Learned Image Compression for Earth Observation: Implications for Downstream Segmentation Tasks

2025-12-01

AI Summary

To address the growing transmission and storage burden caused by the surge in satellite remote sensing data, this work investigates task-oriented image compression tailored to downstream semantic segmentation tasks—specifically wildfire, cloud, and building detection. We propose a learned compression model based on discretized mixed Gaussian likelihood and conduct systematic comparisons against JPEG 2000 on multi-channel optical remote sensing imagery. Results demonstrate that our method significantly improves segmentation accuracy while preserving reconstruction quality, especially for large-scale, multi-channel data; however, conventional JPEG 2000 remains competitive on small-scale, single-channel thermal infrared imagery. This study provides the first empirical validation of task-driven learned compression for remote sensing segmentation. Notably, end-to-end joint optimization of compression and segmentation yields no additional gain under the current experimental setup. Our findings establish a new paradigm for efficient remote sensing data processing, emphasizing task-aware compression design over monolithic end-to-end learning.

Technology Category

• Computer Vision: Segmentation
• Data Mining & Knowledge Management: Data Compression
• Machine Learning: Learning on the Edge & Model Compression

Application Category

• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Systems and Infrastructure for Web, Mobile and WoT: Data management and stream processing for Web, mobile and wireless applications

Abstract

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.

Problem

• Evaluates learned compression for Earth observation data transmission
• Compares traditional versus learned compression on segmentation tasks
• Assesses impact on reconstruction quality and segmentation accuracy

Innovation

• Learned compression outperforms JPEG 2000 for large optical imagery
• Traditional codecs remain competitive on small thermal datasets
• Joint end-to-end optimization does not improve performance over standalone