DVGBench: Implicit-to-Explicit Visual Grounding Benchmark in UAV Imagery with Large Vision-Language Models

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Zhou, Yue, Chen, Jue, Zhang, Zilun, Huang, Penghui, Ding, Ran, Zou, Zhentao, Gao, PengFei, Wei, Yuchen, Li, Ke, Yang, Xue, Jiang, Xue, Yang, Hongxin, Li, Jonathan
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866911351001055232
author Zhou, Yue
Chen, Jue
Zhang, Zilun
Huang, Penghui
Ding, Ran
Zou, Zhentao
Gao, PengFei
Wei, Yuchen
Li, Ke
Yang, Xue
Jiang, Xue
Yang, Hongxin
Li, Jonathan
author_facet Zhou, Yue
Chen, Jue
Zhang, Zilun
Huang, Penghui
Ding, Ran
Zou, Zhentao
Gao, PengFei
Wei, Yuchen
Li, Ke
Yang, Xue
Jiang, Xue
Yang, Hongxin
Li, Jonathan
contents 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
format Preprint
id arxiv_https___arxiv_org_abs_2601_00998
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle DVGBench: Implicit-to-Explicit Visual Grounding Benchmark in UAV Imagery with Large Vision-Language Models
Zhou, Yue
Chen, Jue
Zhang, Zilun
Huang, Penghui
Ding, Ran
Zou, Zhentao
Gao, PengFei
Wei, Yuchen
Li, Ke
Yang, Xue
Jiang, Xue
Yang, Hongxin
Li, Jonathan
Computer Vision and Pattern Recognition
I.4.9
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
title DVGBench: Implicit-to-Explicit Visual Grounding Benchmark in UAV Imagery with Large Vision-Language Models
topic Computer Vision and Pattern Recognition
I.4.9
url https://arxiv.org/abs/2601.00998