Orthogonal l1 technique for image restoration

A novel approach to the restoration of an image is considered when both the point spread function and the blurred image suffer from errors. The usual approaches such as l₁ and l₂ are found to be inappropriate. A robust method, the orthogonal l₁ approach, is suggested for the restoration of images corrupted by errors in the point spread function and measurement. Since the orthogonal l₁ error function has several local minima, the use of the Solis and Wets random optimization algorithm is proposed. Simulation results are presented to illustrate the merit of the approach