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

Abstract

In the present research 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 Ra over drill life. The proposed measurement system, using colour image processing and an artificial neural network, can detect corner wear of a drill accurately and predict the surface roughness Ra of the hole to be drilled with mean and maximum errors of 0.32μm and -1.00μm, respectively. The presence of a built-up edge does not influence the results.