Investigating the Effect of Spatial Context on Multi-Task Sea Ice Segmentation

2025-07-28

AI Summary

This study investigates how spatial context—specifically receptive field size—affects multi-task sea ice segmentation performance, with emphasis on the alignment between remote sensing resolution (Sentinel-1 SAR vs. AMSR2) and intrinsic sea ice properties (concentration, developmental stage, floe size). Method: We propose a novel Adaptive Atrous Spatial Pyramid Pooling (ASPP) module with variable atrous rates to systematically modulate multi-scale spatial context, embedded within a multi-task segmentation network jointly processing SAR and AMSR2 data. Contribution/Results: Experiments demonstrate that small receptive fields best suit high-resolution SAR for fine-grained concentration estimation; medium-sized fields significantly improve developmental-stage segmentation; and low-resolution AMSR2 channels are indispensable for global sea ice mapping. Grad-CAM visualizations reveal complementary feature learning across modalities. Our work provides a reproducible methodology and empirical evidence for receptive field design and multi-source remote sensing fusion in geospatial deep learning.

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Multimodal Learning

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs

Abstract

Capturing spatial context at multiple scales is crucial for deep learning-based sea ice segmentation. However, the optimal specification of spatial context based on observation resolution and task characteristics remains underexplored. This study investigates the impact of spatial context on the segmentation of sea ice concentration, stage of development, and floe size using a multi-task segmentation model. We implement Atrous Spatial Pyramid Pooling with varying atrous rates to systematically control the receptive field size of convolutional operations, and to capture multi-scale contextual information. We explore the interactions between spatial context and feature resolution for different sea ice properties and examine how spatial context influences segmentation performance across different input feature combinations from Sentinel-1 SAR and Advanced Microwave Radiometer-2 (AMSR2) for multi-task mapping. Using Gradient-weighted Class Activation Mapping, we visualize how atrous rates influence model decisions. Our findings indicate that smaller receptive fields excel for high-resolution Sentinel-1 data, while medium receptive fields yield better performances for stage of development segmentation and larger receptive fields often lead to diminished performances. The fusion of SAR and AMSR2 enhances segmentation across all tasks. We highlight the value of lower-resolution 18.7 and 36.5 GHz AMSR2 channels in sea ice mapping. These findings highlight the importance of selecting appropriate spatial context based on observation resolution and target properties in sea ice mapping. By systematically analyzing receptive field effects in a multi-task setting, our study provides insights for optimizing deep learning models in geospatial applications.

Problem

• Optimizing spatial context for multi-task sea ice segmentation
• Exploring receptive field impact on sea ice property segmentation
• Evaluating sensor fusion benefits in sea ice mapping

Innovation

• Uses Atrous Spatial Pyramid Pooling for multi-scale context
• Combines Sentinel-1 SAR and AMSR2 data fusion
• Analyzes receptive field effects via Gradient-weighted Mapping