Yasmeen et al. (2026) An adaptive deep learning framework for multi-temporal crop and drought stress monitoring in precision agriculture

Identification

• Journal: Scientific Reports
• Year: 2026
• Date: 2026-09-09
• Authors: Gausiya Yasmeen, Tasneem Ahmed
• DOI: 10.1038/s41598-026-70304-z

Research Groups

• Advanced Computing Research Laboratory, Department of Computer Application, Integral University, Lucknow, India.

Short Summary

The study proposes a hybrid deep learning framework combining Convolutional Neural Networks (CNN) and Vision Transformers (ViT) to monitor crop and drought stress using multi-temporal Sentinel-2 satellite imagery.

Objective

• To develop an adaptive deep learning-based framework for assessing crop stress dynamics and drought monitoring to improve agricultural productivity and regional food security in semi-arid landscapes.

Study Configuration

• Spatial Scale: Haldharmau village, Gonda District, Uttar Pradesh, India.
• Temporal Scale: Multi-temporal analysis covering the years 2023, 2024, and 2025, with specific monthly stress maps generated from January to June.

Methodology and Data

• Models used: Hybrid DL-based classification model (CNN+ViT) integrated with NDVI (Normalized Difference Vegetation Index) and NDWI (Normalized Difference Water Index).
• Data sources: Multi-temporal Sentinel-2 satellite images.

Main Results

• Successfully generated crop stress and drought maps that provide quantitative insights into the spatial and temporal patterns of crop vulnerability and resilience.
• The framework effectively integrates local spectral/spatial details with broader contextual features to improve crop classification and stress detection.

Contributions

• Advances precision agriculture by integrating satellite imagery, hybrid deep learning (CNN+ViT), and phenological analysis.
• Provides a transferable methodology for drought monitoring in other semi-arid agricultural regions with similar crop systems.

Funding

• No external funding was received for this research (Internal support provided by Integral University; MCN: IU/R&D/2026-MCN0004367).

Citation

@article{Yasmeen2026adaptive,
  author = {Yasmeen, Gausiya and Ahmed, Tasneem},
  title = {An adaptive deep learning framework for multi-temporal crop and drought stress monitoring in precision agriculture},
  journal = {Scientific Reports},
  year = {2026},
  doi = {10.1038/s41598-026-70304-z},
  url = {https://doi.org/10.1038/s41598-026-70304-z}
}