GDSR: Global-Detail Integration through Dual-Branch Network with Wavelet Losses for Remote Sensing Image Super-Resolution

2024-12-31

AI Summary

To address the insufficient joint modeling of global structures and local details, as well as the high computational cost for large-scale remote sensing images in super-resolution (RS-SR), this paper pioneers the integration of RWKV into RS-SR. We propose a Global-Detail dual-branch CNN-Transformer parallel architecture and a Global-Detail Reconstruction Module (GDRM), alongside a novel wavelet-domain Wavelet Loss to enhance high-frequency detail recovery. A lightweight feature fusion mechanism is incorporated to achieve high reconstruction fidelity while substantially reducing computational overhead. Experiments on multiple remote sensing benchmarks demonstrate that our method achieves an average PSNR gain of +0.05 dB over HAT, with 37% fewer parameters, 49% lower FLOPs, and 2.9× faster inference speed. Moreover, the Wavelet Loss exhibits strong cross-architecture generalizability.

Technology Category

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

Application Category

• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

In recent years, deep neural networks, including Convolutional Neural Networks, Transformers, and State Space Models, have achieved significant progress in Remote Sensing Image (RSI) Super-Resolution (SR). However, existing SR methods typically overlook the complementary relationship between global and local dependencies. These methods either focus on capturing local information or prioritize global information, which results in models that are unable to effectively capture both global and local features simultaneously. Moreover, their computational cost becomes prohibitive when applied to large-scale RSIs. To address these challenges, we introduce the novel application of Receptance Weighted Key Value (RWKV) to RSI-SR, which captures long-range dependencies with linear complexity. To simultaneously model global and local features, we propose the Global-Detail dual-branch structure, GDSR, which performs SR reconstruction by paralleling RWKV and convolutional operations to handle large-scale RSIs. Furthermore, we introduce the Global-Detail Reconstruction Module (GDRM) as an intermediary between the two branches to bridge their complementary roles. In addition, we propose Wavelet Loss, a loss function that effectively captures high-frequency detail information in images, thereby enhancing the visual quality of SR, particularly in terms of detail reconstruction. Extensive experiments on several benchmarks, including AID, AID_CDM, RSSRD-QH, and RSSRD-QH_CDM, demonstrate that GSDR outperforms the state-of-the-art Transformer-based method HAT by an average of 0.05 dB in PSNR, while using only 63% of its parameters and 51% of its FLOPs, achieving an inference speed 2.9 times faster. Furthermore, the Wavelet Loss shows excellent generalization across various architectures, providing a novel perspective for RSI-SR enhancement.

Problem

• Remote Sensing Image
• Super-resolution
• Computational Cost

Innovation

• RWKV Model
• Wavelet Loss Function
• Dual-Branch Architecture