Multispectral Image Restoration of Historical Document Images

Culture is preserved through various documents which is a part of the civilization and heritage. Due to extinction and single document copies available for the future generations about the ancient scripts, the archiving of these documents in the digital process is the solution for these problems. In this paper, the aim is to restore the historical document from tears, stains and poor visibility using multispectral imaging tool. The proposed system consists of image fusion, binarization using niblack algorithm and restoration using inpainting process using multispectral imaging tool. The binarization process is done using adaptive thresholding niblack algorithm. The inpainting process fills the missing holes, degraded areas and stains and restores the image. Thus the image can be restored from its degraded condition and is enhanced to remove noise, tears and ink bleed.