回转库档案实时定位中基于鱼眼镜头图像的处理识别技术  

Processing and Recognition Technology Based on Fisheye Lens Image in Real-Time Positioning of Rotary Library Files

作　　者：成杰 叶文武 徐寅林[1] Cheng Jie;Ye Wenwu;Xu Yinlin(School of Physics and Technology,Nanjing Normal University,Nanjing 210023,China)

机构地区：[1]南京师范大学物理科学与技术学院

出　　处：《南京师范大学学报（工程技术版）》2019年第2期75-81,共7页Journal of Nanjing Normal University(Engineering and Technology Edition)

基　　金：江苏省研究生科研与实践创新计划项目(SJCX19_0186)

摘　　要：介绍了一种基于鱼眼镜头图像处理识别的回转库档案盘点技术,采用鱼眼镜头宽视角拍摄回转库中排列成行的带有数字标签的多本档案;利用图像矫正、分割技术处理得到标签上的每一个数字图;由构造的卷积神经网络识别出图像内容;以标签数字列在镜头中的位序实现档案定位.该技术实时性较好,识别准确且成本低.This paper introduces a rotary library file inventory technology based on fisheye lens image processing and recognition. A wide angle of view of the fisheye lens is used to capture multiple files with digital labels arranged in rows. Image correction and segmentation technology are used to process each label. Image content is recognized by constructed Convolutional Neural Network. The file position is realized by the bit sequence of the label number listed in the lens. The technology has good real-time performance,accurate identification and low cost.

关 键 词：鱼眼镜头 图像识别 卷积神经网络 档案识别 

分 类 号：TH711[机械工程—测试计量技术及仪器]