A Hybrid Image Binarization Technique for Enhancement of Degraded Document Images

Image processing is a process of analysis and manipulation of digital images, which improves the quality of an image. Image binarization is a pre-processing step which improves the quality of the image. Image binarization is used to improve the image quality and it segment the pixel values into two categories; i.e. black pixel as foreground and white pixel as background. In this research work, binarization techniques are applied to improve the quality of the degraded document images. Different types of image binarization methods used in this work are Otsu, Nick, Niblack and Sauvola. In addition to this, a new hybrid binarization method is proposed which is a combination of local and global method. Performance factors used are PSNR (Peak Signal to Noise Ratio) and NRM (Negative Rate Metric). From the results, it is observed that the proposed hybrid binarization has produced good results than existing methods.