Parsa Shamsaddini - Faculty of Intelligent Systems Engineering and Data Sciences, Persian Gulf University, Busher, Iran
Ahmad Keshavarz - Faculty of Intelligent Systems Engineering and Data Sciences, Persian Gulf University, Busher, Iran
Hojat Ghimatgar - Faculty of Intelligent Systems Engineering and Data Sciences, Persian Gulf University, Busher, Iran
Stefano Zecchetto - Istituto di Scienze Polari, Consiglio Nazionale delle Ricerche ,Padova, Italy

This paper investigates sea-ice discrimination using SAR images through the utilizationof GLCM (Gray-Level Co-occurrence Matrix) feature extraction coupled with L-scorefeature selection. By focusing on the specific challenge of distinguishing between sea and ice,we aim to streamline the process while maintaining accuracy. Our approach efficiently extractstexture features from Sentinel-۱ images and employs L-score feature selection to mitigate computationalburden without compromising discrimination efficacy. This methodology offers apromising avenue for expediting sea-ice discrimination tasks, essential for various remote sensingand environmental monitoring applications. At the end, this offers significant time savings byapplying the feature selection method, which can happen with almost the same accuracy.

SAR, L-score, GLCM, Sea-ice discrimination, Feature Selection, Sentinel-۱