Real-Time Satellite Image Segmentation Using YOLOv8

Abstract

The project focuses on real-time satellite image segmentation using YOLOv8 for detecting and classifying land regions into three categories: Agriculture Lands, Water Bodies, and Urban Buildings. By leveraging advanced deep learning and object detection models, the system automates segmentation, providing higher accuracy and speed compared to traditional methods. The backend is developed using Flask, while the frontend uses HTML, CSS, and JavaScript to offer a user-friendly web interface for uploading images and visualizing results. This solution can significantly contribute to urban planning, environmental monitoring, and resource management, enabling timely insights into land usage for sustainable development.

Introduction

This project presents a real-time satellite image segmentation system that uses the YOLOv8 deep learning model to automatically identify land-cover features such as agricultural areas, water bodies, and urban regions. The system is designed as a web-based application with a Flask backend and a frontend built using HTML, CSS, and JavaScript, allowing users to upload satellite images and view segmentation results easily.

The motivation behind the work is the growing need for fast, accurate, and automated analysis of satellite data, as manual interpretation is slow and error-prone. The system addresses challenges like varying image quality, lighting conditions, and complex landscapes by using advanced preprocessing, data augmentation, and deep learning techniques.

The methodology includes data collection from sources like NASA and Sentinel, preprocessing (resizing, normalization, augmentation), feature extraction using CNN-based models, model training with techniques like transfer learning and GANs, and evaluation using metrics such as IoU, precision, recall, and F1-score. YOLOv8 is selected for its speed and accuracy in real-time detection and segmentation.

The system follows a client–server architecture where the frontend handles user interaction and visualization, while the Flask backend performs preprocessing, model inference, and result generation. Outputs are refined using post-processing techniques and displayed as segmented maps for better interpretation.

The dataset consists of annotated satellite images covering agriculture, water bodies, and urban areas in YOLO segmentation format. The system is scalable, can be deployed on cloud platforms, and may be extended in the future to include additional data types like hyperspectral images and NDVI maps.

Conclusion

The Satellite Image Segmentation System successfully transforms complex satellite imagery into clear, actionable land-use maps. By harnessing the power of YOLOv8, the system accurately identifies key features such as vegetation, water bodies, agricultural fields, bare land, and urban areas, even across diverse geographical and atmospheric conditions. With a fast and reliable backend, a clean and intuitive web interface, and full support for georeferenced inputs and outputs, users can simply upload an image and receive professional-quality results in seconds. The entire solution is designed to work both on a local computer and on the cloud, making it accessible to individuals, organizations, and governments a like.More than just a prototype, this system delivers a practical, ready-to-usetool that brings advanced artificial intelligence to real-world challenges in environmental monitoring,urban planning, agriculture, and disaster response.It empowers userswhether experts or beginnersto understand and manage our changing planet

References

[1] M. Barthakur and K.K. Sarma, Deep Learning Based Semantic Segmentation Applied to Satellite Image, In book: Data Visualization and Knowledge Engineering, pp.-79-108, Springer, 2019. [2] A. Krizhevsky, I. Sutskever, G.E. Hinton, ImageNet classification with deep convolutional neural networks, in: Advances in Neural Information Processing Systems, pp. 1097-1105, 2012. [3] K. Simonyan, A. Zisserman, Very Deep Convolutional Networks for Large-Scale Image Recognition, arXiv:1409.15, 2014. [4] K. He, X. Zhang, S. Ren, J. Sun, Deep residual learning for image recognition, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition pp. 770-778, 2016. [5] R. Girshick, J. Donahue, T. Darrell, J. Malik Rich feature hierarchies for accurate object detection and semantic segmentation,In: CVPR,2014. [6] G. Cheng, Z. Li, J. Han, X. Yao and L. Guo, Exploring Hierarchical Convolutional Features for Hyperspectral Image Classification, in IEEE Transactions on Geoscience and Remote Sensing, vol. 56, no. 11, pp. 6712-6722, Nov. 2018. [7] P. Zhou, J. Han, G. Cheng and B. Zhang, Learning Compact and Discriminative Stacked Autoencoder for Hyperspectral Image Classification, in IEEE Transactions on Geoscience and Remote Sensing, vol. 57, no. 7, pp. 4823-4833, July 2019. [8] M. Wurma, T. Starkb, X. X. Zhubc,M. Weigandad, H. Taubenbocka, Semantic segmentation of slums in satellite images using transfer learning on fully convolutional neural networks, ISPRS Journal of Photogrammetry and Remote Sensing, Elsevier, vol. 150, pp. 59-69 April 2019, [9] J. Long, E. Shelhamer and T. Darrell, Fully Convolutional Network for Semantic Segmentation, In CPVR, 2015. [10] L. Zhu, Y. Chen, P. Ghamisi and J. A. Benediktsson, Generative Adversarial Networks for Hyperspectral Image Classification, IEEE Transaction on Geoscience and Remote Sensing, VOL. 56, NO. 9, pp. 5046-5062, 2018. [11] Bochkovskiy, A., Wang, C.-Y., & Liao, H.-Y. M. (2020). YOLOv4: Optimal speed and accuracy of object detection. arXiv:2004.10934. [12] Jocher, G., Chaurasia, A., & Qiu, J. (2023). YOLOv8: Next-generation vision AI. Ultralytics. Retrieved from https://docs.ultralytics.com [13] Long, J., Shelhamer, E., & Darrell, T. (2015). Fully convolutional networks for semantic segmentation. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 3431–3440. [14] Ronneberger, O., Fischer, P., & Brox, T. (2015). U-Net: Convolutional networks for biomedical image segmentation. Lecture Notes in Computer Science. [15] Zhang, L., & Ma, J. (2021). Remote sensing image segmentation based on deep learning: A survey. Information Fusion, 73, 145–168. [16] Ma, L., Liu, Y., Zhang, X., Ye, Y., Yin, G., & Johnson, B. A. (2019). Deep learning in remote sensing applications: A meta-analysis and review. [17] Kamavisdar, P., Saluja, S., & Agrawal, S. (2013). A survey on image classification approaches and techniques. International Journal of Advanced Research in Computer and Communication Engineering [18] Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep Learning. MIT Press. [19] Chollet, F. (2017). Deep Learning with Python. [20] Singh, A. (1989). Digital change detection techniques using remotely-sensed data. International Journal of Remote Sensing, 10(6), 989–1003. [21] Tuia, D., Persello, C., & Bruzzone, L. (2016). Domain adaptation for the classification of remote sensing data: An overview of recent advances. IEEE Geoscience and Remote Sensing Magazine, 4(2), 41–57. [22] Zhang, Z., Liu, Q., & Wang, Y. (2018). Road extraction by deep residual U-Net. IEEE Geoscience and Remote Sensing Letters, 15(5), 749–753. [23] Audebert, N., Le Saux, B., & Lefèvre, S. (2017). Semantic segmentation of Earth observation data using multimodal and multi-scale deep networks. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 1–9. [24] Simonyan, K., & Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv:1409.1556. [25] Kingma, D. P., & Ba, J. (2015). Adam: A method for stochastic optimization. International Conference on Learning Representations (ICLR

Copyright

Copyright © 2026 Mrs. R. Prathyusha, G. Humsika, P. Saketh, S. Yashwant Raj Reddy . This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.