Polar Transformations Application in Characters Recognition System Based on Machine Vision

In many situations, we have to read characters on the object we are inspecting. Typically, the character text is horizontal, in this case, we can recognize the character easily, and the recognition methods are very mature. But, sometimes the characters distribute on the circle or ellipse, and the characters are distortion, which bring us huge trouble to recognize. This paper main about how to read the characters on ellipse that distortion, instructing polar transformation, rotate transformation, feature extraction, character segmentation and character recognition and MLP related algorithm. At the end of this paper, the results analysis are given.