A new flexible high-resolution vision sensor for tool condition monitoring

From a critical review of defect morphology and image analysis techniques from the literature it seems that a method to recognise any kind of defect and the algorithms to measure all wear types are not available. This article is divided into two main parts: (i) a possible exhaustive classification of defects in cutting inserts and (ii) the design of an automated sensor to recognise defects and to measure wear. The morphology characterisation has led to the definition of a limited number of classes and recognition criteria that occur for different types of cutting materials and working conditions for milling and turning operations. They represent the main requirements of recognition and measurement algorithms. The global logic flow for decision making is also provided. The sensor configuration is outlined with the necessary views and lighting devices. The identification of the worn out areas is performed by software segmentation to detect the texture differences between damaged and undamaged zones and has been tested on different types of carbide inserts. A resolution enhancement method is also proposed.