Investigations of Infrared thermography and Image Processing for detecting tool conditions of Turning Processes

Abstract

Tool wear in machining could have considerable effect on the quality of the machined features in relation to
dimensional accuracy and surface roughness. Condition monitoring systems could provide the solution to monitor
conditions of the process and also, provide diagnostic and prognostic decisions regarding process faults. Because of the
small scale dimensions of the tool, one-dimensional sensors such as force signals might not be sufficiently sensitive to
measure changes in tool conditions. Image processing could provide a solution in the application of condition monitoring
system if implemented in a suitable way. In this study, novel evaluation techniques are proposed to predict tool health in a
turning process. Four types of image processing algorithms, namely temperature profile analysis; threshold, canny detector
and gradient techniques are used in order to evaluate their capabilities when used to process an infrared image data to detect
tool conditions. The results indicate that the infrared images, when integrated with appropriate image processing algorithms
can provide a powerful monitoring system for the prediction tool conditions of turning operations.