释义 |
constrained least squares restoration =HF+N, where G,F and N are the vectors ofm,m×1 dimensions, which represent the degeneration image, the ideal image and the noise separately, and H is the degeneration matrix ofm,m×m,mdimensions.This method is to find an estimation F,making the criterion function JF=11QF11平方+α11G-HF11平方-11N11平方minimum, where Q is the linear operator of F, α is the Lagrange multiplier.The solution is F=HH+γQQ-1次方HG, where γ=1/2, In general, an iterativep rocedure is used to adjust the solution until the constrained condition 11G-HF11平方=11N11平方 is satisfied.;一种约束复原方法。设退化模型的向量代数结构形式为G=HF+N,其中G,F和N为m,m×1维向量,分别代表退化图象、理想图象和噪声,H为m,m×m,m维退化矩阵。该方法旨在寻求一个估计F,使准则函数JF=11QF11平方+α11G-HF11平方-11N11平方为最小,式中Q为F的线性算子,α为拉格朗日乘子。解得 F=HH+γQQ-1次方HG,其中γ=1/2。通常,迭代过程调整解直到满足约束条件11G-HF11平方=11N11平方为止。 |