CloudFindr: A Deep Learning Cloud Artifact Masker for Satellite DEM Data

Kalina Borkiewicz, Viraj Shah, J.P. Naiman, Chuanyue Shen, Stuart Levy, Jeffrey D Carpenter

External link (DOI)
View presentation:2021-10-27T15:00:00ZGMT-0600Change your timezone on the schedule page
2021-10-27T15:00:00Z
Exemplar figure, described by caption below
When creating a cinematic scientific visualization for documentary films or museums, visualizations must be not only accurate, but also understandable to audiences of all ages and aesthetically pleasing. Data artifacts are therefore undesirable in these situations. CloudFindr is a method for masking out cloud artifacts in digital elevation model (DEM) data gathered by satellites, which would otherwise create unrealistic and distracting spikes on visualized landscapes. CloudFindr performs image segmentation using a U-Net based machine learning model on GLCM pre-processed data run with different parameters and ensemble voting for final mask results.
Fast forward

Direct link to video on YouTube: https://youtu.be/-Xox4RsgoVw

Keywords

Data Management, Processing, Wrangling, Feature Detection, Extraction, Tracking & Transformation, Machine Learning Techniques, Cartography, Maps, Image and Signal Processing, General Public, Application Motivated Visualization, Geospatial Data, Image and Video Data

Abstract

Artifact removal is an integral component of cinematic scientific visualization, and is especially challenging with big datasets in which artifacts are difficult to define. In this paper, we describe a method for creating cloud artifact masks which can be used to remove artifacts from satellite imagery using a combination of traditional image processing together with deep learning based on U-Net. Compared to previous methods, our approach does not require multi-channel spectral imagery but performs successfully on single-channel Digital Elevation Models (DEMs). DEMs are a representation of the topography of the Earth and have a variety applications including planetary science, geology, flood modeling, and city planning.