Decision Automation for Orchards and Vineyards
Type: MBIE Research Programme
Funding and dates: $16,800,000, Oct. 2018 – Sep. 2023
Research collaborators: Universities of Auckland, Canterbury and Otago; Lincoln Agritech and Plant & Food Research
Industry and external partners: New Zealand grower industries, individual growers and New Zealand agricultural manufacturers
Waikato Researchers involved: Professor Mike Duke, Dr Shen Hin Lim, Dr Benjamin McGuinness
Abstract: The research programme will create new technologies for the high-value fruit industries and beyond: Wearable devices and AI-systems that capture spatial and temporal data in fruit crops (with the examples of grapes & apples) as well as from the manual activities of staff at the canopies (e.g. hand thinning, pruning). Decision support systems, based on artificial intelligence and supported with latest technologies in ‘Augmented Reality’ translate this into support of staff or later automated machines to make smart decisions within the orchard and vineyard. The decision-support system will be partnered with a suite of technologies for equipment that human-assist humans or automate manual activities. At the end, the project Human Assist will help to optimise various labour intensive tasks and create a tech-augmented expert workforce in the orchard and vineyard.
The Development of Robotic Asparagus Harvester
Type: Callaghan Innovation R&D Fellowship Grants (in collaboration with RoboticsPlus Limited)
Funding and dates: $90,999, Sep. 2016 to Sep. 2019
Industry and external partners: Tendertips
Waikato Researchers involved: Dr Shen Hin Lim, Professor Mike Duke, Mr Matthew Peebles, Mr Joshua Barnett
Abstract: Asparagus harvesting is a back breaking and labour intensive task.We have developed a proof-of-concept robotic asparagus harvester that first detect and locate harvestable spears using a smart vision system and a cutting and picking tool to harvest the located spears. The successful demonstration was conducted in Los Banos, California, USA in 2019 and this has been filmed to showcase New Zealand technology and innovation to both national and international audiences. This project is currently developing the next iteration of the robotics asparagus harvester, towards the goal of a commercial robotics asparagus harvester.
Multipurpose Orchard Robotics - The future for horticulture
Type: MBIE 2014 Endeavour Fund
Funding and dates: $2,618,598.00 , Oct. 2014 to Sep. 2018
Research collaborators: University of Auckland
Industry and external partners: RoboticsPlus Limited
Waikato Researchers involved: Professor Mike Duke, Dr Chi Kit Au, Mr Joshua Barnett
Abstract: This Programme results in new orchard automation technology which will be used to deliver pollination and harvesting services internationally using a new multi-purpose autonomous robot. The new technology addresses significant issues in the orchard industry: global labour shortages, rising orchard costs, and lack of yield security (for example due to bee Colony Collapse Disorder, which affects fruit pollination, reducing fruit size and yields). The research also benefits the NZ orchard industry through increasing returns and efficiency. It enables precision horticulture that fosters: a) environmental sustainability through best land practices; and b) the integration of innovation on Maori owned orchards. The participating trusts benefit through increased orchard returns and improved skill levels in their orchard workforce. The research develops an Autonomous Multipurpose Mobile Platform (AMMP) modular robot capable of navigating autonomously in orchards and includes vision sensing of flowers and fruit for kiwifruit and apples in orchards, arms and grippers for harvesting kiwifruit and apples, fast-acting directional control mechanisms for precision targeted spraying of pollen and other material on moving flowers, and soft robotic handling of apples and kiwifruit. The research team is a collaboration between Plus Group's RoboticsPlus Ltd, the University of Auckland, Plant and Food Research, and the University of Waikato.
Development of Robust and Reliable Light-based Plant Detection and Characterisation Instrumentation for Precision Agriculture and Environmental Sustainability
Type: Rutherford Discovery Fellowship
Funding and dates: 2019 - 2024
Research collaborators: Unitec Institute of Technology, Massey University, Rua Bioscience
Industry and external partners: Callaghan Innovation
Waikato Researchers involved: Associate Professor Melanie Ooi, Dr Ye Chow Kuang, Professor Mike Duke,
Abstract: The research project will consist of a collaborative approach between biologists, engineers and measurement specialists. Specifically, spectral parameters of targeted herbaceous plant species will be measured using hyper-spectral camera imaging during the life cycle of these plants in order to extract relevant identifying spectral features. The obvious benefit of precisely locating single plants is the ability to measure specific parameters in a plant’s lifecycle, which allows individualised treatment. This directly reduces the associated treatment costs, increases agricultural productivity, and improves environmental sustainability. In the New Zealand context, kaitiakitanga or stewardship/ guardianship is a very relevant concept, whereby any use of organic/inorganic treatment in has an effect on its surrounding environment and therefore has to be clearly justified. The capability to perform an individual plant treatment promotes sustainable land management while recognizing the importance and integrating the kaitiakitanga compliant practices.
The plant life cycle will be verified by the molecular plant scientists through standard laboratory tests. The extracted spectral features will then be used to inform the design of the light spectrum that can best distinguish between different species of plants using commercially available narrow-band light emitting diodes (LED). This will eliminate the need for deploying an expensive and fragile hyperspectral camera into the pasture, since the distinguishing spectral features are embedded into the light design. Thus, the plant identification system is “light-based”. The light-based plant evaluation system is anticipated to require a specially designed light panel, an industrial grade monochrome camera and a microcontroller, all of which are robust to the rigors of farming. The key to the system’s success is to ensure that the designed light can adequately account for known sensitivity ranges of the hyperspectral camera used to measure the spectral parameters and the properties of commercially available LEDs which emit specific bands of light with limited intensity.