Data checked 2026-08-13. Data may be stale - beyond the review cycle.
Review cycles are documented on the Methodology page. Methodology
IDEA-Research/groundingdino
AI Model Tier A perception_ai Apache-2.0
[ECCV 2024] Official implementation of the paper "Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection"
Engineering Snapshot
Use cases & tasks 4
- Best suited for
-
- UAV perception requiring zero-shot detection of arbitrary objects
- Vision-language grounding of text commands to bounding boxes
- Primary tasks
-
- Text-prompted object detection from drone camera feeds
- Open-vocabulary scene understanding for low-altitude AI
Stack & ecosystem
- Resource type
- AI Model
- Ecosystem
- object-detection · open-world · open-world-detection · vision-language · vision-language-transformer
License & compliance
- License
-
Apache-2.0(Inferred)
Lifecycle & freshness Show
- Maintenance
- Active
- Latest version
- Not recorded
- Last activity
- 2026-08-13
- Last checked
- 2026-08-13
- Verification
- Official confirmed
What It Solves
Enable flexible open-set object detection in UAV/low-altitude imagery using natural language prompts without retraining for closed-set classes.
Primary use cases
- Text-prompted object detection from drone camera feeds
- Open-vocabulary scene understanding for low-altitude AI
Secondary use cases
- Research reproduction of ECCV 2024 open-set detection method
When to Use
Consider when
- Target classes are not known a priori
- Natural language descriptions of objects are available
Verify before adopting
- No official ONNX/TensorRT export documented
- Training data composition not fully disclosed
- Two-stage encoder latency vs closed-set detectors
Start Here
Adoption Checklist
- Needs verification No official ONNX/TensorRT export documented
- Needs verification Training data composition not fully disclosed
- Needs verification Two-stage encoder latency vs closed-set detectors
Each check stays "needs verification" until an official source confirms it; unconfirmed items are never marked verified.
Known Limitations & Unknowns
Known limitations
- Two-stage encoder (image + text) increases latency vs closed-set detectors
- Performance depends on text prompt quality
- No official ONNX/TensorRT export documented in source
- Training data composition not fully disclosed in repo
Not publicly verified
- Framework
- Model size
- Runtime
- Edge feasibility evidence
- Benchmark source
- Specific training data
Alternatives & Related Tools
Alternatives
How is it used?
Start from the recorded entry points below, then validate against the technical checklist.
- Official site https://arxiv.org/abs/2303.05499
- Repository https://github.com/IDEA-Research/groundingdino
- Documentation https://arxiv.org/abs/2303.05499
Technical checklist
- OK License identified Recorded: Apache-2.0
- OK Maintenance signal Active
- OK Verification status Official confirmed
- OK Source evidence attached 1 source record(s)
- NEEDS REVIEW Latest version recorded Not recorded
Official Links
Metadata & Governance
| License | Apache-2.0 — Inferred |
|---|---|
| Commercial model | Unknown |
| Maintenance status | Active |
| Verification status | Official confirmed — Confirmed via the official repository API responses in SourceRefs below. |
| Latest version | Not recorded |
| Latest release | Not recorded |
| Last activity | 2026-08-13 |
| Last checked | 2026-08-13 |
| First seen | Not recorded |
Dataset facts
Facts above come from the official dataset card only; unconfirmed fields stay unknown.
Related Resources & Dependencies
Recent Activity
New resource: IDEA-Research/groundingdino
New repository resource added by the sprint promote pipeline.
Related Knowledge
Collections
- Aerial Perception Models and Datasets — Source-backed aerial dataset and model repositories for detection and tracking research and prototyping.
- Visual Perception Models and Data — Indexed model and dataset repositories for aerial visual perception research and prototyping.