Blind deconvolution of bi-level images with successive filtering

Abstract

Many imaging systems are involved in the capture of bi-level objects such as documents and bar codes. However, due to optical aberrations and other degradations such as from motion, the resulting images are no longer bi-level. Restoration is therefore necessary to produce a nearly bi-level output, which is much more convenient for further processing including recognition and identification. In this work, we tackle the problem using successive filtering, where each step is formulated as a quadratic programming optimization problem. This has the properties of fast convergence and good numerical stability. Simulation results show that by integrating the computation in the imaging system, this method is able to restore weak signals.