Deconvolution based on the Wiener–Lucy chain algorithm: an approach to recover local information losses in the deconvolution procedure

A Wiener–Lucy (W-L) chain scheme, which is a combination of a Wiener filter in the spatial-frequency domain and a Richardson–Lucy algorithm in the image domain, is proposed as a new deconvolution procedure. The most serious problem of the deconvolution procedure is related to local information losses in its application domain. The W-L chain tends to recover the local information losses in both the frequency and the image domains. The effectiveness of the W-L chain is proved theoretically, with the scheme being shown to converge to the solution of a simple inverse filter without noise enhancement. The W-L chain is applied to an IR shift-and-add image, and it is shown that the higher-frequency components, which are lost in the Wiener filtering, are well recovered.