Automatic drill wear measurement using colour image processing and artificial neural network

The corner wear of drills is measured automatically in order to predict end of drill life, using hole quality as criterion. Drilling experiments show a strong correlation between the progress of maximum hole diameter and hole surface roughness R-a over drill life. The proposed measurement system uses colour image processing and an artificial neural network. It can detect the corner wear of a drill accurately and predict the surface roughness R-a of the hole to be drilled with mean and maximum errors of 0.32 mu m and -1.00 mu m, respectively. The presence of a built-up edge does not influence the measurement results.