.png){kind=icon}
.png){kind=icon}
Return
Two–stage multimodal 3D point localization framework for automatic grape harvesting
.png){kind=icon}
OA
AI
DOI：10.1016/j.atech.2025.101062.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• A Two–Stage Multimodal 3D Harvesting Point Localization Framework is proposed. • A depth filtering and irregular depth completion method based on mask guidance is proposed. • A lightweight model for instance segmentation, and pose estimation based on YOLOv11 is proposed. • The processing speed reaches 100.6 FPS on a GPU and 27.6 FPS on a CPU, with only 15.9 GFLOPs. • Achieving 99.2% harvesting point accuracy and a 99.2% recall rate for depth within the 600 mm.
Keywords:
Grape
3D harvesting point
Multimodal
Depth filtering and completion
Lightweight and multi-scale model
Journal
IF:
5.7
Papers:
2.4K
Citations:
2.5K