Researchers in China have developed a digital twin greenhouse system designed to improve the efficiency of robotic tomato harvesting. The study, titled "Digital twin-driven system for efficient tomato harvesting in greenhouses," was carried out by the Agricultural Information Institute of the Chinese Academy of Agricultural Sciences and published in the journal Computers and Electronics in Agriculture.
According to Chai Xiujuan, chief scientist of the institute's machine vision and agricultural robot innovation team, the research focuses on challenges in automated tomato harvesting, such as limited camera views, fruit occlusion, and complex fruiting patterns. "Efficient and low-damage harvesting remains a major challenge in modern greenhouse tomato production, particularly in dense planting environments," Chai said. "Our study presents a digital twin-driven system for intelligent tomato harvesting."
The system uses a slidable depth camera mounted on a harvesting robot to create a 3D digital twin of the greenhouse. This model captures the spatial distribution and growth states of tomatoes. The research team then built a learning-based framework to optimize robot positioning, arm trajectory planning, fruit selection order, and operational modes.
Trials conducted in Beijing and the Inner Mongolia autonomous region showed reductions in average picking time to 7.4 seconds per fruit and fewer collisions during harvesting.
Team member Lang Yining said the work integrates perception, simulation, and decision-making into one framework. "Traditionally, a depth camera is installed on the robotic arm to capture the picking view and make harvesting decisions," Lang said. "However, such decisions are usually based only on the local field of view from the current camera position, which may contain just a few tomatoes."
"In our approach, a depth camera mounted on a sliding rail scans dynamically to reconstruct the overall structure of the greenhouse plants," he added. "This creates a digital twin of the entire tomato-growing environment and gives the picking decision algorithm a much broader scope for optimization."
The team reports that the system could be applied to other greenhouse crops. They plan to expand their research by using digital twin technology to simulate growth and harvesting environments for additional crop types to support the development and evaluation of harvesting decision algorithms.
Source: ChinaDaily
