AgroVG: A Large-Scale Multi-Source Benchmark for Agricultural Visual Grounding
Haocheng Li, Juepeng Zheng, Zenghao Yang, Kaiqi Du, Guilong Xiao, Gengmeng Pu, Haohuan Fu, Jianxi Huang, [-1mm] China Agricultural University Sun Yat-sen University Tianjin University, Tsinghua University Southwest Jiaotong University, National Supercomputing Center in Shenzhen · 2026
Read AgroVG: A Large-Scale Multi-Source Benchmark for Agricultural Visual Grounding by Haocheng Li, Juepeng Zheng, Zenghao Yang, Kaiqi Du, Guilong Xiao, Gengmeng Pu, Haohuan Fu, Jianxi Huang, [-1mm] China Agricultural University Sun Yat-sen University Tianjin University, Tsinghua University Southwest Jiaotong University, National Supercomputing Center in Shenzhen (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-22034.txt.
Introduction
Haocheng Li and co-authors present AgroVG, a large multi-source benchmark for agricultural visual grounding, the task of locating objects described in natural language. Reliable evaluation is hard because farm targets are often small, repetitive, occluded, or oddly shaped, and instructions may name one, many, or no objects. AgroVG supplies protocols for box-set matching and query mask coverage. Zero-shot tests of 26 configurations reveal persistent gaps: the best multi-target Set-F1 reaches only 0.35, and the best positive-query mask success at IoU 0.75 stays below 0.17.
Catalog record
| CREATOR | Haocheng Li, Juepeng Zheng, Zenghao Yang, Kaiqi Du, Guilong Xiao, Gengmeng Pu, Haohuan Fu, Jianxi Huang, [-1mm] China Agricultural University Sun Yat-sen University Tianjin University, Tsinghua University Southwest Jiaotong University, National Supercomputing Center in Shenzhen |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 135625 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.22034