基于深度学习的实时防汛图像可视化传输方法  

Visual transmission method of real-time flood control image based on deep learning

作　　者：李海亮 LI Hai-liang(Yinchuan Huangyangtan Flood Control Management Institute,Yinchuan 750021,China)

机构地区：[1]银川市黄羊滩防洪管理所,银川750021

出　　处：《信息技术》2023年第3期91-96,共6页Information Technology

摘　　要：针对接收端接收到的图像峰值信噪比偏小,导致图像传输质量差的问题,提出基于深度学习的实时防汛图像可视化传输方法。采用主成分分析技术,通过矩阵计算、特征变换、数据投影、色彩合成四个步骤,处理防汛图像,设计了图像传输信息保护步骤;运用Fast算法,小规模迭代处理可视化图像像素,重构可视化图像,实现了防汛图像可视化传输。实验结果表明,该方法算法简单,图像信息完整度高,具有较优的图像可视化传输质量。Based on the problem that the peak signal-to-noise ratio of the image received by the receiver is too small,resulting in poor image transmission quality,a real-time flood control image visual transmission method based on deep learning is proposed.Using principal component analysis technology,the flood control image is processed through four steps,including matrix calculation,feature transformation,data projection and color synthesis.The image transmission information protection steps are designed.The visual image pixels are processed iteratively in a small scale by using Fast algorithm,and the visual image is reconstructed to realize the visual transmission of flood control image.The experiment results show that the image information integrity is high and the algorithm is simple,and has better image visual transmission quality.

关 键 词：深度学习 实时防汛 防汛图像 可视化传输 迭代处理 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]