YOLO oriented object detection
Detect objects at any angle with rotated bounding boxes. Ultralytics YOLO26 OBB models fit ships, vehicles, buildings, and text tightly where axis-aligned boxes overlap and fail — ideal for aerial and satellite imagery.
YOLO26 OBB models
YOLO26-obb models predict rotated bounding boxes at 1024-pixel resolution, built for aerial and satellite imagery where objects appear at arbitrary angles. Five sizes trade speed against accuracy, and all fine-tune on custom rotated-box datasets with the standard workflow.
pip install ultralyticsfrom ultralytics import YOLO
# Load a YOLO26 OBB model
model = YOLO("yolo26n-obb.pt")
# Run inference
results = model("image.jpg")| Model | Size (px) | mAP 50-95 (e2e) | mAP 50 (e2e) | CPU ONNX (ms) | T4 TensorRT (ms) | Params (M) | FLOPs (B) |
|---|---|---|---|---|---|---|---|
| YOLO26n-obb | 1024 | 52.4 | 78.9 | 97.7 ± 0.9 | 2.8 ± 0.0 | 2.5 | 14.0 |
| YOLO26s-obb | 1024 | 54.8 | 80.9 | 218.0 ± 1.4 | 4.9 ± 0.1 | 9.8 | 55.1 |
| YOLO26m-obb | 1024 | 55.3 | 81.0 | 579.2 ± 3.8 | 10.2 ± 0.3 | 21.2 | 183.3 |
| YOLO26l-obb | 1024 | 56.2 | 81.6 | 735.6 ± 3.1 | 13.0 ± 0.2 | 25.6 | 230.0 |
| YOLO26x-obb | 1024 | 56.7 | 81.7 | 1485.7 ± 11.5 | 30.5 ± 0.9 | 57.6 | 516.5 |
YOLO26 oriented detection benchmarks from the Ultralytics docs.
Trusted by the world's leading organizations
Our models' impact
Streamline processes across industries with our cutting-edge vision AI models. Speed, accuracy and ease-of-use powered by Ultralytics.
The evolution of Ultralytics YOLO models
See how Ultralytics YOLO evolved from the practical YOLOv5 workflow to edge-ready YOLO26 inference.
Label up to 10x faster with smart annotation
Ultralytics Platform gives you the image annotation tool to build high-quality datasets faster. From smart annotation to precise manual editing, these features are designed to reduce image labeling time without sacrificing quality.
- SAM-powered smart annotation: Masks and bounding boxes in one click.
- Full AI task coverage: Detection, instance segmentation, semantic segmentation, classification, pose, OBB.
- Universal format support: Your choice of YOLO, COCO, and more.
- Team review and versioning: Clear collaboration at every step.
Train on the Best GPUs for Less
26 NVIDIA GPUs starting at $0.24/hr — from Ampere to Blackwell. No markup, no minimums, no commitment.
Explore industry solutions
See how teams apply Ultralytics computer vision across production environments.
Agriculture
Automotive
Healthcare
Logistics
Manufacturing
Retail
Robotics
Agriculture
Automotive
Healthcare
Logistics
Manufacturing
Retail
Robotics
Agriculture
Automotive
Healthcare
Logistics
Manufacturing
Retail
Robotics
Frequently asked questions
Oriented object detection predicts rotated bounding boxes that match each object's actual orientation instead of axis-aligned rectangles. For elongated or densely packed objects — ships in a harbor, vehicles in a parking lot, text lines on a document — rotated boxes fit tightly and avoid the overlap that makes standard boxes ambiguous.
Use OBB when objects appear at arbitrary rotations and their orientation or tight extent matters: aerial and satellite imagery, drone inspection, document analysis, and industrial scenes viewed from above. If objects are mostly upright and orientation is irrelevant, standard object detection is simpler and equally effective.
Ultralytics YOLO26 OBB models are the recommended starting point, pretrained on the DOTA aerial imagery benchmark and available in sizes from nano to extra-large. They share the same training, validation, and export workflow as every other Ultralytics YOLO task.
Annotate rotated boxes in YOLO OBB format and run train mode, or use Ultralytics Platform to annotate, train on cloud GPUs, and evaluate results in the browser. Fine-tuning from DOTA-pretrained weights is the fastest path for most aerial datasets.
Export OBB models with export mode to ONNX, TensorRT, OpenVINO, and other formats. Rotated-box outputs are produced directly by the exported model, keeping inference pipelines simple on edge devices, drones, and servers.
Build oriented detection with Ultralytics
Annotate rotated boxes, train YOLO26 OBB models, and deploy them anywhere with Ultralytics Platform.