一种基于MSER及CNN的车牌文字定位新方法  

A NEW METHOD OF LICENSE PLATE TEXT LOCATION BASED ON MSER AND CNN

作　　者：顾恭 Gu Gong(College of Computer,Beijing University of Technology,Beijing 100124,China)

机构地区：[1]北京工业大学计算机学院,北京100124

出　　处：《计算机应用与软件》2021年第8期206-213,279,共9页Computer Applications and Software

摘　　要：车牌定位是车辆信息识别中的关键和基础。为解决在复杂无约束场景下存在的车牌定位精度不高,噪点和干扰因素较强等问题,提出一种基于最大稳定极值区域和卷积神经网络的车牌精准定位新方法。利用最大稳定极值区域找出车辆图像中二值化参数较为稳定的子图像区域;根据车牌的先验知识,滤掉明显不符合车牌字符特征的子图像区域;对保留下的子图像进行相应的启发式搜索和卷积神经网络识别,找出确切的多个车牌字符位置;通过滑动窗口和卷积神经网络搜索到车牌的始末位置,从而在复杂自然环境下完整获得车辆的牌照区域。实验结果表明,该算法在各类复杂场景下受到的环境影响小,鲁棒性强,定位准确率高。License plate text location is a key and fundamental step in license plate recognition(LPR).A new method of license plate location based on maximally stable extremal regions(MSER)and convolutional neural network(CNN)is proposed to solve the problems of low accuracy,strong noise and interference factors in complex scenes.The method took advantage of MSER to find more stable binary sub-images;according to prior knowledge of license plate,some sub-images that do not conform to features of license plate characteristics obviously were filtered out;it made use of heuristic knowledge and CNN to recognize true license plate character and find out their location in the complete image;sliding window of CNN was utilized to search for beginning and ending positions of a license plate.The experimental results indicate that the proposed algorithm has strong robustness,harder to impact by complex environment and improves accuracy of license plate text location.

关 键 词：车牌定位 最大稳定极值区域 卷积神经网络 滑动窗口算法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]