A THERMOGRAPH IMAGE EXTRACTION BASED ON COLOR FEATURES FOR INDUCTION MOTOR BEARING FAULT DIAGNOSIS MONITORING

In this study, an approach of extraction analysis for bearing fault diagnosis of rotating machinery based on thermogram investigation using color features is proposed in this paper. This research was proposed since condition monitoring and motor failures are great concern in industries. Early fault detection in machineries can avoid production lost and reducing maintenance costs. Therefore, in this work, infrared thermography (IRT) is used as a tool to detect early sign for bearing fault since this infrared thermography (IRT) is one of the most effective non-destructive testing techniques of condition monitoring and fault diagnostics. By using this infrared thermography (IRT) technology, the information of machine condition can be analyzed. In the present study, 300 thermal images are used in this simulation process whereby the images are classified into two classes namely nonnal and abnonnal. The first class consists of 150 images nonnal bearing while another 150 images denote abnonnal bearing class. SURF feature-based algorithm, RGB color space and active contour segmentation are employed in this paper in order to process and differentiate between nonnal and abnonnal bearing image by means of color features called statistical technique. The experiment results indicate that this statistical features of RGB color space able to distinguish the differences between nonnal and abnonnal features of bearing in machinery system.