FLEXIGROBOTS Grape Cluster Detection and Tracking in Drone Video

Grapevine phenotyping using drone-based video analysis enables efficient monitoring of canopy structure, phenology, and leaf traits for vine health assessment. Traditional ground surveys are time-consuming, but UAVs equipped with multispectral sensors and computer vision techniques—such as object detection, tracking, and 3D reconstruction—offer scalable alternatives. Methods like YOLO, Mask R-CNN, and TrackR-CNN improve detection accuracy in woody crops, with recent studies achieving 91% precision using VGG19 on RGB-D data. Applications in grapevines and other fruit trees demonstrate the potential of deep learning for automated phenotyping.