光电子技术, 2020, 40 (1): 6, 网络出版: 2020-04-26
基于残差神经网络的道路提取算法研究
Research on Road Extraction Algorithm Based on Residual Neural Networks
道路提取 遥感图像 空洞卷积 多尺度特征融合 road extraction remote sensing image dilated convolution multi-scale feature fusion
摘要
针对遥感图像道路提取信息丢失问题,提出了一种基于残差神经网络的道路提取算法。首先构建编码器?解码器网络,结合预编码器以及空洞卷积模块进行训练,提取更多的语义信息;其次并联设计的空洞卷积模块加在编码器?解码器结构的中间部分,它可以对不同感受野的特征图进行特征提取;最后编码器?解码器之间采用跳连的方式进行多尺度的特征融合,学习更多低维和高维的特征。实验结果表明,在Massachusetts道路数据集上,该方法相比其他算法在Precision、Recall和F1?score性能指标上分别有11 %、0.3 %和7.4 %的提升;同时在Accuracy指标上也达到了97.9 %,相比于其他算法,该算法有一定的应用价值。
Abstract
In order to solve the problem of road extraction information loss in remote sensing images, a road extraction algorithm based on residual neural networks was proposed. Firstly, an encoder-decoder network was constructed, combined with pre-coder and dilated convolution module to extract more semantic information.Secondly, the parallel designed dilated convolution module was added to the middle part of the encoder-decoder structure, which could extract features of different receptive field features. Finally, the encoder-decoder used jumper to perform multi-scale feature fusion, learning more low-dimensional and high-dimensional features.In the Massachusetts road dataset, this method had 11 %, 0.3 %, and 7.4 % improvement in Precision, Recall, and F1-score performance indicators. At the same time, it also achieved 97.9 % in the Accuracy index. Compared with other algorithms, the algorithm has certain application value.
熊炜, 管来福, 童磊, 王传胜, 刘敏, 曾春艳. 基于残差神经网络的道路提取算法研究[J]. 光电子技术, 2020, 40(1): 6. Wei XIONG, Laifu GUAN, Lei TONG, Chuansheng WANG, Min LIU, Chunyan ZENG. Research on Road Extraction Algorithm Based on Residual Neural Networks[J]. Optoelectronic Technology, 2020, 40(1): 6.