Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This study addresses the challenge of accurately combining multimodal satellite imagery for ice labeling in the dynamic Antarctic marginal ice zone, where surface features shift rapidly between image acquisitions. By introducing a novel architecture based on mutual information warping, the authors align data from different satellite sensors (Sentinel-1 and MODIS) across various modalities (visible, thermal, radar). The key finding is that this approach significantly enhances classification accuracy and allows for effective sea ice segmentation using sparse expert-labeled data, overcoming the limitations of traditional coarse-resolution maritime ice charts.
Aligning multimodal satellite data using mutual information warping leads to a dramatic boost in sea ice classification accuracy, even with sparse expert annotations.
Multimodal satellite imagery provides complementary information for Earth Observation, but accurately combining heterogeneous sensors remains challenging in dynamic environments. Fast-changing regions, such as the Antarctic marginal ice zone, cannot fully exploit multimodal information from different satellite sensors because surface features move between image acquisitions. This spatial and temporal mismatch challenges effective perceptual grounding, violating the assumption of pixel-level correspondence that underpins most multimodal reasoning and downstream classification pipelines. Antarctic sea ice provides a challenging benchmark due to the rapid, heterogeneous drift of individual ice floes and the differing responses of sea ice to radar, visible and thermal sensing modalities. Accurate, dense supervision of sea ice remains scarce because generating pixel-wise labels requires time-consuming expert interpretation of noisy data, leading to historical reliance on coarse-resolution maritime ice charts for model training. This paper presents a novel architecture based on mutual information warping to align multi-satellite (Sentinel-1 and MODIS platforms) multimodal (visible, thermal, radar) satellite scenes. To demonstrate the approach, we introduce a sparse expert-labeled dataset of 2,088 pixel-wise annotations (7,046 expert point classifications) located at the ice-water margin interface across 43 scenes. Our results demonstrate that spatially grounding and aligning modalities prior to segmentation improves classification accuracy, and enables accurate, dense sea ice segmentation from sparse point-wise supervision.