Research Of Inverse Halftoning For Scanned Image

Inverse halftoning of scanned image is valuable in the application of paper map digitalization and digital library. The thesis introduces the development and actuality of halftoning and inverse halftoning technology, and researches the ideal inverse halftoning and the present inverse halftone of scanned images. On the basis of analysis of the whole print-scan process and its experimental results, firstly, a model for CCD scanning noise is developed, then a comprehensive model for print-scan noise is developed too, lastly, an adaptive inverse halftoning algorithm based on the CCD model is presented. Compared to other inverse halftone algorithms, the proposed method, which considers the whole print-scan process and analyzes the sources of noise in theory, reaches a conclusion that the print noise mainly include halftoning noise with high-frequency characteristics and dot gain noise reducing contrast gradient, scanning noise which belong to Gaussian model can be subdivided into multiplicative and additive components, as well as a little impulse noise, moreover the model for noise is established. The proposed algorithm can remove halftoning noise efficiently, while scanning noise is suppressed fully. Finally, an evaluation is done, and software for scanned image inverse halftoning is developed to fulfill the requirement of paper map digitalization.