Vision based macadamia yield assessment

Abstract

Purpose – To describe the prototype macadamia nut yield monitor. Design/methodology/approach – In this paper, the machine vision-based yield monitor for macadamia nut plantations is described. A summary of sensor fusion procedures is presented. Additionally, a summary of current testing progress is provided. Findings – Using vision to count nuts has the potential to revolutionise yield monitoring for the macadamia industry. Additionally, using a vision sensor for in-field location can provide a low cost, highly accurate method of positioning. Tractor (and nut) location can be determined accurate to 12?mm. Practical implications – This project has culminated in the creation of a working prototype harvester. A commercial unit is in the design stage for operation in 2007 harvest season. Originality/value – This paper describes the solution to a particular problem in the macadamia industry, with potential use in wider fields.