A high-fidelity multimodal synthetic dataset generated in NVIDIA Isaac Sim. The dataset
features four major crop types (soybean, corn, cotton, and sorghum) across nine growth stages,
obstacles, and weeds. The data is captured by a 4-wheel steering (4WS) acquisition platform. Our
data includes synchronized RGB-D, 32-beam LiDAR, IMU, and semantic segmentation, along with
with a unique Bird’s-Eye View (BEV) perspective.
