Tomatoes are an important vegetable crop worldwide, but their picking operations have long relied on manual labor, facing problems such as high labor intensity, high costs, and easy fruit damage. In protected tomato cultivation, fruits often grow in clusters with dense branches and leaves. Traditional mechanical picking equipment is prone to environmental interference, resulting in low picking success rates or high damage rates. How to achieve efficient and non-destructive picking of tomatoes in complex planting environments?
The team of Qizhi Yang from the School of Agricultural Engineering, Jiangsu University, in collaboration with Dr. Min M. Addy et al. from the University of Minnesota (USA), has developed a rigid-flexible coupling end-effector integrating a telescopic suction cup and a three-finger gripper, providing an innovative solution to this problem. The related paper has been published in Frontiers of Agricultural Science and Engineering ( DOI: 10.15302/J-FASE-2025643 ).
This study innovatively adopts a collaborative "adhesion-clamping" operation mode. The end-effector first extends the vacuum sucker to adsorb the target tomato, pulls it out from the branches and leaves to avoid interference, and then activates the three-finger gripper to complete the clamping. This step-by-step operation process solves the defects of the traditional single clamping or adhesion methods——the former is easy to damage the fruit due to improper force, while the latter has insufficient stability affected by surface curvature and humidity.
The team determined the optimal operation parameters through 180 sets of experiments: a 270° rotation angle combined with an 8.36 N compound force. Experimental data show that under these parameters, the picking time is reduced by 40% compared with traditional machinery, and the picking cycle for a single fruit only takes 5.4 seconds. By establishing a composite analysis framework of adhesion force (3.58 N) and clamping force (5.94 N), the system achieves an 88% picking success rate, with the fruit damage rate controlled below 0.5%. In the test under simulated greenhouse environment (22 ℃, 60% humidity), the operation efficiency of this equipment is 55% higher than that of manual picking, and it can work continuously to reduce labor intensity.
In response to the complex light conditions in the greenhouse environment, the study adopts the YOLOv5 + HSV hybrid recognition model. By integrating deep learning algorithms and color space analysis technology, the recognition accuracy and speed of nighttime picking are improved. Performance comparison shows that compared with pure adhesion-type (success rate 84%, damage rate 1.4%) and pure clamping-type (success rate 81%, damage rate 7.7%) end-effectors, the new system has significant comprehensive advantages in success rate, damage rate, and operation speed.
This study provides key technical support for tomato picking robots. Its rigid-flexible coupling design concept has been verified by experiments and can effectively meet the needs of mechanized picking of protected tomatoes. With the intensification of the agricultural labor shortage problem, the application of such intelligent equipment will effectively reduce production costs and promote the development of protected agriculture towards high efficiency and precision.
