固体地球物理学术报告通知-刘琳

The cryosphere, Earth’s surface where water is frozen, has been undergoing strong warming and area reduction and casting substantial impacts on the global climate system and socio-economical development. This talk reviews the use of deep learning for cryospheric studies, aiming to showcase the diverse applications of deep learning for tackling research tasks that are becoming more challenging to conventional approaches. Even though such applications are still in an early stage, deep learning has been utilized to characterize nearly all the cryospheric components in a diverse manner. I will give two examples to illustrate how our research group utilizes state-of-the-art deep learning algorithms in the processing of remote sensing imagery towards automated detection of fast changes in glaciers and frozen ground. Finally, I will offer some thoughts on common strengths and limitations as well as future directions of deep-learning-based cryospheric studies.