Leveraging spatial textures, through machine learning, to identify aerosols and distinct cloud types from multispectral observations

Marais, Willem J.; Holz, Robert E.; Reid, Jeffrey S.; Willett, Rebecca M.

doi:https://doi.org/10.5194/amt-13-5459-2020

Articles | Volume 13, issue 10

https://doi.org/10.5194/amt-13-5459-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/amt-13-5459-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 13, issue 10

Research article

|

14 Oct 2020

Research article |

| 14 Oct 2020

Leveraging spatial textures, through machine learning, to identify aerosols and distinct cloud types from multispectral observations

Willem J. Marais, Robert E. Holz, Jeffrey S. Reid, and Rebecca M. Willett

Model code and software

TensorFlow-Slim image classification model library Google https://github.com/tensorflow/models/tree/master/research/slim

Short summary

Space agencies use moderate-resolution satellite imagery to study how smoke, dust, pollution (aerosols) and cloud types impact the Earth's climate; these space agencies include NASA, ESA and the China Meteorological Administration. We demonstrate in this paper that an algorithm with convolutional neural networks can greatly enhance the automated detection of aerosols and cloud types from satellite imagery. Our algorithm is an improvement on current aerosol and cloud detection algorithms.