一种改进的BP算法及其在车牌识别中的应用

在车牌识别系统中,如何提高字符识别的速度和准确率是很关键的问题。传统的BP算法可以实现非线性函数的映射,经过有监督式的学习规则可以达到比较好的识别效果。但是BP算法识别速度慢,而且容易陷入局部最优。本文借助增量PID算法的思想,提出了一种改进的BP神经网络算法,可以根据识别误差自动调整学习速率,而且不容易陷入局部最优。文章最后给出了仿真结果和程序实现。

The Design of Developed BP Arithmetic and Its Application in Identifying The Number Plate of The Vehicle

How to improve the identifying speed and accuracy in the identifying system of the number plate is a key problem. The nonlinear function map can be realized with the classic BP arithmetic and the identifying result will be relatively good under the study rule of supervising type. But BP arithmetic has a low identifying speed and easy to encounter local optimization. In this article, an improved BP nerve net arithmetic was introduced. The study velocity can modulate automatically according to the identifying error, and not easy to encounter local optimization. At last, an emulate result and program realization is given.