Utilizing ArUco markers to define implement boundaries

Abstract

John Deere and Blue River Technology’s autonomous tillage system combines multidisciplinary efforts and cutting-edge technology to achieve Level 5—Unsupervised Autonomy. To create this engineering marvel, countless parameters need defined to ensure safe operation of the system; some of these parameters are static, while other of these parameters are dynamic. One particular set of parameters define the tillage implement’s boundaries for the software stack to utilize, and today that parameter is static. However, to meet Deere and Company’s public, lofty ambitions to scale autonomous tillage to farms around the world, the parameters that define the implement boundaries need to become dynamic. This research project provides a potential solution to defining these implement boundary parameters in a dynamic fashion.

By utilizing Augmented Reality University of Cordoba (ArUco) fiducial markers in unison with the existing Blue River Technology software and hardware stack, a set of dynamic implement boundary parameters is defined. ArUco markers are strategically located around the implement for a moment in time, where in a static fashion, the existing camera hardware takes a snapshot of the implement and markers, then utilizes that information to create and store the newly defined implement parameters. After this brief moment in time, the ArUco markers may be removed from the implement, and the system may proceed with the newly defined and stored implement boundary parameters.

This paper provides a background on fiducial markers, describes the process of locating the ArUco markers, and steps through the software developed to implement this boundary parameter definition.