Image Processing and Diagnostic Technology for Macroscopic Characteristics of Electrical Corrosion of Railway Freight Car Bearings

In this paper, aiming at the diagnosis of electrical erosion fault of railway freight car bearings, an image processing diagnosis method combining macro-morphology, and bearing temperature rise criterion is proposed. The core of the research is to realize the efficient and accurate diagnosis of electrical erosion fault through macro-image recognition technology and combined evidence verification. In this study, the macro image of the bearing surface is obtained at first by a high-resolution camera while combined with the on-line temperature rise data of the bearings, professional image processing software and algorithms are applied for subsequent processing and analysis. Micro-topography images are used as physical evidence to verify the accuracy of AI algorithm, instead of directly identifying objects. Bearing samples covering various types of electrical erosion were selected in the experiment, and the practicability of the method was verified by comparing the macro-morphology, profile characteristics and temperature rise. The proposed diagnosis method is excellent in recognition accuracy and processing speed. The image processing and diagnosis technology of electrical corrosion of railway bearings, which is successfully developed in this study, provides a new and effective solution for efficient identification of electrical corrosion faults.