Cloud Detection over Sea Ice Using a Neural Network and Multi-Angle Imaging SpectroRadiometer (MISR) Imagery

This manuscript presents a novel cloud detection algorithm utilizing a neural network technique, developed for identifying cloudy and clear pixels over sea ice in MISR images. Our methodology is based on an extensive multi-angular dataset covering various Arctic regions in different seasons, demonstrating strong performance metrics, including F score and Accuracy.

We believe that this research contributes significantly to the remote sensing domain and offers a fresh approach to enhancing cloud detection accuracy over sea ice.