Rural R&D for Profit Program: Digital technologies for dynamic management of disease, stress and yield [RnD4Profit-15-02-018]
Abstract
Digital technologies have the potential to transform Australian agriculture, with one recent study estimating that their adoption would increase the gross value of agricultural production by $20 billion annually, with the cotton and wine sectors accounting for 5% of that improvement. The objective of the project was therefore to develop novel digital technologies and processes for Australian cotton and grape producers so that they could share in these productivity improvements. Eight sub-projects explored suitable technology to support five research goals:
- Disease management in cotton
- Yield forecasting in vineyards
- Measure vineyard canopy, nutrition, microclimate, grape defects and disease
- Optimise vineyard spraying
- Tools for spatial data analysis
The project achieved its objective to develop innovative digital technologies for the cotton and grape sectors — at least 18 novel devices or processes were produced.
Summary
Digital technologies have the potential to transform Australian agriculture, with one recent study estimating that their adoption would increase the gross value of agricultural production by $20 billion annually, with the cotton and wine sectors accounting for 5% of that improvement. The objective of the project was therefore to develop novel digital technologies and processes for Australian cotton and grape producers so that they could share in these productivity improvements. Eight sub-projects explored suitable technology to support five research goals:
- Disease management in cotton
- Yield forecasting in vineyards
- Measure vineyard canopy, nutrition, microclimate, grape defects and disease
- Optimise vineyard spraying
- Tools for spatial data analysis
Diseases are a major constraint to cotton production in Australia, either directly through lost production (estimated at $27-75 million annually), or through the inability to plant fields thought to be at risk. The cotton industry has historically conducted surveys to separately monitor disease severity and spread in New South Wales and Queensland. While this surveillance has provided industry with disease trends, it has not allowed for a holistic systems approach to potentially improve disease management. This project developed a process to capture georeferenced in-field disease survey data across both states and also collated three years of field pathology data and management practices into a new database. Multivariate analyses of the survey data were performed to determine if the previous crop and cotton trash influenced disease incidence and severity. The key finding was that the previous crop had a significant effect on disease in the subsequent cotton crop. This finding has already been communicated to the industry through updates to best practice management guides, grower meetings and workshops. In addition, a new plant tissue assay was developed for the organism that causes Verticillium wilt – Verticillium dahliae – that is faster than prior methods. The development of this molecular tool now provides industry with the ability to quickly diagnose the presence of Verticillium and its strain, enabling disease risk to be factored into a decision to grow cotton.
Seventeen new technologies were developed for the wine sector, at various stages of industry readiness. A package of ‘precision agriculture tools’ (PAT) has already been released. PAT assists with the analysis of spatial data collected for Precision Agriculture. Initial feedback is that while the tools are useful, the platform interface is still too complex to be adopted widely by producers. It is recommended that the remaining technologies be assessed for commercialisation potential with input from the sector. For those systems with lower technology readiness where this assessment is not yet practical, it is recommended to continue research where there is ongoing support.
The technologies developed to estimate yield were rated highly by the Industry Reference Group whose role it was to provide feedback to the project team. An improvement in the accuracy and efficiency of yield estimation has the potential to generate significant productivity gains through the entire wine supply chain. Current methods used to estimate yield are typically inaccurate and time consuming, with many growers simply replying on an ‘educated guess’. A prototype hand-held NIR instrument and calibration set was successfully developed which can predict bud fruitfulness in Chardonnay and Pinot Noir vines, in the vineyard. This information can be used by growers to tailor pruning to maximise their returns. In another project, a data processing pipeline was developed and combined with machine learning to process video captured from Go-Pro cameras to accurately count inflorescences down a vineyard row and thereby provide a yield prediction shortly after budburst. The same method worked reliably for bunch counting but needs further optimisation for vineyards with heavier canopies. Low-power radar was able to detect fruit irrespective of canopy, but its ability to reliably quantify fruit mass still needs to be demonstrated.
An associated sub-project developed technology to provide on-the-go, non-contact measurement of canopy size and structure, whole vineyard nutrition status, and detection of disease. Real-time canopy measurement is potentially of value as canopy size affects water use and fruit quality. A system using LiDAR was developed to define three-dimensional (3D) canopy structure for whole vine rows and another system successfully measured canopy size across a vineyard using consumer drone imagery with photogrammetry or Go-Pro video imagery. All these devices (except the drone) were mounted on vineyard vehicles and used to image commercial vineyards in collaboration with three wine companies. Individual vine nutritional status was able to be estimated for an entire vineyard using hyperspectral imaging and analysis. Measurement of diseased grapes in vineyards was possible but had many challenges, as did the pre-symptomatic detection of disease on leaves. In contrast, hyperspectral imaging was successfully used to measure bunch rot in addition to other defects such as sunburn, berry shrivel and ‘matter other than grapes’ (MOG) in machine-harvested loads of grapes delivered to a winery. Current methods for the quantification of these defects rely on subjective or semi-quantitative assessment, so the potential of these new tools is of interest to the sector, especially for larger wineries.
A prototype system was developed using acoustic atmospheric tomography and long wave infrared thermography to provide a three-dimensional record of vineyard microclimate at a sub-metre scale. The system was successfully tested in frosty conditions and during days of extreme heat (up to 48°C). The value of the 3D microclimatic information is yet to be quantified by the sector. The goal to optimise vineyard spraying succeeded by developing a kit for sprayers which uses LiDAR and radar to automatically assess canopies and adjust application in response. Chemical savings of 50-90% were achieved in sparse canopies, with a substantial reduction in spray drift. An ‘electronic leaf’ to measure spray coverage was also developed.
The project achieved its objective to develop innovative digital technologies for the cotton and grape sectors—at least 18 novel devices or processes were produced. While collaboration between the sectors was limited, collaboration within each sector and across research organisations nationally and internationally contributed to this success. Wine Australia has already commenced the assessment of each of the wine sector project outputs for potential commercialisation, and this will include consultation with growers, vineyard contractors and winemakers.
This project was funded by the Australian Government Department of Agriculture, Water and the Environment as part of its Rural R&D for Profit program, with co-investment from Wine Australia, Cotton Research and Development Corporation, Horticulture Innovation Australia Limited, LasTek Pty Ltd and the University of South Australia.