Search NASA⌕ Search

DOE OSTI · 2376405

A Weakly-Supervised, Multitask Deep Learning Framework for Shadow Mitigation in Remote Sensing Imagery

Abstract

We propose a weakly-supervised, multitask framework for training a convolutional neural network to solve the problem of cloud shadow mitigation given only cloud and shadow masks as labels. The network minimizes the Wasserstein distance between shadows and their proximal sunlit neighborhoods, generating a supervisory signal directly from within the input image. We extract further utility from the shadow mask through multitask learning by introducing an auxiliary task of shadow segmentation. Our approach is advantageous since it performs mitigation in an end-to-end framework which requires only a shadowed image for inference. We apply this process to the Landsat 8 OLI SPARCS validation data set and demonstrate plausible results.

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Couwenhoven, Scott, Ientilucci, Emmett, Park, Byung, Hughes, David. 2022-07-01. A Weakly-Supervised, Multitask Deep Learning Framework for Shadow Mitigation in Remote Sensing Imagery. https://doi.org/10.1109/igarss46834.2022.9883550. https://searchnasa.com/record/osti%3A2376405/a-weakly-supervised-multitask-deep-learning-framework-for-shadow-mitigation-in-remote-sensing-imager

Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite Search NASA by Indiciae: BibTeX · RIS.