基于近红外图像匹配分割的涂改车牌识别算法

Recognition Algorithm for Falsified License Plate Based on Near Infrared Image Matching-segmentation

针对目前错综复杂的涂改车牌违规现象,提出一种利用近红外成像技术实现涂改车牌识别的方法。首先,论述了涂改干扰车牌识别的基本原理和特征;然后,在其可行性分析的基础上,设计出基于图像匹配分割的识别算法。实验结果表明所提算法能够实现涂改车牌的准确识别,具有较高的实际意义和价值。

Due to complicated irregularity phenomenon of falsified license plate, a recognition algorithm for falsified license plate based on near infrared imaging is introduced. Firstly, the principle and characteristic of falsification which interferes performance of license plate recognition （LPR） are presented. Then an image matching-segmentation-based algorithm is devised on the basis of feasibility analysis to recognize the falsified license plate. Experiment results show that falsified license plate recognition can be implemented by the proposed algorithm, which has higher veracity and more valuable than the other.