摘要
该文在考虑图素失真满足二维联立自回归（ＳＡＲ）模型的条件下，推导建立了二维最小二乘（ＬＳ）阶递归快速算法。该算法将矩阵运算减到最小，且收敛速度保持与一维算法同样快。该算法不但可用于图象滤波，还可用于二维参数估计、二维系统辨识及模式识别等领域。文未，计算机仿真实验结果验证了该文得出的结论：信噪比提高２０ｄＢ，运算量是高斯法的二分之一至三分之一。
On the basis of the fact that two-dimensional image deformity meet 2D series autoregressive(SAR)model,the 2D order-recursive least-squares(LS)fast algorithm is presented in this paper. The calculations of matrix is decreased to the least so that the total operations are much decreased. And its convergence speed is as fast as that of 1D algorithm. This algorithm can be used not only to image filtering, but also to 2D parameters estimation,2D Systems identif-cation,and pattern recognition as well.In the end,computer simula-tion experiments’results are coincided with the conclusions obtained in the paper:the Signal to noise ratio(SNR)is increased by 20 dB,the operations are only the half or one third of Gaussican method’s.
出处
《南京理工大学学报》
EI
CAS
CSCD
1994年第6期71-75,共5页
Journal of Nanjing University of Science and Technology
基金
国家自然科学基金
关键词
自适应滤波
图象处理
最小二乘法
递归滤波法
adaptive filtering,image processing,least-square method,recurcive filtering method,algorithm analysis
