抄録
In this paper, an estimation was made of the image quality of the radiographic imaging system, the joint between probability matrix and frequency analysis theory was made clear with mediating the idea of the whitening noise filter. The detectivity d' of the ideal detector is defined as, [numerical formula] Where S_1 denotes the signal matrix, Φ^<-1>_<mn> the variance and co-variance inverse matrix, S (u), the signal spectrum and W(u), the Wiener spectrum of noise. The next fomula used was, Φ^<-1>_<mn>=G^TG Where G is the triangular matrix of the whitening noise filter. G corresponds to √<W (u)^<-1>> and GS_1 to the convolution between the whitening noise filter function on the spatial area and the signal distribution. Thus, Φ_<nn> can be represented by a non singular and symmetric matrix whose normal diagonal elements are all same, and G can be represented by a triangular matrix whose normal diagonal elements are all same. This proved that the noise process of the radiographic imaging system is a necessary ergodic.
