Objective: Automated Pruning System for Vineyards
The ROBOTRIM project, “Mobile Robot for Vineyard Pruning,” aims to design a pruning system that enables both autonomous navigation through vineyard fields and the execution of precise cuts for pruning.
This system is based on robotics and computer vision to automate the pruning of vines. The development consists of several phases, with the objectives being: to develop a computer vision algorithm that detects pruning points according to established criteria and send this information to the robotic arm, which then performs the cut. Another goal is to design a platform to house all the necessary equipment that can navigate the working environment autonomously.
This project has received public funding under the AEI 2017, 2018, and 2022 calls from the Ministry of Industry, Trade, and Tourism, as part of the aid established to support innovative business clusters and improve the competitiveness of small and medium-sized enterprises.
The partners participating in ROBOTRIM I and II include the FEMAC Cluster, CODORNIU, TEYME, Vernis Motors, and ATRIA. The partners in ROBOTRIM III include the FEMAC Cluster, CODORNIU, EARTH ROVER, the INNOVI Cluster, and ATRIA.
ATRIA’s role in the ROBOTRIM project has been to develop the pruning point recognition algorithm and manage the positioning and cutting with the collaborative robot.
Solution: 2D and 3D Computer Vision with a Collaborative Cutting Robot
First, we developed a computer vision algorithm that detects pruning points based on various criteria related to wine quality and grape variety. We used 2D images to detect points of interest and 3D images to determine their 3D coordinates. The algorithm automatically detects these points and translates the coordinates into real-world coordinates, allowing the robot to perform the cut accurately.

The pruning is performed using a collaborative robot equipped with a pneumatic shear as the main cutting tool.
The design includes various pieces of equipment, with the necessary communication protocols established for integration.
Finally, we developed several prototypes of moving platforms, which were tested in field conditions.

As a result, we have developed a prototype that can navigate between vineyard rows, identify points of interest for pruning, and execute cuts using a pneumatic shear mounted on a collaborative robotic arm.
