TUB

The Turbid Underwater Baseline (TUB) dataset addresses the lack of curated data for vision research in highly turbid underwater environments. It supports tasks such as underwater image enhancement and restoration with ground truth, object detection and segmentation under severe visibility degradation, and multi-view reconstruction. TUB consists of 33 static scenes captured inside a cylindrical water tank using four synchronized GoPro11 cameras. Scene geometry, camera poses, and object… See the full description on the dataset page:

Source: Hugging Face Hub (vapaau/TUB). Metadata imported from the dataset’s Hub tags.