Pricing
Scoped to your environments. Priced to your volume.
UAV dataset pricing depends on environment complexity, sensor annotation depth, geographic diversity, and volume. We scope every project and send a proposal — no generic price sheet, no wasted calls.
Choose your scope
Three tiers — from first sim-to-real benchmark to ongoing production data pipeline.
Starter Dataset
For teams benchmarking real vs. sim
Single-environment capture brief with standard annotations. For teams on VisDrone or EuRoC MAV who need a custom-briefed, on-demand alternative with richer sensor metadata — and for initial sim-to-real gap analysis or proof-of-concept training runs.
- Single environment type (urban, agricultural, forest, etc.)
- Standard annotation schema (object detection, scene labels)
- Core sensor metadata (GPS, timestamp, resolution, FPS)
- IMU data (gyroscope + accelerometer) on every clip
- Rights documentation and provenance records included
- API or batch download delivery
- QA report with annotation coverage metrics
- Up to 500 clips per brief
Custom UAV Pipeline
For perception and navigation model teams
Multi-environment capture across diverse geographies with enriched sensor metadata and RLHF-ready annotations. For teams building UAV perception, SLAM, and navigation policy pipelines that need real-world data beyond any academic dataset or simulator.
- Multi-environment briefs (up to 8 environment types)
- Enriched sensor data (IMU, gyroscope, accelerometer, altitude)
- RLHF annotation schema (detection confidence, nav decision quality)
- Geographic and terrain targeting worldwide
- AI pre-screening + human review on every asset
- Semantic segmentation and object tracking labels
- Structured dataset delivery for perception model training
- Custom volume (500–50,000+ clips)
- Dedicated QA report with diversity and coverage analysis
Enterprise Volume
For ongoing autonomous systems data infrastructure
Ongoing data contract with a dedicated contributor pool, priority QA, and custom SLA. Built for defence, aerospace, and enterprise robotics teams with continuous data needs.
- Ongoing capture brief contracts (monthly or quarterly)
- Dedicated contributor pool matched to your environments
- Priority QA with 24-hour turnaround on review cycles
- Full model evaluation pipeline integration (Traiin Data)
- BAA, DPA, and data sharing agreement documentation
- EU AI Act Article 10 + defence procurement compliance docs
- Verifiable provenance audit trail per asset
- Executive reporting and dataset quality dashboards
- Custom SLA with delivery guarantees
Included in every UAV dataset
No add-ons for the basics. Every Pioggia dataset ships with the sensor data and documentation autonomous systems teams need.
GPS + IMU on every clip
GPS coordinates, gyroscope, and accelerometer logs captured natively on contributor devices — no post-processing or estimation.
Rights documentation
Per-asset license terms, contributor consent records, and usage permissions — ready for legal and procurement review.
Verifiable provenance
Tamper-evident capture record: who, when, where, what device, at what coordinates. Audit-ready for defence and enterprise procurement.
QA report
Annotation coverage, sensor completeness checks, quality screening, and environmental diversity assessment.
EU AI Act documentation
Article 10 compliance: provenance, representativeness, and bias records with every dataset.
API or batch delivery
Plug into your training pipeline via API, or receive a structured batch download compatible with PyTorch, TensorFlow, and ROS.
Frequently asked questions
Do you publish exact pricing?
Dataset pricing depends on volume, environment types, sensor annotation depth, and turnaround. We scope every project and send a proposal after your first call — no commitment required.
Can I get a sample dataset before committing?
Yes. Request one through the form and we'll send a real UAV sample — clips with GPS, IMU logs, annotations, and QA documentation — so you can evaluate quality and sensor coverage before scoping a full brief.
How quickly can you deliver data for a specific environment?
Turnaround depends on environment complexity, sensor and annotation depth, geographic targeting, and volume. We commit to a specific delivery SLA in your scope proposal, before any work begins.
Do contributors actually capture IMU data?
Yes. Every submission includes accelerometer + gyroscope logs captured at the device's native sampling rate — the same sensor class used for UAV state estimation in deployed systems. It supports 6-DOF pose estimation pipelines and is structurally equivalent to the IMU data your flight controller depends on.
How does this compare to academic datasets like VisDrone or EuRoC MAV?
Academic datasets were captured once, for one research purpose, often under licensing that prohibits commercial training. Pioggia delivers on-demand, custom-briefed data — you define the environments, annotation schema, and volume. Every asset is rights-cleared for commercial training, with full provenance documentation academic datasets don't provide.
What annotation formats do you support?
JSON, JSONL, CSV, and custom schemas. Our RLHF annotations integrate directly with standard reward model training pipelines, and we're compatible with ROS, PyTorch, and TensorFlow dataset loaders.
Can you cover specific geographies or terrain types?
Yes. Our contributor network is global and covers 8 environment types. Specify precise regions, terrain, lighting conditions, and seasonal requirements in your capture brief.
Do you support EU AI Act compliance documentation?
Yes. Every Pioggia dataset ships with the documentation required by Article 10 of the EU AI Act: provenance records, representativeness assessment, bias analysis, and consent records. Enforcement begins August 2026.
Tell us what your UAV needs to understand
We'll scope a custom dataset brief — environments, sensor requirements, annotation schema — and send a proposal. No commitment required.