In order to join virtual sessions, you must be registered and logged-in(Were you registered for the in-person meeting in Denver? if yes, just log in.) 
Note: All session times are in Mountain Daylight Time.

Leveraging Deep Learning Method in Sub-pixel Cloud Detection and Rainy Cloud Classification

Authors: Qian Liu*, George Mason University, Chaowei Yang, George Mason University, Hui Xu, University of Marryland, Manzhu Yu, Pennsylvania State University
Topics: Geographic Information Science and Systems
Keywords: Precipitation detection; Convective; Stratiform; ABI; Machine learning; Deep Learning
Session Type: Paper
Presentation File: No File Uploaded


Cloud is an important atmospheric factor that influence both the long-term climate change and short-term weather dynamics of the earth. Its distribution and types substantially affects Earth’s energy budget and related to various nature phenomena such as precipitation. This study leverages the up-to-date deep learning method, deep neural network in multiple satellite data sub-pixel cloud detection and rainy cloud classification. For sub-pixel cloud detection, DNN model has a relatively high accuracy with a total probability of detection (POD) of 96.51%, and critical success index (CSI) of 91.02%; for rainy cloud classification, DNN achieves a CSI of 0.71 and a POD of 0.86 in rainy cloud detection, and a CSI of 0.58, POD of 0.72 in convective precipitation delineation.

To access contact information login