A Comprehensive Review Of Spatial Attention And Multi-Scale Dilated Convolutions For Land Use Mapping

Yuvraj Madhukar  Wadghule; Dr. Shashank  Swami

doi:10.29284/g2raby46

Authors

Yuvraj Madhukar Wadghule Department of Computer Science and Engineering ,Vikrant University, MP, India.
Dr. Shashank Swami Department of Computer Science and Engineering, Associate Professor, Vikrant University, MP, India.

DOI:

https://doi.org/10.29284/g2raby46

Keywords:

Land Use Mapping, Remote Sensing, Spatial Attention, Multi-Scale Dilated Convolution, Deep Learning, Semantic Segmentation, High-Resolution Imagery, Land Use and Land Cover Classification.

Abstract

The mapping of land use from remote sensing imagery is a critical component for city planning, environment monitoring, agriculture assessment, and sustainable resource management. The spatial variability of the objects, the land cover diversity, the size of objects and the similarity of their spectra in different land use classes, however, make the mapping of these objects difficult. Traditional CNNs are inadequate for the multi-scale contextual information and the fine spatial details preservation. In order to overcome these issues, a Spatial Attention-Driven Multi-Scale Dilated Convolutional Network (SA-MSDC) has been proposed in this paper to improve land use mapping using high-resolution remote sensing images. The proposed framework uses multi-scale dilated convolutions with various dilation rates to extract contextual information from multiple spatial dimensions without causing the feature-map loss due to resolution increase to be that high. A spatial attention mechanism is introduced to selectively highlight informative spatial regions and suppress irrelevant background features to gain better discrimination between complex and spectrally similar land use categories. The network also incorporates multi-level feature representations, which help to retain detailed boundary information and high-level semantic features. The proposed approach is tested with the benchmark remote sensing datasets in standard training and testing protocol. It is evaluated based on the accuracy, mean Intersection over Union, precision, recall and F1-score. Experimental results show that the proposed method can effectively enhance the accuracy of land-use classification and boundary delineation compared to the traditional convolutional method and attention-based method. It is a comprehensive and adaptable system for automated high-resolution land use classification and could be applied to urban planning, agricultural surveillance, environmental evaluation and geographic information system (GIS)-based decision making.

References

1] J. Long, E. Shelhamer, and T. Darrell, “Fully convolutional networks for semantic segmentation,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 39, no. 4, pp. 640–651, 2017.

[2] O. Ronneberger, P. Fischer, and T. Brox, “U-Net: Convolutional networks for biomedical image segmentation,” in Proc. International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI), 2015, pp. 234–241.

[3] F. Yu and V. Koltun, “Multi-scale context aggregation by dilated convolutions,” in Proc. International Conference on Learning Representations (ICLR), 2016.

[4] L.-C. Chen, G. Papandreou, I. Kokkinos, K. Murphy, and A. L. Yuille, “DeepLab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected CRFs,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 40, no. 4, pp. 834–848, 2018.

[5] L.-C. Chen, Y. Zhu, G. Papandreou, F. Schroff, and H. Adam, “Encoder-decoder with atrous separable convolution for semantic image segmentation,” in Proc. European Conference on Computer Vision (ECCV), 2018, pp. 801–818.

[6] S. Woo, J. Park, J.-Y. Lee, and I. S. Kweon, “CBAM: Convolutional block attention module,” in Proc. European Conference on Computer Vision (ECCV), 2018, pp. 3–19.

[7] K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proc. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, pp. 770–778.

[8] V. Badrinarayanan, A. Kendall, and R. Cipolla, “SegNet: A deep convolutional encoder-decoder architecture for image segmentation,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 39, no. 12, pp. 2481–2495, 2017.

[9] N. Audebert, B. Le Saux, and S. Lefèvre, “Semantic segmentation of earth observation data using multimodal and multi-scale deep networks,” in Proc. Asian Conference on Computer Vision (ACCV), 2016, pp. 180–196.

[10] K. Marmanis, K. Schindler, J. D. Wegner, S. Galliani, M. Datcu, and U. Stilla, “Classification with an edge: Improving semantic image segmentation with boundary detection,” ISPRS Journal of Photogrammetry and Remote Sensing, vol. 135, pp. 158–172, 2018.

[11] X. Wang, R. Girshick, A. Gupta, and K. He, “Non-local neural networks,” in Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2018, pp. 7794–7803.

[12] Z. Liu, Y. Lin, Y. Cao, H. Hu, Y. Wei, Z. Zhang, S. Lin, and B. Guo, “Swin Transformer: Hierarchical vision transformer using shifted windows,” in Proc. IEEE/CVF International Conference on Computer Vision (ICCV), 2021, pp. 10012–10022.

[13] E. Xie, W. Wang, Z. Yu, A. Anandkumar, J. M. Alvarez, and P. Luo, “SegFormer: Simple and efficient design for semantic segmentation with transformers,” in Advances in Neural Information Processing Systems (NeurIPS), vol. 34, pp. 12077–12090, 2021.

[14]“UNetFormer: A UNet-like transformer for efficient semantic segmentation of remote sensing urban scene imagery,” ISPRS Journal of Photogrammetry and Remote Sensing, vol. 190, pp. 196–214, 2022.

[15]“P-Swin: Parallel Swin transformer multi-scale semantic segmentation network for land cover classification,” Computers & Geosciences, vol. 175, Art. no. 105340, 2023.

[16] M. Hao, X. Dong, D. Jiang, X. Yu, F. Ding, and J. Zhuo, “Land-use classification based on high-resolution remote sensing imagery and deep learning models,” PLOS ONE, vol. 19, no. 4, 2024.

[17] X. Li, H. Zhao, D. Wu, Q. Liu, R. Tang, L. Li, Z. Xu, and X. Lyu, “SLMFNet: Enhancing land cover classification of remote sensing images through selective attentions and multi-level feature fusion,” PLOS ONE, vol. 19, no. 5, 2024.

[18] S. Wang, Q. Hu, S. Wang, P. Zhao, J. Li, and M. Ai, “Category attention guided network for semantic segmentation of fine-resolution remote sensing images,” International Journal of Applied Earth Observation and Geoinformation, vol. 127, 103661, 2024.

[19] V. Anitha, D. Manimegalai, and S. Kalaiselvi, “Downstream lingering attention transformer network (DsLATNet) for land use land cover classification: A bicolor deep learning framework,” Applied Soft Computing, vol. 165, 112074, 2024.

[20] X. Ma, R. Lian, Z. Wu, R. Guan, T. Hong, M. Zhao, M. Ma, J. Nie, Z. Du, S. Song, and W. Zhang, “A novel scene coupling semantic mask network for remote sensing image segmentation,” ISPRS Journal of Photogrammetry and Remote Sensing, vol. 221, pp. 44–63, 2025.

[21] Y. Ji, W. Shi, J. Lei, and J. Ding, “DBRSNet: A dual-branch remote sensing image segmentation model based on feature interaction and multi-scale feature fusion,” Scientific Reports, vol. 15, 27786, 2025.

[22] Y. Said, O. Saidani, A. D. Algarni, M. H. Algarni, and A. Flah, “Lightweight multiscale information aggregation network for land cover land use semantic segmentation from remote sensing images,” Scientific Reports, vol. 15, 30265, 2025.

[23] Z. Wang, Y. Li, and K. Ma, “RTAS-Net: A ResNet-transformer-ASPP semantic segmentation network for remote sensing images,” PLOS ONE, vol. 21, no. 4, e0343729, 2026. doi: 10.1371/journal.pone.0343729.