Image inpainting in micrometeorological analysis

Digital image inpainting is the process by which corrupted or defective areas in an image are systematically corrected. New digital image inpainting techniques have been developed in recent years, leading to numerous successful applications, particularly in the area of image restoration. We propose a new image inpainting algorithm based on wavelet sparse representation, and extend its applicability as a new approach for gap-filling in micrometeorological data. Our approach consists of treating the incomplete data set as a structured image that has a sparse representation in the wavelet domain. Therefore, an ℓ₁ minimization problem is formulated in order to characterize the sparsest solution associated with the complete data set. A numerical experimentation on a real micrometeorological data set is conducted, demonstrating the effectiveness of the proposed approach.