Shared-pattern network improves classification of mixed remote sensing images
Shared-Pattern Collaborative Network for Multi-Source Remote Sensing Data Classification
Computer Vision and Pattern Recognition
Summary
Classifying images from different types of remote sensing data, like hyperspectral images and radar or LiDAR, is challenging because they show different features but also share some common structures. The authors propose a new method called SPCNet that captures shared patterns and adapts how it combines these data sources across different regions in images. Their approach uses special blocks to learn both large-scale common structures and local details dynamically. Tests show this method works better than existing ones on several standard remote sensing datasets.
What this means in practice
- •For geospatial data analysts: Improve land cover classification accuracy by combining hyperspectral and radar or LiDAR data more effectively with shared-pattern modeling.
- •For agricultural monitoring teams: Enhance crop and vegetation mapping by leveraging diverse remote sensing data sources more reliably through spatially adaptive feature fusion.
Authors
Chenfang Yang, Chuanzheng Gong, Feng Gao, Junyu Dong, Qian Du
Abstract
Multi-source remote sensing data classification has received growing attention recently. However, existing methods still struggle to jointly model long-range contextual dependencies and shared spatial patterns across heterogeneous modalities. To address this issue, we propose a Shared-Pattern Collaborative Network (SPCNet) for HSI and SAR/LiDAR joint classification. To enhance cross-modal collaborative learning, Shared-Pattern Collaborative Block (SPCB) is designed by introducing a shared large-kernel operator and a dynamic convolution branch, enabling the network to capture common spatial structures and content-adaptive local features. Furthermore, Region-Adaptive Modality Allocator (RAMA) is developed to assign spatially varying modality weights for reliable feature fusion. Experiments on Houston2013, Augsburg, and Houston2018 datasets demonstrate that SPCNet achieves superior classification performance over several state-of-the-art methods. The code will be made publicly available at https://github.com/oucailab/SPCNet