基于深度卷积神经网络的模糊图像重建系统设计  

Design of fuzzy image reconstruction system based on deep convolutional neural network

作　　者：张玉霞 王超 刘艳辉 张志刚 ZHANG Yuxia;WANG Chao;LIU Yanhui;ZHANG Zhigang(School of Information Engineering,Handan University,Handan 056005,China;Laboratory and Training Center of Electronic Information Engineering,Handan University,Handan 056005,China)

机构地区：[1]邯郸学院信息工程学院,河北邯郸056005 [2]邯郸学院电子信息工程实验与实训中心,河北邯郸056005

出　　处：《现代电子技术》2021年第15期81-84,共4页Modern Electronics Technique

基　　金：邯郸市教育科学研究“十三五”规划项目:基于虚拟现实技术构建虚拟双导师的实践教学研究(1926029);邯郸学院校级项目:基于日志大数据邯郸学院校园网智慧运维的研究与实现(XZ2019205);邯郸学院校级项目:基于大数据的邯郸学院教学质量保障体系研究(J201903);邯郸学院校级项目:基于云计算的中小型门户网站群构建的研究(14205)。

摘　　要：传统的模糊图像重建系统重建图像清晰度较低、图像重建耗时较长,为此基于深度卷积神经网络设计了一种新的模糊图像重建系统,系统设计分为硬件设计与软件设计两部分。根据系统的基础状况划分硬件系统处理区域,并不断调整区域内部信息,优化获取的图像信息,过滤错误代码数据,调节数据参数,加大数据整合的力度,实现对数据硬件的结构性处理,以硬件设计的元件数据为基础作为软件数据的操作前提实现软件系统设计,综合数据算法合理分析图像在传输的同时可能产生的传输信息缺失问题,并查找缺失原因,在提升传输速率的同时提高图像传输的完整度,由此实现对模糊图像重建系统的设计。实验结果表明,基于深度卷积神经网络的模糊图像重建系统能够有效提高重建图像的清晰度,缩短重建时间。The traditional fuzzy image reconstruction system has lower image sharpness and longer image reconstruction time.Therefore,a new fuzzy image reconstruction system is designed based on deep convolutional neural network.The system is composed of hardware and software.The processing area of the hardware system is sectioned according to the basic condition of the system.The internal information of the area is constantly adjusted,the acquired image information is optimized,the wrong code data is filtered,the data parameters are adjusted,and the intensity of data integration is increased,so as to realize the structural processing of the data hardware.The design of software system is realized by taking the component data of hardware design as the operation premise of software data.The data and algorithm are integrated to reasonably analyze the transmission information loss that may occur while the image is being transmitted.In addition,the cause of the loss is found out to improve the image integrity while improving the transmission rate.On the basis of the above steps,the design of the fuzzy image reconstruction system is achieved.An experiment was performed.The experimental results show that the fuzzy image reconstruction system based on deep convolutional neural network can effectively improve the sharpness of the reconstructed image and shorten the reconstruction time.

关 键 词：深度卷积神经网络 模糊图像 图像重建 重建系统 神经网络算法 图像分析 模糊重建 

分 类 号：TN911.73-34[电子电信—通信与信息系统] TP753[电子电信—信息与通信工程]