Air-ground collaborative fruit selective picking operation system based on double-arm remote control operation

By combining UAV 3D point cloud mapping with CNN neural network, and integrating multi-sensor fusion positioning and cloud server evaluation, selective fruit picking based on dual-arm remote control operation was achieved through air-ground collaborative operation. This solved the problems of low robot operation efficiency, weak perception, and insufficient accuracy, and improved picking efficiency and fruit quality.

CN121844844APending Publication Date: 2026-04-14SOUTH CHINA AGRICULTURAL UNIVERSITY
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-01-07
Publication Date
2026-04-14

AI Technical Summary

Technical Problem

Existing technologies suffer from low robot operation efficiency, weak visual perception and environmental understanding capabilities, insufficient decision-making and operational precision, and a lack of a dynamic scheduling and collaborative framework for robots based on measured terrain data. This results in low fruit recognition success rates, inaccurate maturity assessments, and difficulty in achieving stable and reliable autonomous harvesting in complex environments.

Method used

A dual-arm remote-controlled air-ground collaborative selective fruit picking system is adopted. It combines UAV 3D point cloud mapping with CNN neural network to identify fruit location and maturity. The robot achieves centimeter-level positioning through multi-sensor fusion. The cloud server evaluates the picking success rate and assigns tasks. The operator remotely controls the dual robotic arms to pick fruit through VR glasses. The autonomous picking algorithm is optimized by reinforcement learning.

Benefits of technology

It improves the accuracy of fruit positioning and recognition, enhances harvesting efficiency and fruit quality, reduces labor costs, enables non-destructive harvesting in complex scenarios, dynamically schedules robot cluster operations, and enhances robot adaptability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121844844A_ABST
    Figure CN121844844A_ABST
Patent Text Reader

Abstract

The invention belongs to the technical field of intelligent agriculture, and discloses an air-ground collaborative fruit selective picking operation system based on double-arm remote control operation, which comprises the steps of constructing an air-ground collaborative operation framework, combining unmanned aerial vehicle three-dimensional point cloud mapping with a CNN neural network, accurately identifying the position and maturity of fruits, and selecting the fruits according to the position and maturity of the fruits. The robot realizes centimeter-level positioning through multi-sensor fusion and extended Kalman filtering, and a visual servo mode controls a tail end error at a millimeter level, so that the positioning and recognition precision is improved. Autonomous picking and double-arm teleoperation are cooperated, and the cloud server evaluates the picking success rate through a BP neural network. And an operator realizes lossless picking by virtue of VR remote control and cooperation of double mechanical arms. According to the method, a remote control operation case and an autonomous operation result serve as teaching data, and the adaptive capacity of the robot is improved through reinforcement learning of a continuous iteration model. The labor cost is effectively reduced, the picking efficiency and the fruit quality are improved, and intelligent agricultural picking automation and intelligent upgrading are promoted.
Need to check novelty before this filing date? Find Prior Art