采用视觉显著性和卷积网络的车牌定位算法

详细信息    查看全文 | 推荐本文 |

英文篇名：A License Plate Location Algorithm Using Visual Saliency and Convolutional Network 作者：刘华春 ; 侯向宁 英文作者：LIU Hua-chun;HOU Xiang-ning;School of Electronic Information and Computer Engineering,Engineering & Technical College of Chengdu University of Technology; 关键词：视觉显著性 ; 卷积神经网络 ; 车牌定位 ; 深度学习 ; 分类 ; 车牌识别 ; AlexNet网络 英文关键词：visual significance;;convolution neural network;;license plate positioning;;deep learning;;classification;;license plate recognition;;AlexNet network 中文刊名：WJFZ 英文刊名：Computer Technology and Development 机构：成都理工大学工程技术学院电子信息与计算机工程学院; 出版日期：2019-03-06 10:25 出版单位：计算机技术与发展 年：2019 期：v.29;No.266 基金：四川省教育科研重点项目(A22017003);; 四川省乐山市科技计划重点项目(16GZD050) 语种：中文; 页：WJFZ201906016 页数：5 CN：06 ISSN：61-1450/TP 分类号：80-84

摘要

针对传统的车牌定位方法存在对环境要求高,容易受干扰,鲁棒性不强等不足,提出将视觉显著性与卷积神经网络相结合应用于车牌定位,设计了一种车牌定位方法。该方法包含2个阶段,第1阶段为提取车牌候选区域,将视觉注意机制引入车牌定位过程,采用自下而上的模型,提取出车辆图像中具有显著特征的车牌区域。采用视觉显著性算法,实现车牌候选区域的快速定位和提取,避免了在车辆图像的各个区域进行扫描。第2阶段为车牌识别,将深度卷积神经网络应用于车牌区域识别,实现候选区域中非车牌和车牌的准确分类。实验结果表明,该方法性能优异,大幅降低了车牌区域漏检率,鲁棒性强,定位准确率比数字图像方法提高约5%。
        In view of the shortcomings of traditional license plate location methods,such as high environmental requirements,easy interference and poor robustness,we design a license plate location method by combining visual saliency and convolutional neural network. The method consists of two stages. The first stage is to extract the license plate candidate area. The visual attention mechanism is introduced into the license plate location process,and the bottom-up model is used to extract the license plate area with significant features in the vehicle image. The visual saliency algorithm is used to realize the rapid positioning and extraction of the license plate candidate area,thus avoiding scanning in various areas of the vehicle image. The second stage is license plate recognition. The deep convolutional neural network is applied to the identification of the license plate area for accurate classification of non-license plates and license plates in the candidate area. The experiment shows that the proposed method has excellent performance,sharply dropping the missed detection rate of the license plate area,with strong robustness,and its accuracy of positioning is improved about 5% compared with the digital image method.

引文

[1] SIVARAMAN S,TRIVEDI M M.Integrated lane and vehicle detection,localization,and tracking:a synergistic approach[J].IEEE Transactions on Intelligent Transportation Systems,2013,14(2):906-917.
    [2] TEOH S S,BRAUNL T.Symmetry-based monocular vehicle detection system[J].Machine Vision and Applications,2012,23(5):831-842.
    [3] 曾丽华,李超,熊璋.基于边缘与颜色信息的车牌精确定位算法[J].北京航空航天大学学报,2007,33(9):1112-1116.
    [4] 黄艳国,赵书玲,许伦辉.基于纹理特征和颜色匹配的车牌定位方法[J].微电子学与计算机,2011,28(9):123-126.
    [5] GEIGER A,LENZ P,STILLER C,et al.Vision meets robotics:the KITTI dataset[J].International Journal of Robotics Research,2013,32(11):1231-1237.
    [6] SIAGIAN C,ITTI L.Rapid biologically-inspired scene classification using features shared with visual attention[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2007,29(2):300-312.
    [7] VIKRAM T N,TSCHEREPANOW M,WREDE B.A saliency map based on sampling all image into random rectangular regions of interest[J].Pattern Recognition,2012,45(9):3114-3124.
    [8] KRIZHEVSKY A,SUTSKEVER I,HINTON G E.ImageNet classification with deep convolutional neural networks[J].Advances in Neural Information Processing Systems,2012,25(2):1097-1105.
    [9] SHI Liang,WANG Jinqiao,XU Lei,et al.Context saliency based image summarization[C]//Proceedings of the 2009 IEEE international conference on multimedia and expo.New York,NY,USA:IEEE,2009:270-273.
    [10] ITTI L,KOCH C,NIEBUR E.A model of saliency based visual attention for rapid scene analysis[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1998,20(11):1254-1259.
    [11] SHAN Ying,SAWHNEY H S,KUMAR R.Unsupervised learning of discriminative edge measures for vehicle matching between non-overlapping cameras[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2008,30(4):700-711.
    [12] 傅鹏,谢世朋.基于级联卷积神经网络的车牌定位[J].计算机技术与发展,2018,28(1):134-137.
    [13] 李庆庆,张燕平.基于模糊边缘检测算法的车牌定位[J].计算机技术与发展,2006,16(12):7-8.
    [14] GUO Y,HSU S,SAWHNEY H S,et al.Robust object matching for persistent tracking with heterogeneous features[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2007,29(5):824-839.