首页 | 本学科首页   官方微博 | 高级检索  
     检索      

移动端身份证号码识别算法实现
引用本文:朱 健,马汉杰,冯 杰,杨芷晴,韩煌达,王 健.移动端身份证号码识别算法实现[J].教育技术导刊,2019,18(12):19-21.
作者姓名:朱 健  马汉杰  冯 杰  杨芷晴  韩煌达  王 健
作者单位:浙江理工大学 信息学院,浙江 杭州 310018
基金项目:国家自然科学基金项目(61501402)
摘    要:为实现多场景下快速便捷的身份证号码提取,提出一种基于移动端的身份证号码识别算法。该算法利用搭载Android操作系统的移动设备,进行身份证号码区域定位和提取;首先利用身份证特殊的颜色分布,选取合适的通道分量;再通过图像阈值分割、噪声处理以及形态学处理,将身份证号码图像二值化;同时针对拍摄过程中可能出现的图像倾斜情况,通过图像旋转和投影法进行水平矫正;然后通过投影法将单个字符提取出来;最后使用卷积神经网络进行图像识别。通过MIX手机测试表明,单次身份证号码图像识别时间约为156ms,准确率约为99.1%,能够满足鲁棒性和实时性要求。

关 键 词:移动设备  轮廓提取  倾斜校正  卷积神经网络  
收稿时间:2019-03-18

Implementation of ID Card Number Recognition Algorithms Based on Mobile Terminal
ZHU Jian,MA Han-jie,FENG Jie,YANG Zi-qing,HANG Huang-da,WANG Jian.Implementation of ID Card Number Recognition Algorithms Based on Mobile Terminal[J].Introduction of Educational Technology,2019,18(12):19-21.
Authors:ZHU Jian  MA Han-jie  FENG Jie  YANG Zi-qing  HANG Huang-da  WANG Jian
Institution:School of Information Science and Technology, Zhejiang Sci-Tech University, Hangzhou 310018,China
Abstract:algorithm uses mobile devices with Android operating system to locate and extract the ID number area. Firstly, it uses the special color distribution of the ID card to select the appropriate channel components. Then, it binarizes the ID number image through image threshold segmentation, noise processing and morphological processing. At the same time, it aims at the possible image tilt during the shooting process. Horizontal rectification is carried out by image rotation and projection method; then single character is extracted by projection method; finally, image recognition is carried out by convolution neural network. The MIX mobile phone test shows that the recognition time of single ID card number image is about 156 ms, and the accuracy is about 99.1%, it can meet the requirements of robustness and real-time.
Keywords:mobile devices  contour extraction  skew correction  convolutional neural network  
点击此处可从《教育技术导刊》浏览原始摘要信息
点击此处可从《教育技术导刊》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号