摘要
车牌区域定位是车牌识别中的核心技术,利用Adaboost分类器选择和组合haar特征,实现复杂背景下的车牌图像快速定位。Adaboost算法的核心思想是利用同一个训练集来训练不同的分类器,得到若干个弱分类器,然后把这些弱分类器构成一个强分类器。实验结果表明,该方法可以在不良天气、遮挡和非线性光照等复杂情况下,快速和有效地定位到图像中的车牌,且每分钟可以处理100帧以上的图像。
License plate location is important for license plate recognition system. A license plate location method is proposed based on statistical pattern recognition knowledge under clutter background, which selects and combines Haar feature using Adaboost classifier. The core idea of Adaboost is to train a different classifier with one training set, to get a number of weak classifiers, and to form a strong classifier with weak classifiers. Experimental results show that the proposed method is faster and more accurate than common license plate location algorithms in bad weather with occlusion and nonlinear illumination etc. The method detects more than 100 frames per minute.
出处
《辽宁工业大学学报（自然科学版）》
2016年第2期81-86,共6页
Journal of Liaoning University of Technology(Natural Science Edition)
基金
国家自然科学基金项目(61502216)
