Sample frames for this project are available under NDA. Write to bilal@wortel.ai and we will share the relevant before and after pairs.
Challenges
- Distinguishing POA from visually similar grasses and surrounding vegetation.
- Dense and complex field backgrounds made precise boundaries hard to place.
- Overlapping plants and partially visible leaves complicated mask drawing.
- Small or irregularly shaped plants were difficult to segment accurately.
- Polygon boundaries had to stay consistent across varied image conditions.
Related work
Aerial Orthomosaic Segmentation
Whole-site orthomosaics segmented into vegetation, water and bare ground across tiles stitched from multiple flights.
Ground Cover Quadrat Segmentation
2,275 cover-crop plot images annotated in Labelbox: plot corners boxed, then every pixel classed as groundcover, weed, residue, crop or soil.
Black Grass Segmentation
Polygon masks isolating black grass from crops and similar-looking grasses across dense agricultural field imagery.