High speed machined surface roughness measurement
Surface roughness monitoring techniques using non-contact methods based on computer vision technology are becoming popular and cost effective. An evolvable hardware configuration using reconfigurable Xilinx Virtex FPGA xcv1000 architecture with capability to compensate for poor illumination environment is proposed in this paper. In this work one of the important parameters of machined components namely surface roughness Ra is estimated with greater degree of accuracy using a new non-contact method of surface roughness measurement. The proposed method uses a machine vision system to capture the image of the surface, which is illuminated by two light sources at an inclined angle. Effective processing of the captured image requires computing architectures that are less complicated, highly flexible and further cost-effective. The experimental results and results obtained from trained artificial neural network are indicating that the proposed technique can be used to evaluate the roughness of the machined surfaces.
Keywords: Surface Roughness, Evolvable hardware, Genetic algorithm, Regression analysis.