Research And Implementation Of Bearing Workpiece Characters Recognition System Based On Machine Vision

Railway transportation is characterized by large transportation capacity, fast transportation speed, long load distance, low transportation cost and less influenced by weather condition. The development of railway transportation is an important index to measure a nation’s economic development and social progress. According to the requirements of China’s related railway management departments, the double-row tapered roller bearing used in passenger and freight trains adopt its own sole workpiece characters to trace the whole range quality in all links including manufacturing, inspection, operation and maintenance. However, applying human eyes to recognize workpiece characters on the bearing production line has bad recognition accuracy and low work efficiency.This paper put forward a bearing workpiece characters recognition system based on the machine vision. This system is composed of industrial camera, telecentric lens, optical filter, LED light source and IPC etc. The recognition procedure consists of image acquisition, image preprocessing, image locating, image segmentation, character identification and workpiece characters output. The image acquisition process is responsible for obtaining the original monochrome image for system. The image preprocessing takes charge of detecting and extracting the ROI and then storing it. The image locating process accounts for locating character position precisely. The image segmentation process is in charge of segmenting various characters after locating. The character recognition is the most important process in bearing workpiece characters recognition, because its identify results directly determines the fair or foul of bearing workpiece characters recognition system. This paper adopts artificial neural network and SVM calculation to recognize the bearing workpiece characters effectively.The bearing workpiece characters recognition is a system with high real-time, better robustness and accuracy. The scheme put forward in this paper satisfied the above requirements to the maximum extent. Under the laboratory environment, eight double-row tapered roller bearings are selected to conduct comprehensive environment testing. The result showed that the scheme put forward in this paper has fast system response speed, high character recognition rate (95%), stable system operation(no breakdown within48hours’constant operation), which also demonstrates its relatively bright application prospect.