Growers could snap up better harvest data thanks to ute-mounted camera
Published:03 February 2014
CQUni researchers Prof Kerry Walsh, Dr Phul Subedi and Dr Alison Payne
Fruit growers need an estimate of crop load to plan their harvest, but with some orchards having tens of thousands of trees, manual counting of fruit can be time-consuming, expensive and inaccurate.
That’s why CQUniversity researchers have been fine-tuning a machine vision system built around a ute-mounted camera.
This technology should enable growers to capture images of all trees in the orchard.
An algorithm in the software can recognise and count the number of fruit on each tree, providing an expected harvest total up to six weeks before the picking season starts.
This early and accurate total is vital for growers sourcing buyers and planning the containers and labour force required for the job. It should also allow a record of yield across seasons, on a per tree basis.
Extending their Horticulture Australian (cat 1 funding) project, CQUniversity researchers Professor Kerry Walsh, Dr Phul Subedi and Dr Alison Payne are providing image sets to research labs in Israel and Thailand to compare count accuracy compared to the algorithms they have developed.
In related news, CQUniversity is offering the new Bachelor of Agribusiness and Food Security, with details via http://bit.ly/1ehmT9Z .