Towards the development of an automatic robotic arm positioning system applied to agriculture

Abstract

The increase in process automation is remarkable, because, according to the Tenth Americas Competitiveness Forum, more than 50% of labor activities in Mexico in 2017 could be automated with existing technologies. Likewise, half of the labor activities in 2017 will be automated by 2055. Now, talking about the agricultural sector, the main annual crops in the state of Guanajuato are white grain corn, feed corn, grain sorghum, broccoli, grain wheat and onion [1], in that order where, speaking in terms of total production in tons, the potato produced in Guanajuato represents 13.32 percent compared to the total production of onion, which ranks sixth. Although, according to data from [2], Guanajuato's yield, measured in tons produced per harvested area, is the highest in the whole country, translating into the state producing more potato per area. In this paper, it is intended to propose an automation solution for the white potato classification process through a robotic arm and linear actuators mounted on a conveyor belt, based on an artificial intelligence algorithm applied to a vision system. When a potato is detected, the robotic arm will mark its imperfections, and the system will classify it with the help of the actuators according to the algorithm, thus reducing the degree of uncertainty in the selection process.