PCBN Tool Performance Evaluation Based on Image Analysis of Machining Surface Texture

Abstract

The Polycrystalline Cubic Boron Nitride (PCBN) cutting tools has have been developed for high speed machining in modern automation manufacture. The machining surface roughness is regarded as an important criterion to assess PCBN cutting tools performance. There are too many problems in conventional detection method. In order to solve that problem, we present a new way that is based on image analysis of machining surface texture to assess surface roughness. The new method is consisted of three steps. It captures surface texture image when machining is finished or pauses. Firstly, RGB histogram is adopted to analyze image pixel information. This means takes advantage of histogram technique and provides more pixel distribution information than gray histogram. Secondly, unsupervised texture segmentation is used based on resonance algorithm. Thirdly, a new estimation parameter E that is the density of surface contour peak is put forward to estimate machining surface roughness.