Research Of Character Recognition Technology In Locomotive Components Detection

As the rapid development of society, various types of locomotives play a more andmore important role in people’s daily life and give people a lot of convenience. The safeand intelligent scheduling management of locomotives has become a vitally important task.Locomotive components must been detected at regular intervals to ensure the safe oflocomotive. As the development of computer science, character recognition technology ismore and more widely applied to detect locomotive components. In this thesis, twoalgorithms about character recognition are proposed and are used to detect locomotivecomponent. The main research works are as follows:(1) Summarized the current research status of character recognition technology and itsapplication status in the detection of locomotive components.(2) Analysed the characteristics of thehigh-speed rail catenary rod number. Then ashape context algorithm, based on principal component analysis, is proposed for rodnumber recognition. By using minimum bounding rectangle, there is rotation invariance inthis algorithm. Experiment results show that the algorithm can recognize the catenary rodnumber better.(3) Analyseddifficulty of train license plate recognition firstly. Then a templatematching algorithm is proposed for it. This algorithm includes extracting the characterregion, segmenting and recognizing the character. Experiment results show that thealgorithm has good recognition results.