一种基于深度相机的光合检测机器人定标方法
By employing a calibration method for photosynthesis detection robots based on depth cameras and deep learning, and using LabelMe and Mask-R-CNN to annotate leaf contours, combined with Zhang Zhengyou's calibration method, the real-time performance and hardware requirements of photosynthesis detection robot calibration are solved, achieving efficient leaf pose calibration.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CAS CENT FOR EXCELLENCE IN MOLECULAR PLANT SCI
- Filing Date
- 2024-04-26
- Publication Date
- 2026-07-17
AI Technical Summary
Existing calibration methods for photosynthesis detection robots lack real-time performance on edge computing platforms, have high hardware requirements, require a large amount of preliminary preparation work, and suffer from limited 2D calibration scenarios and significant resource waste.
A calibration method for photosynthesis detection robots based on depth cameras is adopted. The labelme tool is used to annotate the blade outline, and the mask image is obtained by combining the deep learning image segmentation algorithm mask-rcnn. The transformation relationship from camera to blade is constructed, and the transformation relationship from robot end effector to blade is solved by Zhang Zhengyou calibration method, which reduces hardware requirements and improves real-time performance.
This approach improves the real-time performance and accuracy of calibration for photosynthesis detection robots while reducing hardware requirements, thereby reducing the amount of preliminary data preparation and making it suitable for various application scenarios.
Smart Images
Figure CN118298029B_ABST