摘要
针对利用Gabor小波进行目标识别的过程中参数序列过多的问题,提出一种新的参数组合选取原则.该方法利用Gabor小波参数特性,配合能量函数构成图像的特征点集合.同时提出了基于统计模型的自适应非线性滤波算法,以弥补基本统计模型下卡尔曼滤波算法的缺陷,配合Gabor小波特征点集合能够大大减少特征点序列过多所造成的算法耗时问题.仿真和实验表明,该算法可广泛应用于各种图像目标识别与跟踪的系统中,并具有良好的图像目标识别能力.
To improve too many parameters of Gabor wavelet, a new algorithm for choosing target characteristic based on analysis the parameters of Gabor wavelet sufficiently is introduced. A feature points collection is strueted according to figure parameters with energy function. Adaptive nonlinear filter is proposed to overcome the trap of traditional statistical model Kalman filter, which can reduce much time on calculating. The simulation and experiment show that it has well capacity of recognizing figure target and can be applied to all vision tracking system.
出处
《控制与决策》
EI
CSCD
北大核心
2008年第2期162-166,共5页
Control and Decision
基金
哈尔滨市留学回国人员基金项目(2004AFLXJ009)