Dust Segmentation
The S-Dust Dataset focuses on the topic „Dust Segmentation and Inpainting“. Therefore, it provides over 900 manually labelled images from two fields in northern Bavaria, Germany.
Reference-based Dust Removal v2023
The RB-Dust introduces a reference-based dataset. It consists of 200 paired images, with different backgrounds at different time. [Published at CVPRW]
Single Image Tillage Quality Monitoring
The unpaired dataset aims to remove dust investigating tillage quality monitoring.