Simple fuzzy technique combining sharpening and noise reduction for color images

The development of digital techniques for color image enhancement is very important because color pictures include information that is essential in many research and application fields. Color image enhancement, however, is not a trivial problem, because it typically deals with conflicting operations such as data sharpening and smoothing. The challenging goals consist in avoiding noise amplification, chroma distortion and excessive overshoot on object edges. This paper presents a novel technique for the enhancement of RGB images that is designed to address all these issues. The proposed approach consists in a two-stage recursive operator that adopts fuzzy processing in order to gradually perform smoothing and sharpening of noisy color data. In this approach, chroma distortion is avoided by choosing vector smoothing of RGB data and by applying the sharpening to the luminance channel only. Computer simulations show that the method combines simplicity and effectiveness. It can sharpen the image details and reduce the noise without generating false colors, whereas other techniques cannot. The method is easy to use too. The behavior of the operator is easily controlled by choosing a few parameter values.