PixDLM: A Dual-Path Multimodal Language Model for UAV Reasoning Segmentation

TL;DR

This paper introduces a new UAV reasoning segmentation task, a large-scale benchmark dataset DRSeg, and a multimodal language model PixDLM to address the unique challenges of UAV imagery analysis.

Contribution

It formally defines UAV reasoning segmentation, creates the DRSeg benchmark dataset, and proposes PixDLM as a baseline model for this task.

Findings

PixDLM achieves strong baseline results on DRSeg.

UAV reasoning segmentation presents unique challenges compared to ground-level scenes.

DRSeg contains 10,000 high-resolution aerial images with Chain-of-Thought QA annotations.

Abstract

Reasoning segmentation has recently expanded from ground-level scenes to remote-sensing imagery, yet UAV data poses distinct challenges, including oblique viewpoints, ultra-high resolutions, and extreme scale variations. To address these issues, we formally define the UAV Reasoning Segmentation task and organize its semantic requirements into three dimensions: Spatial, Attribute, and Scene-level reasoning. Based on this formulation, we construct DRSeg, a large-scale benchmark for UAV reasoning segmentation, containing 10k high-resolution aerial images paired with Chain-of-Thought QA supervision across all three reasoning types. As a benchmark companion, we introduce PixDLM, a simple yet effective pixel-level multimodal language model that serves as a unified baseline for this task. Experiments on DRSeg establish strong baseline results and highlight the unique challenges of UAV reasoning segmentation, providing a solid foundation for future research.