抄録
Penumbral imaging technique can be applied to highly penetrating radiations such that of as neutrons. In penumbral imaging, the source image can be recovered from its penumbral image by deconvolution. The method is an efficient imaging technique for fast ignition research. However, the γ rays produced by the fast-heating laser pollute the penumbral image as noise. Conventional deconvolution methods like the Wiener filter cannot obtain a clear reconstructed image from noisy penumbral image. In this paper, we proposea new reconstruction method by principal component analysis (PCA). The method can efficiently remove the noise by “training” images obtained from other experiments. We used the (2D)2PCA method as a noise reduction method, which is one of the PCA methods. The efficacy of the proposed method is demonstrated by computer simulation.
