一种改进AdaBoost算法的车牌检测方法

Method for license plate detection based on improved AdaBoost algorithm

针对传统AdaBoost算法的不足,分析了分类器训练耗时和训练过程中容易出现样本权重扭曲的问题,并提出了解决这一问题的有效方法。新方法主要对特征值和排序结果进行缓存以及对样本权重的更新规则进行适当调整。实验结果表明,使用该方法训练级联车牌检测器能较好地解决传统AdaBoost算法中所出现的权重扭曲及训练时间长的问题,在提高检测率的同时训练时间缩短了50%左右。

Focusing on the disadvantages of classical AdaBoost algorithm, this paper mainly analyzed the issue that the training time for classifiers was time-consuming and in training process the sample weights were easily distorted and a new method was advanced to avoid the problems. The new method was to buffer the computational results of sorted feature values and regulate the updated rules of sample weights appropriately. As a result, using the method to train a cascade license plate, the experimental results show that the new method does not lead to the issue of weight distortion and time consuming like classical AdaBoost often does, and moreover, the training time is shortened to 50 percent with a high detection rate.