[1]单传辉,叶绍华,姚万琪,等.基于深度可分离残差网络的遥感影像路网检测[J].计算机技术与发展,2023,33(04):75-81.[doi:10. 3969 / j. issn. 1673-629X. 2023. 04. 011]
　SHAN Chuan-hui,YE Shao-hua,YAO Wan-qi,et al.Remote Sensing Image Road Network Detection Based on Deep Separable Residual Network[J].,2023,33(04):75-81.[doi:10. 3969 / j. issn. 1673-629X. 2023. 04. 011]
点击复制

基于深度可分离残差网络的遥感影像路网检测()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
33
期数:
2023年04期
页码:
75-81
栏目:
媒体计算
出版日期:
2023-04-10

文章信息/Info

Title:
Remote Sensing Image Road Network Detection Based on Deep Separable Residual Network
文章编号:
1673-629X(2023)04-0075-07
作者:
单传辉; 叶绍华; 姚万琪; 张 欣
安徽工程大学 电气工程学院,安徽 芜湖 241004
Author(s):
SHAN Chuan-hui; YE Shao-hua; YAO Wan-qi; ZHANG Xin
School of Electrical Engineering,Anhui Polytechnic University,Wuhu 241004,China
关键词:
遥感影像; 路网检测; 深度可分离卷积运算; 残差模块; 深度可分离残差网络
Keywords:
remote sensing image;  road network detection;  deep separable convolutional operation;  residual module;  deep separable residual network
分类号:
TP183
DOI:
10. 3969 / j. issn. 1673-629X. 2023. 04. 011
摘要:
从高分辨率遥感影像中提取并检测路网一直都是计算机视觉研究的热点和难点。 目前,基于深度学习的遥感影像路网检测方法大部分都是以卷积运算为基础的卷积神经网络,而以深度可分离卷积运算为基础深度可分离卷积神经网络作为以卷积运算为基础的卷积神经网络的替代神经网络,不仅在特征提取能力上优于卷积神经网络,而且在参数量和计算量方面也低于卷积神经网络。 鉴于此,该文利用深度可分离卷积运算替换卷积运算,并引入残差模块,构造了深度可分离残差网络进行遥感影像的路网自动检测的应用。 实验结果表明,在 RRSI 和 CHN6 -CUG 数据集上,虽然深度可分离残差网络的准确率和损失与相对应的卷积神经网络和残差网络的准确率和损失的区别不大,但是深度可分离残差网络的训练耗时时长远远低于相对应的卷积神经网络和残差网络的训练耗时时长,而且深度可分离残差网络的路网检测实际结果也优于相对应的卷积神经网络和残差网络的路网检测实际结果。
Abstract:
Extracting and detecting road network consistency from high resolution remote sensing images has been a hotspot and difficultyin computer vision research. At present,the existing network remote sensing image detection method based on deep learning is mostlybased on the convolutional operation and convolutional neural network. As an alternative neural network to convolutional neural network,deep separable convolutional neural network based on the deep separable convolutional operation is not only superior to convolutionalneural network in feature extraction ability, but also superior to convolutional neural network in parameter amount and calculationamount.In view of this,we use the deep separable convolutional operation to replace the convolutional operation,and introducing theresidual module, construct a deep separable residual network to detect road network automatically of remote sensing images. Theexperimental results show that on RRSI and CHN6-CUG datasets,although the accuracy and loss of the deep separable residual networkis not significantly different from those of the corresponding convolutional neural network and residual network,the training time of thedeep separable residual network is much longer than that of the corresponding convolutional neural network and residual network.Moreover,the actual results of the road network detection of the deep separable residual network are also better than those of thecorresponding convolutional neural network and residual network.

相似文献/References:

[1]余 威,龙慧云.基于深度卷积网络的遥感影像建筑物分割方法[J].计算机技术与发展,2019,29(06):57.[doi:10. 3969 / j. issn. 1673-629X. 2019. 06. 012]
　YU Wei,LONG Hui-yun.A Building Segmentation Method Based on Deep Convolution Networks for Remote Sensing Imagery[J].,2019,29(04):57.[doi:10. 3969 / j. issn. 1673-629X. 2019. 06. 012]
[2]冯 澳,李鸿翔,刘子曦,等.遥感影像水体提取模型的新型研究[J].计算机技术与发展,2021,31(增刊):56.[doi:10. 3969 / j. issn. 1673-629X. 2021. S. 011]
　FENG Ao,LI Hong-xiang,LIU Zi-xi,et al.Research on Water Extraction Model from Remote Sensing Image Based on Improved Deeplabv3+ Network[J].,2021,31(04):56.[doi:10. 3969 / j. issn. 1673-629X. 2021. S. 011]
[3]王 诏*,王 燕,苏国辉,等.一种阶梯型多尺度神经网络条带噪声降噪模型[J].计算机技术与发展,2023,33(07):68.[doi:10. 3969 / j. issn. 1673-629X. 2023. 07. 010]
　WANG Zhao*,WANG Yan,SU Guo-hui,et al.A Staircase Multi-scale Neural Network Model for Stripe Noise Reduction[J].,2023,33(04):68.[doi:10. 3969 / j. issn. 1673-629X. 2023. 07. 010]

更新日期/Last Update: 2023-04-10