DVGBench: Implicit-to-Explicit Visual Grounding Benchmark in UAV Imagery with Large Vision-Language Models
This work addresses the problem of implicit visual grounding for drone-based agents, which is incremental as it builds on existing large vision-language models by adding a specialized benchmark and method.
The paper tackles the limitation of existing remote sensing visual grounding datasets that rely on explicit referring expressions by introducing DVGBench, a benchmark for implicit visual grounding in UAV imagery across six scenarios, and proposes DroneVG-R1, a model that improves grounding accuracy by converting implicit references to explicit ones using a novel chain-of-thought method.
Remote sensing (RS) large vision-language models (LVLMs) have shown strong promise across visual grounding (VG) tasks. However, existing RS VG datasets predominantly rely on explicit referring expressions-such as relative position, relative size, and color cues-thereby constraining performance on implicit VG tasks that require scenario-specific domain knowledge. This article introduces DVGBench, a high-quality implicit VG benchmark for drones, covering six major application scenarios: traffic, disaster, security, sport, social activity, and productive activity. Each object provides both explicit and implicit queries. Based on the dataset, we design DroneVG-R1, an LVLM that integrates the novel Implicit-to-Explicit Chain-of-Thought (I2E-CoT) within a reinforcement learning paradigm. This enables the model to take advantage of scene-specific expertise, converting implicit references into explicit ones and thus reducing grounding difficulty. Finally, an evaluation of mainstream models on both explicit and implicit VG tasks reveals substantial limitations in their reasoning capabilities. These findings provide actionable insights for advancing the reasoning capacity of LVLMs for drone-based agents. The code and datasets will be released at https://github.com/zytx121/DVGBench