【摘要】 提出一种自适应空域递归快速最小二乘算法,此算法直接从包含在某一滑动数据窗中的图象灰度数据开始做参数估计,数据窗每次滑动更新一行数据,参数估计量也随之迅速地改变,因而能用于非平稳图象。此算法的运算量为18mN^1/2+4 2m^1/2NMADPR（每次递归的乘/除法次数）,m,N分别指模型参数个数和包含在数据窗中的数据个数。更多还原

【Abstract】 This paper presents a new fast spatially recursive algorithm for general two—dimensional adaptive LS finite impulse response （FIR） filtering and autoregressive model fitting of a sliding windowed image,the new scheme requires roughly 18m N^1/2+4（2m）^1/2 multiplications/divisions to recursively update the estimates of the parameters,where N and m are the numbers of data in the window and the estimated parameters respectively.更多还原