Random valued impulse denoising using robust direction based detector

Impulse noise is a spark that affects the contents of digital images. It also occurs during image acquisition, transmission due to malfunctioning of pixel elements in the camera sensors, faulty memory locations, timing errors in analog to digital conversion and bit errors during transmission. Removal of impulse noise involves detection followed by filtering mechanism. There are two types of impulse noise namely Salt and pepper impulse noise and Random valued impulse noise. Among them, random valued noise is difficult to detect. Among existing detection methods, most of them provide good results for low noise density images. Few methods deals with high density noise and provide less visual quality. In this paper, we propose a new detector; Robust Direction based Detector (RDD) which is based on ROR (Robust Outlyingness Ratio) and Standard Deviation (SD). Based on the ROR, the pixels are classified into noisy and noise free pixels. Then, different decision rules are adopted on the corrupted pixels again based on standard deviation. Then the corrupted pixels are filtered by using Edge Strength Interpolation Algorithm.