What question did this study set out to answer?

To resolve fragmented predictions and inaccurate boundary delineation during road extraction from high-resolution optical remote sensing imagery using a novel state-space architecture.

August 20, 2026International Journal of Pattern Recognition and Artificial Intelligence

GLP-Mamba: A Global-Local Perception Mamba Model for Road Extraction from Optical Remote Sensing Imagery

View Full Paper

Ask AI

Bookmark

Authors

SSSheng SunShanghai University YCYu ChenGuizhou University XYXu YuGuangdong University of Technology

Discussion

Member takes

Implication

Comparative evaluation demonstrates superior road extraction performance across three benchmark satellite datasets, indicating enhanced boundary precision and structural continuity.

Key Points

To resolve fragmented predictions and inaccurate boundary delineation during road extraction from high-resolution optical remote sensing imagery using a novel state-space architecture.
Built the GLP-Mamba architecture upon RS-Mamba, integrating a Global-Local Perception (GLP) module to refine shallow features and a Structural Consistency Enhancement (SCE) block for residual structural decoding.
Evaluated segmentation performance and performed component ablation studies across three benchmark collections: the DeepGlobe Roads, Massachusetts Roads, and Jilin-1 Satellite datasets.
GLP-Mamba achieved the highest F1 score and Intersection over Union (IoU) metrics among all compared methods across the three benchmark datasets.
Ablation confirmed that long-range context is captured by the OSS block, shallow feature representation is strengthened by the GLP module, and road boundary fragmentation is reduced by the SCE block.

Cite This Study

Sun et al. (2026) studied this question.

synapsesocial.com/papers/6a86b5eb8a91293e6a1cd8be https://doi.org/10.1142/s0218001426590408

View Full Paper

Ask AI

Bookmark

Implication

Comparative evaluation demonstrates superior road extraction performance across three benchmark satellite datasets, indicating enhanced boundary precision and structural continuity.

GLP-Mamba: A Global-Local Perception Mamba Model for Road Extraction from Optical Remote Sensing Imagery

Discussion

Member takes

Implication

Key Points

Cite This Study

GLP-Mamba: A Global-Local Perception Mamba Model for Road Extraction from Optical Remote Sensing Imagery

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider