Multi-branch and Dual-task Method for Road Extraction from Multimodal Remote Sensing Images

被引：0
作者：
Lin Y. [1 ]
Jin F. [1 ]
Wang S. [1 ]
Zuo X. [1 ]
Dai L. [2 ]
Huang Z. [1 ]
机构：
[1] Institute of Geospatial Information, Information Engineering University, Zhengzhou
[2] College of Computer Science and Artificial Intelligence, Wenzhou University, Wenzhou
来源：
Journal of Geo-Information Science | 2024年 / 26卷 / 06期
基金：
中国国家自然科学基金;
关键词：
convolution neural network; data fusion; intelligent interpretation; multimodal; optical image; road extraction; SAR image;
D O I：
10.12082/dqxxkx.2024.240101
中图分类号：
学科分类号：
摘要：
Optical images and SAR images have rich complementary attributes, and an effective data fusion strategy can provide a solid information base for objects interpretation. Roads, as strip features, their topology, distribution patterns, and application scenarios often pose challenges to the interpretation results. Based on this, this paper proposes a multi-branch and dual-task method for road extraction from multimodal remote sensing images. First, encoding-decoding networks with the same structure but independent parameters are constructed for feature extraction of optical and SAR images, respectively, and road surface segmentation labels are used for supervised training. Second, the coding layer features of the SAR images are introduced for road edge detection, and their intermediate features are input to the decoding layer features of the SAR image, so as to optimize the discrimination effect between the road and the background. Finally, the designed Channel Attention-Strip Spatial Attention (CA-SSA) is utilized to fully fuse the shallow and deep features of optical and SAR images to predict the final road extraction results. In the experiment, using the Dongying data set as the reference, it is proved that the method of this paper is superior to the comparative methods based on quantitative evaluation metrics, has obvious advantages in challenging areas such as road intersection and low- grade roads, and has best road extraction results when optical images is affected by clouds. © 2024 Science Press. All rights reserved.
引用
收藏
页码：1547 / 1561
页数：14
相关论文
共 52 条
  • [1] Wang C.W., Xu R.T., Xu S.B., Et al., Toward accurate and efficient road extraction by leveraging the characteristics of road shapes[J], IEEE Transactions on Geoscience and Remote Sensing, 61, (2023)
  • [2] Han L.Y., Hou L., Zheng X.X., Et al., Segmentation is not the end of road extraction: An all-visible denoising auto-encoder for connected and smooth road reconstruction[J], IEEE Transactions on Geoscience and Remote Sensing, 61, (2023)
  • [3] Shelhamer E., Long J., Darrell T., Fully convolutional networks for semantic segmentation[C], IEEE Transactions on Pattern Analysis and Machine Intelligence, pp. 640-651, (2017)
  • [4] Ronneberger O., Fischer P., Brox T., U- net: Convolutional networks for biomedical image segmentation[C], International Conference on Medical Image Computing and Computer- Assisted Intervention, 2015, pp. 234-241, (2015)
  • [5] Badrinarayanan V., Kendall A., Cipolla R., SegNet: A deep convolutional encoder-decoder architecture for image segmentation[J], IEEE Transactions on Pattern Analysis and Machine, 39, 12, pp. 2481-2495, (2017)
  • [6] Li X., Zhang G., Cui H., Et al., MCANet: A joint semantic segmentation framework of optical and SAR images for land use classification[J], International Journal of Applied Earth Observation and Geoinformation, 106, (2022)
  • [7] Sukawattanavijit C., Chen J., Zhang H.S., GA- SVM algorithm for improving land- cover classification using SAR and optical remote sensing data[J], IEEE Geoscience and Remote Sensing Letters, 14, 3, pp. 284-288, (2017)
  • [8] Zhang H.S., Li J., Wang T., Et al., A manifold learning approach to urban land cover classification with optical and radar data[J], Landscape and Urban Planning, 172, pp. 11-24, (2018)
  • [9] Zhang H.S., Wan L.M., Wang T., Et al., Impervious surface estimation from optical and polarimetric SAR data using small- patched deep convolutional networks: A comparative study[J], IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 12, 7, pp. 2374-2387, (2019)
  • [10] Li X., Lei L., Sun Y.L., Et al., Multimodal bilinear fusion network with second-order attention-based channel selection for land cover classification[J], IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 13, pp. 1011-1026, (2020)
123456