Skip to Main Content
This paper presents a new algorithm based on subspace and extended constrained least squares (CLS) for multichannel blind image restoration. Subspace-based methods are good solutions for multichannel blind image restoration, but always sensitive to noise. An adaptive singular value decomposition (SVD)-based denoising technique is used to improve the degraded images quality, and a priori distribution of blurs derived from subspace approach is added to the extended CLS multichannel blind image restoration method, which makes a new regularization function to recover images. Experimental results show that, this algorithm is convergent and works well even on very noisy circumstances.