Paper ID: 2206.13042

A Strategy Optimized Pix2pix Approach for SAR-to-Optical Image Translation Task

Fujian Cheng, Yashu Kang, Chunlei Chen, Kezhao Jiang

This technical report summarizes the analysis and approach on the image-to-image translation task in the Multimodal Learning for Earth and Environment Challenge (MultiEarth 2022). In terms of strategy optimization, cloud classification is utilized to filter optical images with dense cloud coverage to aid the supervised learning alike approach. The commonly used pix2pix framework with a few optimizations is applied to build the model. A weighted combination of mean squared error and mean absolute error is incorporated in the loss function. As for evaluation, peak to signal ratio and structural similarity were both considered in our preliminary analysis. Lastly, our method achieved the second place with a final error score of 0.0412. The results indicate great potential towards SAR-to-optical translation in remote sensing tasks, specifically for the support of long-term environmental monitoring and protection.

Submitted: Jun 27, 2022