Albury NSW, 2-3 September 2010
Hosted by the Australian Centre for Precision Agriculture and SPAA Precision Agriculture Australia
Summary Report by Dan Bloomer and James Powrie, LandWISE Inc.
Presentations at the symposium tended to refer to practical applications of Precision Agriculture technologies with a research focus. They ranged across topics from grains, to remote sensing, and from livestock to viticulture.
The opening by outgoing SPAA President Mark Branson, referred to population increases, tightening oil supply and environmental regulation, crop losses through climate change and future restrictions on synthetic fertilisers. Precision Agriculture was offered as an opportunity for the farming industry to invest in technologies to address these challenges.
Note: The names credited here are those of the speakers. Other authors also contributed to most of the papers presented. We assume full papers will be made available through SPAA or ACPA websites.
Assessment of sugar cane yield monitoring technology
Troy Jensen- National Centre for Engineering in Agriculture
Systems were trialled for monitoring sugar cane yield. Significant variation was noted between yield monitor types. While we don’t grow sugar cane in NZ there may be implications for forage harvesters.
Are passive sensors passé
Eileen Perry- DPI Victoria
Passive sensors rely on sunlight as the source that is reflected, so there are issues dealing with readings as light conditions change. Active sensors have their own light source, so they get a stable result in changing light conditions – and can be used at night if desired. Passive sensors have a place in remote sensing. Satellites, for example, cannot supply active light so remain passive. Passive sensors are also being used in permanent positions to monitor temporal changes in crop status.
Some links to other information: Tetracam.com and headworkphotonix.com and Quantalab in Spain may be interesting to review. The FLIR handheld imaging sensors for carotenoids and chlorophyll looks interesting: www.psi.cz/products/pocket-sized-instruments/plantpen.
David Lamb- Precision Agriculture Research Group, University of New England
Experiments have been using GPS information to track pasture quality and grazing behavior. By tracking what animals are doing when, i.e, moving, still or eating, they are gaining insight into which areas are most productive. This enables improved understanding for integrated landscape management. Tracking animals shows places where they are not grazing at all. Pasture quality patterns show up, with different pastures preferred morning to afternoon.
RFID animal tags are also being trialled to enable automatic walkover weighing and fleece weight recordings to assess growth achieved in relevant parts of the paddock. RFID eartags with GPS (estimated 3 year battery life) are about 2 yrs from market.
Other studies have investigated training animals to allow remote control using sirens. This will potentially enable isolated stock to be managed and encouraged to move to new paddocks by a farmer anywhere on the globe.
UNEPARG research has now shown that Cropcircle sensors with enhanced LEDs can now be used up to 50m above the ground for improved (and safer) aerial biomass measurement.
Once biomass variability has been mapped using, for example, optical sensors or CDAX pasture meter they believe you can define transects within a field that give you a shortcut to getting whole paddock biomass estimates (truncated biomass transects).
David Lamb also noted LIDAR may augment the use of sensors by adding more detailed knowledge of biomass. David Manktelow’s Hawke’s Bay work with orchard canopies was mentioned.
Why do grapevine yield and quality vary in time and space?
Andrew Hall- National Wine and Grape Industry Centre, Charles Sturt University
Andrew reported that their research showed there are big annual deviations in yield. Data from a normal year predicts variability in other years better than same-season sampling in an off-year. Spatial variation was found to be greatest early and late in season. Understanding the effects on weather induced change on reproductive processes and carbohydrate dynamics within the plant appears to be the key to making accurate productivity forecasts. [See also Rob Bramley’s report]
Remote sensing applications for the Australian cotton industry.
Andrew Robson- Queensland Department of Employment, Economic Development and Innovation.
Much of this focused on review of different resolutions of available remote imagery. Some satellite imagery is relatively coarse but is cheap for large areas so may prove appropriate in the cotton industry. Strategic purchase of imagery at only necessary resolution can save cost.
UAV technology with PA to improve productivity and sustainability of macadamia orchards
Leasie Felderhof- Skyview Solutions Pty Ltd.
Regulations are holding back use of UAVs (model aircraft with cameras) in Australia and in US. A commercial pilot licence is required for UAVs! A macadamia orchard of 100ha was mapped in 4 hours flying time. Wind can limit flight potential.
Image quality can be very high (3cm pixel size) and can be used to show individual leaves and identify early stages of pest attack. This is possible using relatively easily available cameras. New photo stitching technology makes it possible to easily and rapidly create mosaics from individual images.
Preliminary cost analysis suggests that nutrition maps derived from aerial images can lower production inputs and increase profits in the macadamia sector.
Fixing low pH soils and solving drainage problems with PA in NSW and Victoria
Tim Neale- precisionagriculture.com.au
Farmers can gain more value from their GPS purchase by using their guidance system to create farm topography maps. Drainage analysis and cut and fill maps can then be created from these. Software advances allowed them to reduce soil moved to ensure drainage by 90% compared to traditional laser levelling techniques.
High resolution satellite images were being used to identify areas suspected of low pH. They compared targeted sampling to grid sampling and decided that zones drawn from satellite imagery were a good intermediate step to start on variable rate liming. A targeted approach to pH can reduce the number of samples required.
John Deere Industry Update
Andrew Bremner John Deere AMS
A number of new John Deere PA products are being launched for 2011. Remote access to tractors equipped with new communications features, will allow for fleet management, diagnostics and online backup.
An Automated system for rapid in field soil nutrient testing
Craig Lobsey- Australian Centre for Precision Agriculture
This work is bridging the gap between conventional lab sampling and lab analysis and current proximal soil sensors. Wet chemistry is desirable for measuring soil nitrate, and potassium and sodium. But rapid processing is essential for on-the-go in-field analysis. The research goal was to get wet chemistry sample processing done in 30 seconds to make in-field processing economic. They found (the cunning bit) that by carefully measuring the rate of change of the chemical reaction at early stages they were able to predict a result that would be gained by the reaction going to completion (which takes 30+ minutes).
SPAA Precision Ag Association Grower group project
Leighton Wilksch- SPAA Precision Agriculture Association
The Grower Group project is focusing on site specific crop management. SPAA has 16 farm discussion groups running, with three meetings a year are held per group.
SPAA also ran their first annual Advanced Training Courses in PA for advisors and key farmers (70 attendees). People we spoke to said this was stunningly good. We’ll find out if we can arrange a New Zealand course if there is interest.
Wireless Sensor Networks
Rakesh Devadas- RMIT University, Melbourne
Effective calibration of remote sensing by satellites, aeroplanes etc needs ground based data. A system was set up to capture continuous ground based measurements, to give data that coincides exactly with aerial and satellite observations. To collect that data, the researchers built an in-paddock sensor network that was wirelessly connected to a base computer, so that when satellite image received they had the necessary data to calibrate the image. This cost $2K per node.
Identifying scales at which yield and soil attributes are related
Brett Whelan- Australian Centre for Precision Agriculture
This statistical presentation looked at the resolution (density of sampling points) of data collection, and the analysis of that data to explain the causes of variability. The amount of variability seen in data can relate more to the amount of sampling than the actual variability present. Some analyses can show highly correlated results, but only explain a little of the variability. By changing the scales, you may have lower correlation but explain more of the variability.
Integrating temporal variation into the management of spatial variability of Precision Viticulture
Rob Bramley- CSIRO
Work was completed at Stoneleigh Squires vineyard in Marlborough. This was led by Marlborough Wine Research Centre’s Mike Trought with John Paul Praat also involved in the project. They analysed spatial vineyard information gained over several years to quantify relationships. EM appeared useful for describing soil variation but detected very low EC values over a very narrow range due to shallow stony soils. EC was closely correlated with trunk circumference.
They used an array of Cropcircle sensors from side of canopy to derive the Plant Cell Density (PCD) index. This uses the same wave bands (near infrared and red) as NDVI, but is a simple ratio rather than the difference ratio used for NDVI index. In contrast to Australian work, vine vigour assessments did not correlate with yield. The researchers believe this is due to the high degree of bud selection at hand pruning. Australian vines are typically machine pruned.
With assistance from a panel of winemakers they determined a juice score based on soluble solids, juice pH and titratable acidity. They then predicted the change in juice score through time using the proximally sensed vine vigour data. This gave a spatial fruit quality and maturity map of the vineyard changing over time. This allowed them to create a map showing the date on which optimum harvest score was attained.
“Identifying those parts of the vineyard that produce the best fruit, and the date for best balance of flavour and aroma is the holy grail of viticulture research” (Mike Trought)
Onion Precision Agronomy
Trevor Twigden- National Onion Labs
Trevor’s team has been working on nutrient and plant stress impacts on onion pungency, sweetness and shelf life. There is a significant price premium for “mild onions” (called “sweet onions” in USA) which are actually low pungency rather than sweet.
To avoid pungency growers must avoid plant stress as the chemical involved (pyruvic acid) results from a stress response. There is “Mild Onion Certification” in Australia – Australia’s only flavour certified fresh horticultural product on the market. The research has demonstrated that precision agronomy e.g. GPS mapping and GPS generated sample sites with infield onion pungency testing makes it feasible to isolate areas with low pungency onions. The marketing system is fully traceable, so the industry can assure quality is not degraded by untested onions being represented as mild.
A very large amount of spatially referenced sampling has enabled the researchers to identify soil, nutrient and nutrient combination effects on pungency and on storage quality. Pungency and sugars are independent. Industry belief was that higher yields caused low storability, but this work shows good nutrition (at all points across the full site) allows high yields (100+ t/ha) and good storability.
Martin Keye- Topcon
The CropSpec optical sensor has been through numerous upgrades. It is currently being widely used in Montana to manage protein levels in wheat, because protein premiums are at record highs.
New TopCon products have remote communications ability allowing on-line service and support. Farmers can see exactly where each machine is and where it has been, from the office computer. They can send prescription maps directly to the tractor and equipment. Martin sees great future for such connectivity allowing access to weather data etc from the cab. The move to wireless connectivity avoids the need for USB data sticks (thumb drives) or swapping data cards to transfer data – this is seen by many as a vital improvement.
Developing capacity via paddock learning – growers and groups
Simon Craig- Birchip Cropping Group (BCG)
Simon reported that use of PA technology for guidance is widespread among BCG farmers, but use of PA information for input decisions has been rare. Barriers of initial investment and aging gear are being displaced by drivers of reduced input costs and risks. A major limitation is software compatibility between office system and various different equipment protocols.
The BCG has been working with farmers to reduce inputs where possible, delay inputs such as nitrogen until later in the season when seasonal forecasts and yield potential predictions are more reliable. They have taught farmers to use paddock strips: e.g. leaving out an input in a pass that crosses over representative zones then monitoring yield at harvest time. These are very effective for building confidence to push the boundaries with little risk to production. An example relevant to NZ maize, would be leaving out P in starter fertiliser. They noted the benefits of running inter-seasonal nutrient budgets, reflecting previous crop extraction and anticipated yields.
PA education and training modules
Brett Whelan- Australian Centre for Precision Agriculture
Brett and colleagues have prepared a series of training modules based on the 2006 GRDC PA Training Manual. They identify different levels of knowledge sought, and structure the material accordingly. Each module covers a separate aspect of PA, and starts with a general introduction, allows an interested person to access more detailed information, and includes references to very details references should those be wanted. This information is intended to be made freely available, downloadable or CD, and open for use by educators and advisors as they wish to generate their own resources.
Future Trends in PA
James Hassall- Farmer and Nuffield Scholar in Precision Agriculture
James is a very long term PA farmer who built his own auto steer system years ago, and other pieces of equipment since. James presented a view of the future of PA, based on his experience, his Nuffield research and hours sitting in his tractor thinking. He says ISOBUS communication between tractor and equipment is coming. However some manufacturers think it is already becoming redundant as it won’t cope with the required volume of data moving between multiple sensors on the tractor and equipment. He says manufacturers are sharing information and testing compatiblity between gear by holding ‘plugfests’.
James is building a robot to move through his fields. He suggested that in terms of technology, the 80s introduced the personal computer, the 90’s the internet. We are currently seeing miniaturisation, and the next decade will see widespread introduction of robotics. He believes he’ll be able to use his robot for many applications, including soil sampling, protein monitoring, weed spraying and more.
He noted work by David Slaughter using a robot to drop micro-doses of herbicide onto weed leaves. The Danish researchers have combined optical colour and shape technology to identify weed species and apply different chemicals at specific rates using bubble jet printer technology.
James is convinced that wireless data transfer and diagnostics will change farm management. No more having data cards slip out of your top pocket into the spray tank. He is keen to upload maps from office to tractor to manage equipment. He employs a lot of casual labour and remote access can assist the farmer ensure correct operation, and help inexperienced operators easily.
Trade Displays (a sample)
Apogee is a defence company which has moved into agriculture, bringing a different spin on technology. They are highly specialised in satellite radar which penetrates cloud and works at night and can provide a lot of information not available through standard light sensing from satellites. Radar allows identification of different materials, through polarizing of signals. http://www.apogee.com.au
Trimble was marketing their equipment range and their base station network services.
Leica had a display of their ag products along with video on their mining and construction products with laser driven drafting of existing structures and earthworks.
Crop-Optics had a display of Weedseeker and Greenseeker sensors (formally manufactured by Ntech, now owned by Trimble). The display included a moving banner of maize plants being scanned by Greenseeker and showing calculated variable rate against the NDVI score changing as the plant images passed. Contact Scott Jameson
Tim Neale and Andrew Whitlock had their imaging, training packages and consulting services on display.
This is a South African firm now in Australia. They have an easily mounted multi-spectral camera for aircraft. Contact Andrew Coleman