3D Remote Sensing of Clouds  

Recent advances in multi-view high-resolution instruments and computation power enable, in principle, 3D volumetric recovery of clouds. This is in contrast to the current operational remote-sensing paradigm which relies heavily on 1D radiative transfer. The atmosphere is modeled as plane-parallel horizontal slabs which do not express the true 3D nature of clouds, thus biasing retrievals. Our work models the atmosphere as a full 3D grid and proposes a novel approach for passive (relying on solar radiation) scatterer tomography. We fit a 3D volumetric model of scatterers (cloud droplets) to multi-view/multi-spectral polarimetric images. The forward model is a numerical 3D radiative transfer solver. Model-to-data fit is posed as a high-dimensional optimization problem that is made computationally tractable on large scales, thanks to an efficient algorithm we developed.

CloudCT is a space mission, directly derived from my Ph.D. research, to launch 10 pico-satellites that will orbit in a formation and gather multi-view observations of clouds. CloudCT, lead by Yoav SchechnerIlan Koren, and Klaus Schilling, has won ERC Synergy funding of 14 million euros.​ For more information see the ClouCT page. Below are two (very short) clips highlighting the CloudCT mission concept and our breakthrough computational approach. 

Below are longer more technical videos, detailing the mathematical formulation and algorithms, given at SPIE 2021 and ICCV 2015. The more recent work incorporates information from the polarized scattered signal to uncover the droplet size distribution within the cloud.



As part of an effort to make this research open, useful, and reproducible, we have spent many hours coding, re-coding, and documenting. The core radiative transfer routines are sourced from the Fortran SHDOM (Spherical Harmonic Discrete Ordinate Method) code by Frank K. Evans. I am particularly grateful for the guidance of Amit Aides in the initial stages of this project and the wonderful rigorous work of Jesse Loveridge.



  1. Aviad Levis, Anthony B. Davis, Jesse R. Loveridge, and Yoav Y. Schechner "3D cloud tomography and droplet size retrieval from multi-angle polarimetric imaging of scattered sunlight from above", Proc. SPIE, Polarization Science and Remote Sensing X, 2021.

  2. Aviad Levis, Yoav Y. Schechner, Anthony B. Davis, Jesse Loveridge, "Multi-View Polarimetric Scattering Cloud Tomography and Retrieval of Droplet Size", Remote Sens. 2020.

  3. Tamar Loeub, Aviad Levis, Vadim Holodovsky, and Yoav Schechner, "Monotonicity Prior for Cloud Tomography" ECCV, 2020.

  4. Amit Aides, Avid Levis, Vadim Holodovsky, Yoav Schechner, Dietrich Althausen, and Adi Vainiger, "Distributed Sky Imaging Radiometry and Tomography", Proc. IEEE ICCP, 2020.

  5. Felipe A. Mejia, Ben Kurtz, Aviad Levis, Íñigo de la Parra, Jan Kleissl, "Cloud tomography applied to sky images: A virtual testbed", Solar Energy, 2018.

  6. Aviad Levis, Yoav Y. Schechner, Anthony B. Davis, Multiple-Scattering Microphysics Tomography, Proc. IEEE CVPR, 2017.

  7. Vadim Holodovsky, Yoav Y. Schechner, Anat Levin, Aviad Levis, Amit Aides, “In-Situ Multi-View Multi-Scattering Stochastic Tomography, Proc. IEEE ICCP, 2016.

  8. Aviad Levis, Yoav Y. Schechner, Amit Aides, Anthony B. Davis, Airborne three-dimensional cloud tomography, Proc. IEEE ICCV, 2015.

  9. Danny Veikherman, Amit Aides, Yoav Y. Schechner, and Aviad Levis,Clouds in The Cloud”, Proc. ACCV, 2014.