Multispectral Images Deblurring via Interchannel Correlation Relationship
Yuetian Shi, Nan Wang, Fanchao Yang, Geng Zhang, Siyuan Li, Xuebin Liu · 2021
Image blur is very common in multispectral images, especially out-of-focus blur when the multispectral cameras are well focused on a specific channel. Therefore, it is important to recover the latent clear images from the blurred image. Many researchers have proposed a large number of methods for deblurring multispectral images. These algorithms assume that the relationship between channels is consistent. In fact, the relationship between different channels in multispectral images suffers from different blurring. Therefore it needs different processing for different channels. We propose a joint spatial and spectral filtering algorithm for multispectral images deblurring. We have introduced an adaptive Gaussian distribution as a constraint relationship between channels, which can effectively deblur the multispectral images. Extensive experiments show that our method achieves state-of-the-art results.