Russian Student Builds AI-Powered Tomato-Harvesting Robot
A young engineer from Novosibirsk has designed an intelligent robotic manipulator that can identify ripe tomatoes, predict harvest timing, and replace manual labor in greenhouses.
Anton Vlasenko, a student at Novosibirsk State University, has developed a robotic manipulator powered by artificial intelligence to automate tomato harvesting in industrial greenhouses. The device not only distinguishes ripe tomatoes from green ones but also predicts the optimal time to pick them—helping farmers plan harvest schedules more efficiently.
The introduction of such automated systems could promote sustainable agricultural growth in Russia and worldwide, increasing productivity and reducing dependence on manual labor.
The robot relies on computer vision algorithms to analyze each fruit’s condition. Instead of a simple red-versus-green classification, the system evaluates ripeness by analyzing image color channels and saturation levels. It can estimate how many days remain until full maturity, ensuring only ripe tomatoes are harvested while leaving the rest for later collection.
According to Novosibirsk State University’s press office, the robot uses a YOLOv8 neural network for precise tomato localization. To prevent fruit damage, the manipulator is equipped with pressure sensors that control grip strength, while ultrasonic sensors help it avoid collisions. The software is written in Python using the OpenCV and NumPy libraries.
The robot’s structure was 3D-printed and consists of 115 components, each carefully post-processed. At this stage, developers are fine-tuning the manipulator’s movement algorithms. The next step is creating a mobile platform that will allow the robot to move autonomously between greenhouse rows.
The idea originated at a hackathon and has already caught the interest of Tolmachevsky Greenhouse Complex, which plans to conduct pilot trials. The system is also expected to be adapted for harvesting other vegetable crops in the future.
