GeoSeg: Training-Free Reasoning-Driven Segmentation in Remote Sensing Imagery

TL;DR

GeoSeg is a training-free, reasoning-driven segmentation framework for remote sensing imagery that leverages large language models and spatial cues to achieve zero-shot localization without supervision.

Contribution

It introduces GeoSeg, a novel zero-shot, training-free approach combining LLM reasoning with spatial cues for remote sensing segmentation, and provides a new diagnostic benchmark.

Findings

GeoSeg outperforms baseline methods in remote sensing segmentation tasks.

The dual-route prompting mechanism effectively fuses semantic and spatial information.

Component ablations confirm the importance of each part of GeoSeg.

Abstract

Recent advances in MLLMs are reframing segmentation from fixed-category prediction to instruction-grounded localization. While reasoning based segmentation has progressed rapidly in natural scenes, remote sensing lacks a generalizable solution due to the prohibitive cost of reasoning-oriented data and domain-specific challenges like overhead viewpoints. We present GeoSeg, a zero-shot, training-free framework that bypasses the supervision bottleneck for reasoning-driven remote sensing segmentation. GeoSeg couples MLLM reasoning with precise localization via: (i) bias-aware coordinate refinement to correct systematic grounding shifts and (ii) a dual-route prompting mechanism to fuse semantic intent with fine-grained spatial cues. We also introduce GeoSeg-Bench, a diagnostic benchmark of 810 image--query pairs with hierarchical difficulty levels. Experiments show that GeoSeg consistently outperforms all baselines, with extensive ablations confirming the effectiveness and necessity of each component.