摘要
针对传统算法难以克服天气多变等外界因素的干扰,且识别率低的不足,研究了智能监控中如何有效提高车牌识别率的问题。根据车牌字符图像的特点,提出了一种基于局部HOG特征的稀疏表达车牌识别算法。方法采用字符图像5个ROI的归一化HOG特征为基础建立特征向量,构建字典,并利用稀疏表达思想求解字符图像特征的稀疏系数,进而完成车牌识别。实验结果表明,与传统改进的BP神经网络法相比,该方法不仅有很高的正确识别率,而且对于有部分缺失的字符图像也有很高识别可靠性,具有很大的应用价值。
The traditional license plate recognition algorithm suffered from the changed weather and low proper recognition rate,which seriously degraded the reliability of the intelligent supervisory system.In order to improve the rate of recognition,making use of the features of license plate's character,this paper proposes an algorithm with sparse representation based on local HOG.After getting the normalized HOG vector of 5 ROI in character image,the feature-dictionary is constructed depending on the HOGs,and then the sparse coefficient of test image on the feature-dictionary is achieved with the spares representation theory.The experiment results demonstrate that this method has the superiority of reliability and stability compared with the traditional optimized BP Neural Network,even if the character image is incomplete.
出处
《计算机仿真》
CSCD
北大核心
2011年第4期367-369,407,共4页
Computer Simulation