M. Lazo and A. Sanfeliu (Eds.): CIARP 2005, LNCS 3773, pp. 786 – 793, 2005. © Springer-Verlag Berlin Heidelberg 2005 Tool Insert Wear Classification Using Statistical Descriptors and Neuronal Networks E. Alegre, R. Aláiz, J. Barreiro, and M. Viñuela Escuela de Ingenierías Industrial e Informática, Universidad de León, 24071, León, España {enrique.alegre, rocio.alaiz, joaquin.barreiro}@unileon.es mavilo92@hotmail.com Abstract. The goal of this work is to automatically determine the level of tool insert wear based on images acquired using a vision system. Experimental wear was carried out by machining AISI SAE 1045 and 4140 steel bars in a precision CNC lathe and using Sandvik inserts of tungsten carbide. A Pulnix PE2015 B/W with an optic composed by an industrial zoom 70 XL to 1.5X and a diffuse lighting system was used for acquisition. After images were pre-processed and wear area segmented, several patterns of the wear area were obtained using a set of descriptors based on statistical moments. Two sets of experiments were carried out, the first one considering two classes, low wear level and high wear level, respectively; the second one considering three classes. Performance of three classifiers was evaluated: Lp 2 , k-nearest neighbours and neural networks. Zernike and Legendre descriptors show the lowest error rates using a MLP neu- ronal network for classifying. 1 Introduction Measuring of wear in tools for machining has been in the scope of many studies. Depending on the method for acquiring values and their implementation, methods to wear measuring are classified in direct or indirect, and according to the monitoring in continuous and intermittent [1]. Direct methods measure change of actual parameters values as shape and location of the cutting edge [2] (optical methods: CCD cameras or optic fibber sensors), tool material volumetric loss, electrical resistance at the part-tool interface (voltage measuring of a specific conductive covering), part dimensions (dimensional measuring with optic devices or with micrometers, pneumatic, ultrasonic or electromagnetic transducers) or distance between tool and part. Indirect methods contrast the wear with process parameters, which are easier of measuring. However, the computational effort later on is bigger. Examples are cutting forces evaluation (effort measuring devices, sensors, piezoelectric plates or rings, bearings with force measuring, torque measuring, etc.), tool or tool-holder vibration (accelerometer), acoustic emissions (transducers integrated in the tool-holder or microphones), current or power consumption in the screw or motor (ammeters or dynamometers), temperature (thermocouples or pyrometers, colour reflectance or chip surface) or roughness of machined surface (mechanical or optical methods) [3].