Using machine vision tech to optimise mango harvest

A Central Queensland University (CQ) researcher has used machine vision to optimise mango harvesting, after receiving a $180,000 Industry Research Fellowship to contribute to a cloud-based system for planning the mango harvest. Dr Zhenglin Wang carried out the project with fruit grower-marketer Perfection Fresh Australia (PF Australia) and mango growers in the Burdekin, Mareeba and Childers regions in Queensland.

“Horticulture is Queensland’s second-largest primary industry, employing more than 25,000 people. Farmers and marketers have the challenge of knowing how much crop is on a tree and when it will be ready for harvest. China, Korea, Japan, Hong Kong and Singapore love Queensland mangoes and this project will support export marketing,” said Kate Jones, Minister for Innovation and Tourism Industry Development.

As part of the project, researchers will collect data about fruit maturity and crop load on farms, using light detection and ranging (LiDAR), machine vision and time-of-flight cameras. Dr Wang said current estimates of fruit load are based on a manual count of sample trees, but new technologies that allow assessment of the whole orchard have become robust and cheaper.

“We plan to automate how mango growers estimate the spread of flowering as well as estimation of fruit number, weight and maturity. Machine vision rigs on farm vehicles will move through the orchards mapping fruit attributes and, when the images and data are processed, growers will view results via a mobile phone app. This Fellowship will allow exploration of LiDAR and Time of Flight technologies in such a system,” Dr Wang said.

PF Australia, which holds the rights to the Calypso mango, and the Department of Agriculture and Fisheries have joined the project to test the technologies on trees planted in different densities.

“The intent of the work is to create an online platform that shares information through the Calypso mango supply chain to improve harvest planning, fruit management and marketing,” Dr Wang said.

The research carried out by CQ University will help growers and wholesalers manage seasonal variability. The research is undertaken within a CQ University team supported by a complementary research project that aims to implement these technologies into an automated mango harvester.

“The system he is building could be applicable to many other crops and harvesting situations,” said Francesco Oliveri, Head of ICT at Perfection Fresh Australia.

Queensland and the Northern Territory produce approximately 95% of Australia’s mangoes crop, which is worth $180 million a year.