Machine Vision based Surface Roughness measurement with Evolvable Hardware Filter

Surface roughness is an important parameter for machined components. Different techniques for surface roughness measurement include the contact and non-contact schemes. The stylus instrument based approach is a contact method but suffers from the drawback that, it may not represent the real characteristics of a surface. The optical measuring methods are alternative to stylus method, but, they are also sensitive to lighting and noise. In this paper, as an alternative method, Evolvable Hardware (EHW) architecture is proposed to filter the noise present in the machine vision grabbed image. The scheme is proposed in such a way as to preserve more image features with less computation effort. The algorithms presented in this paper shall form the basic frame for EHW based implementation of on-line adaptive noise removal systems. The proposed EHW approach provides a better performance when compared to the traditional adaptive filter techniques. Also, the EHW filter provides an improved PSNR and the surface roughness can be estimated with greater accuracy.