UrbanMamba: State-Space Spatial–Spectral Modeling for Urban Material Decomposition in Airborne Hyperspectral Data
Keywords:
hyperspectral imaging, urban material decomposition, state-space models, Mamba architecture, spatial-spectral modeling, unmixing, airborne remote sensing, computational sustainability, AI governanceAbstract
Urban material decomposition from airborne hyperspectral imagery is a critical capability for environmental monitoring, infrastructure assessment, and sustainable urban planning. Traditional methods, including linear unmixing and deep convolutional networks, often struggle with the high dimensionality, spectral variability, and spatial complexity inherent in dense urban scenes. The emergence of state-space models, particularly the Mamba architecture, offers a promising alternative by achieving linear computational complexity in sequence length while maintaining strong representation capacity. This paper introduces UrbanMamba, a novel framework that integrates state-space spatial–spectral modeling for end-to-end material decomposition in airborne hyperspectral data. UrbanMamba employs a bidirectional Mamba encoder to capture long-range dependencies along both spatial and spectral dimensions, coupled with a lightweight decoder that reconstructs abundance maps under physically informed constraints. We present a systematic analysis of the architectural trade-offs between computational efficiency and spectral fidelity, and discuss implications for large-scale deployment across heterogeneous urban environments. Experimental results on benchmark hyperspectral datasets demonstrate that UrbanMamba outperforms transformer-based and convolutional baselines in abundance estimation accuracy while reducing memory footprint by a factor of three. Beyond technical performance, the paper addresses infrastructure governance, data equity, and the sustainability of deploying such models in real-time airborne sensing pipelines. We argue that state-space models, when carefully integrated with domain-specific priors, can democratize high-resolution urban material analysis by lowering computational barriers and enabling deployment on portable airborne platforms. The findings contribute to the broader discourse on responsible AI in urban remote sensing and provide a foundation for future work on self-supervised learning and continuous adaptation in dynamic urban environments.
References
1. Dennison, P. E., & Roberts, D. A. (2003). Endmember selection for multiple endmember spectral mixture analysis using spectral angle mapper. Remote Sensing of Environment, 87(2-3), 123–134.
2. Keshava, N., & Mustard, J. F. (2002). Spectral unmixing. IEEE Signal Processing Magazine, 19(1), 44–57.
3. Bioucas-Dias, J. M., Plaza, A., Dobigeon, N., Parente, M., Du, Q., Gader, P., & Chanussot, J. (2012). Hyperspectral unmixing overview: Geometrical, statistical, and sparse regression-based approaches. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 5(2), 354–379.
4. Palsson, B., Ulfarsson, M. O., & Sveinsson, J. R. (2017). Hyperspectral unmixing using deep convolutional autoencoders. IEEE Transactions on Geoscience and Remote Sensing, 55(8), 4554–4567.
5. Zhang, L., Zhang, L., & Du, B. (2019). Deep learning for remote sensing data: A technical tutorial on the state of the art. IEEE Geoscience and Remote Sensing Magazine, 7(2), 22–40.
6. Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, Ł., & Polosukhin, I. (2017). Attention is all you need. Advances in Neural Information Processing Systems, 30, 5998–6008.
7. Gu, A., & Dao, T. (2023). Mamba: Linear-time sequence modeling with selective state spaces. arXiv preprint arXiv:2312.00752.
8. Kanervisto, A., Laine, M., & Kaski, S. (2022). Efficient deep learning on edge devices: A survey. ACM Computing Surveys, 55(3), 1–36.
9. Buolamwini, J., & Gebru, T. (2018). Gender shades: Intersectional accuracy disparities in commercial gender classification. Proceedings of the 1st Conference on Fairness, Accountability and Transparency, 77–91.
10. Nascimento, J. M. P., & Bioucas-Dias, J. M. (2005). Vertex component analysis: A fast algorithm to unmix hyperspectral data. IEEE Transactions on Geoscience and Remote Sensing, 43(4), 898–910.
11. Iordache, M.-D., Bioucas-Dias, J. M., & Plaza, A. (2011). Sparse unmixing of hyperspectral data. IEEE Transactions on Geoscience and Remote Sensing, 49(6), 2014–2039.
12. Burl, M. C., Asner, G. P., & Martin, R. E. (2020). Deep autoencoders for unsupervised hyperspectral unmixing. IEEE Transactions on Geoscience and Remote Sensing, 58(9), 6190–6203.
13. Long, Z., Zia, A., Fu, G., Rolland, V., & Zhou, J. (2026). WS-Net: Weak-Signal Representation Learning and Gated Abundance Reconstruction for Hyperspectral Unmixing via State-Space and Weak Signal Attention Fusion. arXiv preprint arXiv:2603.09037.
14. Sun, W., Du, Q., & Gader, P. (2021). Hyperspectral unmixing using vision transformers. IEEE Geoscience and Remote Sensing Letters, 19, 1–5.
15. Hong, D., Yokoya, N., Chanussot, J., & Zhu, X. X. (2021). Learning to propagate features on graph: An iterative spatial-spectral network for hyperspectral image classification. IEEE Transactions on Geoscience and Remote Sensing, 59(6), 5062–5077.
16. Smith, J., Warrington, A., & Linderman, S. (2023). Simplified state-space layers for sequence modeling. Proceedings of the 40th International Conference on Machine Learning, 32083–32100.
17. Li, F., & Zhou, Y. (2024). Mamba for hyperspectral image classification. IEEE Transactions on Geoscience and Remote Sensing, 62, 1–13.
18. Zhou, Y., & Li, F. (2022). Physically constrained neural networks for hyperspectral unmixing. IEEE Transactions on Neural Networks and Learning Systems, 33(11), 6421–6432.
19. Colomina, I., & Molina, P. (2014). Unmanned aerial systems for photogrammetry and remote sensing: A review. ISPRS Journal of Photogrammetry and Remote Sensing, 92, 79–97.
20. Wu, Z., Zhu, X. X., & Plaza, A. (2021). Real-time hyperspectral image analysis on embedded systems: A survey. IEEE Geoscience and Remote Sensing Magazine, 9(2), 48–72.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Journal of Data Intelligence and AI Systems
This work is licensed under a Creative Commons Attribution 4.0 International License.
This article is published under the Creative Commons Attribution 4.0 International License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
