University of Kufa Official Website for the University of Kufa
A research team from the University of Kufa has published a study titled “Developing Aeolian Sand Spectral Index Using Multispectral Imagery and Machine Learning Models: A Representative Case Study in Iraq.”
Prof. Dr. Ahmed Al-Sultani, Assistant Prof. Dr. Ihsan Ali (Faculty of Physical Planning), and Prof. Dr. Furqan Hassan (Faculty of Computer Science and Mathematics) co-authored the study.
The research aimed to develop a novel spectral index for detecting aeolian sands in the Najaf–Samawah dune field using Landsat OLI satellite imagery and advanced machine learning (ML) models. The team generated 15 normalized difference (ND) spectral indices and tested them across nine ML algorithms, including LibSVM, LibLINEAR, and the Generalized Linear Model (GLM).
Two indices were successfully developed: the comprehensive DSI-C and the simplified DSI-R. The latter demonstrated exceptional performance, achieving 93.617% overall accuracy, 87.233% mean Kappa coefficient, and 93.331% mean F1-score, establishing its efficacy in high-precision aeolian sand mapping.
This innovation holds significant potential for environmental monitoring and desertification mitigation in arid regions, aligning with global sustainability goals.
