Vinculación de un sistema electromecánico y visión artificial para la implementación de un prototipo recolector de fresas en un cultivo hidropónico

Abstract

This research project describes the development of an electromechanical system based on artificial vision in order to automate the strawberry harvesting process in hydroponic crops, The mechanical design of the prototype was carried out using SOLID WORKS 2020, obtaining an electromechanical model of two operations, a robotic arm for the harvesting of strawberries and a traction system for the displacement of the prototype through flat fields in controlled environments. Arduino MEGA was employed to control the displacements of both traction and servomotors of the robotic arm, Python programming language was used for the artificial vision, and it can be executed both automatically and manually. The field tests were carried out in an experimental hydroponic crop designed based on the dimensions of hydroponic strawberry structures that are produced in Ecuador with an average height of 55 cm, the hydroponic furrows have 30 plants, each plant has an average of 5 to 6 strawberries. The results obtained after the evaluation of the prototype showed that the most suitable environment for the detection of strawberries is partially cloudy because in this environment the vision correctly differentiates between a ripe and unripe strawberries obtaining an efficiency of 98.5% with an error of 1.5% Concerning the positioning of the robotic arm on the milling cutter based on inverse kinematics, it has an efficiency of 98.14% with an error of 1.86%. The cutting and gathering system express an average time of 51 seconds, being this an extensive time for harvesting. A different type of tool could be experimented with for greater efficiency with respect to time.