Paired RGB and Thermal Synthetic Images for Adverse Weather Perception

30,781 paired RGB-to-Thermal images address the scarcity of real thermal training data. Robot Flow Labs built this 27GB dataset as part of the ANIMA Perception Suite for robotics. The dataset was last updated in March 2026.

Use Cases

• Training image-to-image translation models to convert RGB images to thermal images using the paired data.
• Developing sensor fusion algorithms by aligning features from the paired RGB and thermal modalities.
• Benchmarking object detection models' performance in adverse weather conditions using the thermal images.
• Creating data augmentation pipelines for night vision systems by leveraging the synthetic thermal data.

Strengths

• Contains 30,781 paired image samples, providing a substantial volume for training.
• Dataset size is 27GB, indicating high-resolution or numerous image files.
• Part of a suite with 96 AI modules, suggesting integration into a larger, tested system.

Limitations

• All data is synthetic, which may not perfectly capture real-world thermal sensor noise and artifacts.
• Specific column names, file formats, and sample data are unknown, hindering initial assessment.
• License information is unavailable, creating uncertainty for commercial use.

Provenance

Source

Robot Flow Labs / AIFLOW LABS LIMITED

Collection Method

Synthetic generation.

Freshness

Last updated March 2026.