Current and Potential Robotic Applications to Improve Cotton Production

Rapid advances in computer vision, deep learning, and autonomous equipment have resulted in commercial robots for autonomous weed control and point to a future where robotic systems will play a larger role in agricultural production.  Over the last two years, Cotton Incorporated has sponsored a small number of projects to look at the potential use of robots for cotton harvest and more recently, weed control. These projects have collectively resulted in an autonomous robot capable of traveling through the field using a combination of proximal sensors and GPS; a small tractor using machine vision that can autonomously harvesting cotton bolls under defoliated conditions; a draft economic model that incorporates regional climate data to compare the value of a robotic harvest system to once-over harvesters; and data from plots in four locations where cotton was hand-harvested twice per week to estimate the potential yield and quality benefits to provide estimates for the economic model.  Initial progress has also been made in evaluating new mechanism for removing seed cotton from the plant, developing an image library of weed species significant to cotton, exploration of genetic manipulations that could facilitate robotic harvest, and concepts around material handling in the field. In addition to providing an overview of recent progress, this paper will also explore other areas of cotton production that could benefit from the use of robotic systems.