Research on Gating Fusion Algorithm for Power Grid Survey Data Based on Enhanced Mamba Spatial Neighborhood...

Area of Science:

• Remote Sensing
• Geospatial Analysis
• Computer Vision

Background:

• Power grid surveying requires fusing panchromatic and multispectral imagery for effective power line design.
• Existing deep learning fusion methods often result in spectral information loss and structural blurring.

Purpose of the Study:

• To develop an advanced fusion model for power grid surveying that enhances spatial-spectral feature representation.
• To address spectral distortion and spatial detail loss in remote sensing image fusion.

Main Methods:

• Introduced a novel fusion model incorporating a TransforRS-Mamba module for integrated spatial-spectral feature merging.
• Implemented an improved spatial proximity-aware attention mechanism (SPPAM) for complex object relationship recognition.
• Utilized an optimized spatial proximity-constrained gated fusion module (SPCGF) to enhance key object feature recognition.

Main Results:

• The proposed model demonstrated superior fusion effectiveness compared to 11 existing methods on GF-2 and QuickBird datasets.
• Qualitative and quantitative analyses confirmed significant reductions in spectral distortion and spatial detail loss.
• The model effectively enhances the representation of spatial-spectral features in remote sensing images.

Conclusions:

• The developed fusion model offers a significant advancement for power grid surveying applications.
• The method shows strong potential for improving the accuracy and detail of fused remote sensing imagery.
• Further research is needed to evaluate the model's generalization across diverse datasets and environmental conditions.