A VISION SYSTEM FOR SURFACE-ROUGHNESS ASSESSMENT USING NEURAL NETWORKS

In this study we use machine vision to assess surface roughness of mac hined parts produced by the shaping and milling processes. Machine vis ion allows for the assessment of surface roughness without touching or scratching the surface, and provides the flexibility for inspecting p arts without fixing them in a precise position. The quantitative measu res of surface roughness are extracted in the spatial frequency domain using a two-dimensional Fourier transform. Two artificial neural netw orks, which take roughness features as the input, are developed to det ermine the surface roughness. The first network is for test parts plac ed in a fixed orientation, which minimises the deviation of roughness measures. The second network is for test parts placed in random orient ations, which gives maximum flexibility for inspection tasks. Experime ntal results have shown that the proposed roughness features and neura l networks are efficient and effective for automated classification of surface roughness.