The Potential Application of Deep Machine Learning Technique for the Space Lidar Retrievals
Abstract
The Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP), on-board The Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) is a space-borne polarization sensitive lidar developed by the collaboration of NASA and CNES (The American and French Space Study Centers). It is profiling from the space the Earth-Atmosphere system since June 2006 which greatly helps the community to study and understand the Earth-Atmosphere system and the Climate changes. The traditional retrieval way for the space lidar is to use the layer integrated signal with a technique to correct lidar attenuation for the detection and the classification of atmospheric features, and for the retrievals of their microphysical properties. Since the development of the computer power and speed, the deep machine learning technique has been unprecedentedly improved. Benefit from this new method that could combine both the channel and texture information to recognize feature pattern, in this study, we will discuss the application of this potential new technique in the space lidar retrievals. We will show that the deep machine learning methods can improve the feature detection from the single shot observations, to improve the feature classification and the feature properties retrievals by using the lidar profile observations. The results will not only show that this new method takes in additional information from vertical texture of the feature instead of using averaged layer information where texture information has been missed due to integration or average. But also, the new method can better deal with the instrument noise and the lidar attenuation from above layers.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2019
- Bibcode:
- 2019AGUFM.A33M2972Z
- Keywords:
-
- 3360 Remote sensing;
- ATMOSPHERIC PROCESSES;
- 0555 Neural networks;
- fuzzy logic;
- machine learning;
- COMPUTATIONAL GEOPHYSICS;
- 1610 Atmosphere;
- GLOBAL CHANGE;
- 1942 Machine learning;
- INFORMATICS