Researchers Pursue Automated Staking for Ag Labor Savings

With the North Carolina fresh-market tomato season in full swing, N.C. Plant Sciences Initiative‘s Makerspace team was sweating it out at North Carolina State University’s horticulture field station recently, assembling an imaging system that could be key to lowering the amount of labor it takes to bring fresh, locally grown tomatoes to market.

NC State Extension specialist Emmanuel Torres Quezada, of the Department of Horticultural Science, is working with Makerspace Director Andrea Monteza to automate the labor-intensive process of staking tomatoes and other crops.

Goal: Reduce the Need for Manual Labor

Tomato growers need to place a minimum of 2,000 wood or metal stakes per acre to keep plants from sprawling on the ground, where they’re more susceptible to damage and diseases and harder to harvest.

With growers finding it increasingly difficult to find workers to carry out the repetitive and physically demanding work, Torres and Monteza are collaborating with Fernando Fuentes, a biosystems engineer with the Universidad de Talca in Chile, to develop a tractor-pulled device that’s smart enough to set accurately spaced stakes without damaging young plants or irrigation lines.

A mechanical arm will pick up individual stakes from a storage area, and sensing devices paired with artificial intelligence will tell a hammer or piston precisely when to drive the stakes into the ground.

Next Step: Collecting Images Need to Train AI-Based System

Up to 100,000 images will be needed to train the computer vision system that will be part of the device, and the Makerspace team has been designing and testing a tractor-mounted system capable of capturing and uploading such images from farm fields.

In addition to Torres and Monteza, the project team includes the N.C. PSI’s Brad Lewis, manager of the North Carolina Ag Analytics Platform, Connery Tran, Vy Hoang, John Glenn II and Ryan Babcock.

At a Glance: The MakerSpace Assembles the Imaging System

This post was originally published in Plant Sciences Initiative.