基于移动窗口定位和复合神经网络的车牌识别算法

Vehicle License Plate Recognition Algorithm Based on Moving Target Location and Compound Neural Network

针对一般车牌定位算法存在多个车牌识别区域,以及单个人工神经网络识别复杂、运算量大、不容易收敛等问题,提出采用移动窗口定位算法和复合BP神经网络模型,可以较好的解决以上问题,在实际应用中取得了预期效果。

In view of the multi - location areas problem of normal vehicle license plate location algorithms and complexity and great operation and un -convergence of single neural network,this article brings up a LPR system using moving target location algorithm and compound neural network model, which can succeed in solving those problems. The application of our LPR system in fieldwork has a good result.