GRE Suite: Geo-localization Inference via Fine-Tuned Vision-Language Models and Enhanced Reasoning Chains

TL;DR

Experimental results demonstrate that GRE significantly outperforms existing methods across all granularities of geo-localization tasks, underscoring the efficacy of reasoning-augmented VLMs in complex geographic inference.

Abstract

Recent advances in Visual Language Models (VLMs) have demonstrated exceptional performance in visual reasoning tasks. However, geo-localization presents unique challenges, requiring the extraction of multigranular visual cues from images and their integration with external world knowledge for systematic reasoning. Current approaches to geo-localization tasks often lack robust reasoning mechanisms and explainability, limiting their effectiveness. To address these limitations, we propose the Geo Reason Enhancement (GRE) Suite, a novel framework that augments VLMs with structured reasoning chains for accurate and interpretable location inference. The GRE Suite is systematically developed across three key dimensions: dataset, model, and benchmark. First, we introduce GRE30K, a high-quality geo-localization reasoning dataset designed to facilitate fine-grained visual and contextual analysis. Next, we present the GRE model, which employs a multi-stage reasoning strategy to progressively infer scene attributes, local details, and semantic features, thereby narrowing down potential geographic regions with enhanced precision. Finally, we construct the Geo Reason Evaluation Benchmark (GREval-Bench), a comprehensive evaluation framework that assesses VLMs across diverse urban, natural, and landmark scenes to measure both coarse-grained (e.g., country, continent) and fine-grained (e.g., city, street) localization performance. Experimental results demonstrate that GRE significantly outperforms existing methods across all granularities of geo-localization tasks, underscoring the efficacy of reasoning-augmented VLMs in complex geographic inference. Code and data will be released at https://github.com/Thorin215/GRE.

Figures (15)

• Figure 1: Performance comparison of our reasoning-based GRE versus traditional alignment-based approaches and MLLM baselines on image geo-localization.
• Figure 2: Summary of current image geo-localization model architectures.
• Figure 3: Overview of our GRE framework. The geographical reasoning pipeline begins with data preparation, incorporating automated CoT generation, regular expression matching, and manual filtering. Based on our constructed GRE30K dataset, we employ a post-training procedure that consists of supervised fine-tuning to learn reasoning patterns, followed by two-stage rule-based reinforcement learning to enhance image geo-localization reasoning capabilities.
• Figure 4: Indicators distribution of GREval-Bench.
• Figure 5: A detailed illustration of the evaluation pipeline.