FREE Teledyne FLIR Thermal Dataset for Algorithm Training
This free Teledyne FLIR ADAS Dataset provides fully annotated thermal and visible spectrum frames for the development of object detection systems using convolutional neural networks (CNNs). This data was constructed to encourage research on visible + thermal sensor fusion algorithms ("RGBT") and to empower the automotive community to create safer and more efficient ADAS and driverless vehicle systems.
The ability to sense thermal infrared radiation, or heat, provides both complementary and distinct advantages to existing sensor technologies such as visible cameras, Lidar, and radar systems. The Teledyne FLIR thermal sensors can detect and classify in challenging conditions including total darkness, most fog, smoke, inclement weather, and glare. When combined with visible light data and distance scanning data from Lidar and radar, thermal data paired with machine learning creates a more comprehensive detection and classification system.
Dataset Specifications
| Content | A total of 26,442 fully annotated frames with 520,000 bounding box annotations across 15 different object categories. |
| Images | 9,711 thermal and 9,233 RGB training/validation images with a suggested training/validation split. Includes 16-bit pre-AGC frames. |
| Videos | 7,498 total video frames recorded at 24Hz. 1:1 match between thermal and visible frames. Includes 16-bit pre-AGC frames. |
| Frame Annotation Label Totals | Over 375,000 annotations in the thermal and visible spectrum. |
|
Video Annotation Label Totals |
Over 145,000 annotations in the thermal and visible spectrum |
| Label Categories |
|
| Thermal Camera Specifications |
Teledyne FLIR Tau 2 640x512, 13mm f/1.0 (HFOV 45°, VFOV 37°) |
| Visible Camera Specifications | Teledyne FLIR Blackfly S BFS-U3-51S5C (IMX250) camera and a 52.8° HFOV Edmund Optics lens |
|
Dataset File Format |
|
Accelerate AI Deployment with Conservator
Conservator™, an affordable, cloud-based dataset development software to help perception engineers accelerate AI development using thermal infrared and visible image datasets to train neural networks. Conservator includes dataset workflow tools for annotation, curation, quality assurance, and dataset version control to meet the demands of data scientists in automotive, defense, security, and smart cities applications.
Integrate Thermal Cameras Faster with Prism
Prism AI software framework provides classification, object detection, and object tracking enabling perception engineers to quickly start integrating thermal cameras for advanced driver assistance systems (ADAS) and autonomous vehicle (AV) systems. Developers can use Prism AI as the primary perception software or as reference while taking advantage of camera calibration tools, along with visible-and-thermal fusion and advanced image processing capabilities that offer superior perception capability, especially for pedestrian and animal detection.
Related Products
Related Documents
FLIR Boson Software IDD
FLIR ADK Quickstart Guide
FLIR ADK Datasheet