Learned Image Compression for Earth Observation: Implications for Downstream Segmentation Tasks
This work addresses data transmission and storage challenges for satellite-based Earth observation systems, though it is incremental as it compares existing methods on specific tasks.
The study evaluated learned compression algorithms for Earth observation data, finding that they outperformed JPEG 2000 in reconstruction quality and segmentation accuracy for large-scale optical imagery, but traditional methods remained competitive for smaller thermal infrared datasets.
The rapid growth of data from satellite-based Earth observation (EO) systems poses significant challenges in data transmission and storage. We evaluate the potential of task-specific learned compression algorithms in this context to reduce data volumes while retaining crucial information. In detail, we compare traditional compression (JPEG 2000) versus a learned compression approach (Discretized Mixed Gaussian Likelihood) on three EO segmentation tasks: Fire, cloud, and building detection. Learned compression notably outperforms JPEG 2000 for large-scale, multi-channel optical imagery in both reconstruction quality (PSNR) and segmentation accuracy. However, traditional codecs remain competitive on smaller, single-channel thermal infrared datasets due to limited data and architectural constraints. Additionally, joint end-to-end optimization of compression and segmentation models does not improve performance over standalone optimization.