Synthetic Crop-Weed Image Generation and its Impact on Model Generalization
By: Garen Boyadjian , Cyrille Pierre , Johann Laconte and more
Potential Business Impact:
Creates fake farm pictures to train weeding robots.
Precise semantic segmentation of crops and weeds is necessary for agricultural weeding robots. However, training deep learning models requires large annotated datasets, which are costly to obtain in real fields. Synthetic data can reduce this burden, but the gap between simulated and real images remains a challenge. In this paper, we present a pipeline for procedural generation of synthetic crop-weed images using Blender, producing annotated datasets under diverse conditions of plant growth, weed density, lighting, and camera angle. We benchmark several state-of-the-art segmentation models on synthetic and real datasets and analyze their cross-domain generalization. Our results show that training on synthetic images leads to a sim-to-real gap of 10%, surpassing previous state-of-the-art methods. Moreover, synthetic data demonstrates good generalization properties, outperforming real datasets in cross-domain scenarios. These findings highlight the potential of synthetic agricultural datasets and support hybrid strategies for more efficient model training.
Similar Papers
Enabling Plant Phenotyping in Weedy Environments using Multi-Modal Imagery via Synthetic and Generated Training Data
CV and Pattern Recognition
Helps farmers see weeds in plants using heat.
A Hybrid CNN-ViT-GNN Framework with GAN-Based Augmentation for Intelligent Weed Detection in Precision Agriculture
CV and Pattern Recognition
Helps robots find weeds to save crops.
Synthetic Data for Robust Runway Detection
Machine Learning (CS)
Teaches self-flying planes to land safely.