Machine vision system for counting macadamia nuts

摘要

A need has been identified for more advanced technology to assist harvesting in the macadamia nut industry. An accurate automated method for counting nuts as they are harvested is required to replace manual labour in varietal strain investigation. A machine vision approach of counting the nuts in real time as they are collected has been chosen. This paper presents a machine vision system that implements fast, efficient algorithms for circle detection and object tracking. Experimental results from initial field tests show that the method is a viable option and further field trials will be undertaken.