes a tracking algorithm based on LocalBinary Pattern ( LBP) featurewhich is combined with color feature. In the process of object tracking, H channel is separated from theHSV space and is converted to LBP mi age at first. Then Camshift operates on a back- projection mi age produced from object histogram mode.l The expermi ent result show s that the tracking algorithm proposed can track object stably and in real- tmi e, regardless of the existence of covering, changed appearance and illumination changing. KEYW RDS: Object tracking; Localbinary pattern( LBP); H istogram; Camshift algorithm [2, 3] 1 , Camshift( Continuously daptive M ean Shift [ 4] ) HSV H , H , , , , Camshift, , , Camshift , , , LBP, LBP Camshift,