Application of a Partial Convolutional Neural Network for Estimating Missing Remote Sensing Data (sample data and model outputs)

We implemented a partial convolutional neural network for the imputation of missing remote sensing data. The data set is saved as comma-separated values (CSV) format and comprise of: 1. 54 sample cases of GOCI input data (with NaN regions) for 2018 2. Output of Kriging and Partial CNN imputation of the respective GOCI input samples 3. Topology (used as input for the Partial CNN model), Latitude, and Longitude of the respective GOCI data.