EcoSense AI: An AI-Based Environmental Change Detection in Coorg Using Satellite Data and U-Net Deep Learning | IJCT Volume 13 – Issue 5 | IJCT-V13I5P66
IJCT
International Journal of Computer Techniques
ISSN 2394-2231 · Peer-Reviewed · Open Access
📚 Volume 13, Issue 5
📅 October 26, 2026
📄 Pages 567–575
🔖 ID: IJCT-V13I5P66
EcoSense AI: An AI-Based Environmental Change Detection in Coorg Using Satellite Data and U-Net Deep Learning
Author(s)
Tejaswini B, Mamatha G, Megha S N, Shreyas S
Abstract
Environmental changes such as vegetation loss, deforestation, and land-use changes can gradually affect the ecological condition of a region. Monitoring these changes over large areas through manual inspection or basic satellite-image comparison can be time-consuming and may provide limited information about the spatial extent of the changes. This paper presents EcoSense AI, an AI-based environmental monitoring system developed to detect and analyze environmental changes in the Coorg region of Karnataka using satellite imagery and environmental data. The proposed system combines Sentinel-2 satellite observations with Normalized Difference Vegetation Index (NDVI) analysis to assess vegetation conditions across different time periods. A U-Net based deep learning model is incorporated for pixel-level analysis of satellite imagery, while environmental parameters such as temperature, rainfall, and soil moisture are used to provide additional context for ecosystem assessment. Temporal comparison of satellite-derived vegetation information is used to identify areas showing degradation or improvement. The system integrates these analytical components into a web-based platform for geospatial visualization and environmental assessment. Model performance is evaluated using standard metrics including accuracy, precision, recall, F1-score, Intersection over Union (IoU), and Dice coefficient. The proposed approach provides an automated framework for localized environmental monitoring and can assist in identifying areas that require further ecological investigation.
Keywords
Environmental Change Detection, Sentinel-2, Satellite Imagery, U-Net, Deep Learning, NDVI, Remote Sensing, Environmental Monitoring, Coorg.
Conclusion
EcoSense AI presents an AI-based framework for monitoring environmental changes in the Coorg region of Karnataka using satellite imagery, deep learning, vegetation indices, environmental observations, and geospatial visualization. The proposed framework integrates land-cover segmentation with temporal NDVI analysis to provide spatially interpretable information about vegetation and environmental changes. A U-Net-based deep learning model was implemented for pixel-level land-cover segmentation. The model achieved an overall accuracy of 92.4%, Precision of 91.8%, Recall of 92.4%, F1-score of 91.9%, Mean IoU of 83.6%, and Mean Dice coefficient of 90.9% on the evaluated test dataset. These results demonstrate the capability of the segmentation component to identify major land-cover patterns. The segmentation output is combined with temporal NDVI analysis to distinguish vegetation degradation, improvement, and relatively stable conditions. Environmental parameters such as temperature, rainfall, and soil moisture are incorporated as supporting information for interpreting observed vegetation variations. Overall, the proposed framework provides a foundation for localized environmental monitoring in Coorg by combining remote sensing, deep learning, environmental data analysis, and geospatial visualization. The system can support researchers and environmental stakeholders in identifying areas requiring further investigation.
References
[1] O. Pesek, L. Brodsky, L. Halounová, M. Landa, et al., “Convolutional neural networks for urban green areas semantic segmentation on Sentinel-2 data,” Remote Sensing Applications: Society and Environment, vol. 36, Art. no. 101238, 2024, doi: 10.1016/j.rsase.2024.101238.
[2] H. He, J. Yan, D. Liang, Z. Sun, J. Li, and L. Wang, “Time-series land cover change detection using deep learning-based temporal semantic segmentation,” Remote Sensing of Environment, vol. 305, Art. no. 114101, 2024, doi: 10.1016/j.rse.2024.114101.
[3] N. Dahiya et al., “A novel deep learning change detection approach for estimating spatiotemporal crop field variations from Sentinel-2 imagery,” Remote Sensing Applications: Society and Environment, vol. 35, Art. no. 101259, 2024, doi: 10.1016/j.rsase.2024.101259.
[4] M. Devara, V. K. Maurya, and R. Dwivedi, “Landslide extraction using a novel empirical method and binary semantic segmentation U-NET framework using Sentinel-2 imagery,” Remote Sensing Letters, vol. 15, no. 3, pp. 326 338, 2024, doi: 10.1080/2150704X.2024.2320178.
[5] R. Miao, G. Meng, K. Zhou, Y. Li, R. Chang, et al., “DASUNet: A deeply supervised change detection network integrating full-scale features,” Scientific Reports, 2024, doi: 10.1038/s41598-024-63257-8.
[6] E. A. Peker, M. Çalışkan, E. Bora, Ç. İnan, and M. E. Uçaner, “CORINE Level-2 Land Cover Classification from Sentinel-2 Images Using U-Net Deep Learning Model,” International Journal of Environment and Geoinformatics, vol. 12, no. 2, pp. 34–44, 2025
[7] J. Jagannathan, M. Thanjai Vadivel, and C. Divya, “Land use classification using multi-year Sentinel-2 images with deep learning ensemble network,” Scientific Reports, vol. 15, Art. no. 29047, 2025, doi: 10.1038/s41598-025-12512-7.
[8] T. F. Deressu, A. K. Bojer, T. G. Debelee, W. G. Negera, S. Nadarajah, and K. W. Gebissa, “Enhancing land use and land cover classification with deep learning-based satellite imagery segmentation,” International Journal of Applied Earth Observation and Geoinformation, 2025, Art. no. 104839, doi: 10.1016/j.jag.2025.104839.
[9] “Change detection in Sentinel-2 images using deep learning ensembles,” Remote Sensing Applications: Society and Environment, vol. 40, Art. no. 101764, 2025, doi: 10.1016/j.rsase.2025.101764.
[10] J. G. Borges et al., “A 3D convolutional neural network approach for monitoring land cover change using Sentinel-2 satellite imagery,” Frontiers in Forests and Global Change, 2025.
[11] P. M. Priya, “SAU-MTF: Siamese attention U-Net with multimodal temporal fusion for accurate deforestation detection,” Frontiers in Remote Sensing, vol. 7, 2026, Art. no. 1701194, doi: 10.3389/frsen.2026.1701194.
[12] “Hybrid ConvLSTM U-Net Deep Neural Network for Land Use and Land Cover Classification from Multi Temporal Sentinel-2 Images: Application to Yaoundé, Cameroon,” Mathematics and Computer Applications, vol. 31, no. 1, Art. no. 18, 2026, doi: 10.3390/mca31010018.
[2] H. He, J. Yan, D. Liang, Z. Sun, J. Li, and L. Wang, “Time-series land cover change detection using deep learning-based temporal semantic segmentation,” Remote Sensing of Environment, vol. 305, Art. no. 114101, 2024, doi: 10.1016/j.rse.2024.114101.
[3] N. Dahiya et al., “A novel deep learning change detection approach for estimating spatiotemporal crop field variations from Sentinel-2 imagery,” Remote Sensing Applications: Society and Environment, vol. 35, Art. no. 101259, 2024, doi: 10.1016/j.rsase.2024.101259.
[4] M. Devara, V. K. Maurya, and R. Dwivedi, “Landslide extraction using a novel empirical method and binary semantic segmentation U-NET framework using Sentinel-2 imagery,” Remote Sensing Letters, vol. 15, no. 3, pp. 326 338, 2024, doi: 10.1080/2150704X.2024.2320178.
[5] R. Miao, G. Meng, K. Zhou, Y. Li, R. Chang, et al., “DASUNet: A deeply supervised change detection network integrating full-scale features,” Scientific Reports, 2024, doi: 10.1038/s41598-024-63257-8.
[6] E. A. Peker, M. Çalışkan, E. Bora, Ç. İnan, and M. E. Uçaner, “CORINE Level-2 Land Cover Classification from Sentinel-2 Images Using U-Net Deep Learning Model,” International Journal of Environment and Geoinformatics, vol. 12, no. 2, pp. 34–44, 2025
[7] J. Jagannathan, M. Thanjai Vadivel, and C. Divya, “Land use classification using multi-year Sentinel-2 images with deep learning ensemble network,” Scientific Reports, vol. 15, Art. no. 29047, 2025, doi: 10.1038/s41598-025-12512-7.
[8] T. F. Deressu, A. K. Bojer, T. G. Debelee, W. G. Negera, S. Nadarajah, and K. W. Gebissa, “Enhancing land use and land cover classification with deep learning-based satellite imagery segmentation,” International Journal of Applied Earth Observation and Geoinformation, 2025, Art. no. 104839, doi: 10.1016/j.jag.2025.104839.
[9] “Change detection in Sentinel-2 images using deep learning ensembles,” Remote Sensing Applications: Society and Environment, vol. 40, Art. no. 101764, 2025, doi: 10.1016/j.rsase.2025.101764.
[10] J. G. Borges et al., “A 3D convolutional neural network approach for monitoring land cover change using Sentinel-2 satellite imagery,” Frontiers in Forests and Global Change, 2025.
[11] P. M. Priya, “SAU-MTF: Siamese attention U-Net with multimodal temporal fusion for accurate deforestation detection,” Frontiers in Remote Sensing, vol. 7, 2026, Art. no. 1701194, doi: 10.3389/frsen.2026.1701194.
[12] “Hybrid ConvLSTM U-Net Deep Neural Network for Land Use and Land Cover Classification from Multi Temporal Sentinel-2 Images: Application to Yaoundé, Cameroon,” Mathematics and Computer Applications, vol. 31, no. 1, Art. no. 18, 2026, doi: 10.3390/mca31010018.
📋 How to Cite This Paper
Tejaswini B, Mamatha G, Megha S N, Shreyas S (2026). EcoSense AI: An AI-Based Environmental Change Detection in Coorg Using Satellite Data and U-Net Deep Learning. International Journal of Computer Techniques, 13(5), 567–575. ISSN: 2394-2231. DOI: https://doi.org/10.5281/zenodo.23671582
