移动通信2024，Vol.48Issue(11)：57-62,6.DOI:10.3969/j.issn.1006-1010.20240927-0001
基于深度学习的海洋色散信道信号处理研究
Signal Processing for Marine Dispersive Channels Based on Deep Learning
摘要
Abstract
In marine channels formed by non-uniform media,electromagnetic wave dispersion results in distinct harmonic characteristics in the frequency domain.However,traditional spectral line detection methods typically focus on identifying individual spectral lines without a global analysis of harmonic characteristics,and the dispersion of spectral line energy leads to significant performance degradation of these traditional methods.To overcome these challenges,a convolutional neural network(CNN)detection method based on prior Hough enhancement is proposed.This method utilizes the Hough transform for spectral line pre-localization,and leverages the global perspective and powerful feature learning capabilities of deep learning to achieve accurate detection of dispersion characteristics,significantly improving detection performance.关键词
非均匀介质/色散现象/谐波特征/霍夫变换/卷积神经网络Key words
non-uniform medium/dispersion phenomena/harmonic characteristics/Hough transform/convolutional neural network分类
电子信息工程引用本文复制引用
刘子玄,陈智勇,余白石,钱良..基于深度学习的海洋色散信道信号处理研究[J].移动通信,2024,48(11):57-62,6.基金项目
国家重点研发计划项目"高性能海洋电场传感器设计与研制"(2022YFC3104001) （2022YFC3104001）
国家自然科学基金优秀青年科学基金项目"移动计算通信网络理论与方法"(62222111) （62222111）
