运用由粗至精策略的车牌识别系统

his paper deals with problematic from field of image processing, artificial intelligence and machine vision in construction of a license plate recognition system. This issue includes mathematical principles and algorithms. The significant of this system is its robustness for Chinese license plate detection and recognition. Authors represent a coarse-to-fine strategy: license plate region's rough detection and accurate localization of the region of interest (ROI). For the Optical Character Recognition (OCR) task, a Probabilistic Neural Network (PNN) is trained to identify Chinese and alphanumberic characters. It turns out a high accuracy is achieved in experiments and the system has a good application prospect.