| A
Roundup of Developments from the 6th International Conference on Precision
Agriculture- Minnesota July 2002
Brett Whelan
Australian Centre for Precision Agriculture
The 6th International Conference on PA was
held as usual in the international metropolis of Minneapolis. There
were 549 participants involved in presenting or attending 136 general session
papers, 101 poster presentations, 26 commercial displays, 25 interactive
sessions designed for growers and the agribusiness community, 8 software
demonstration presentations, 4 hands-on workshops and 1 conference dinner
that now allows participants to exchange cash for beer.
The crops under consideration included wheat,
corn, soybeans, cotton, canola, barley, peanuts, grapes, rice, coffee,
sweet potato, oil palm and pasture species. Evidence of the broad
potential impact PA has on world agricultural production.
So breifly, what happened?
• There was a number of soil sensors
shown under development or in production: a real-time soil spectrophotometer;
soil profile compaction sensor; profile compaction/EC sensor; bipolar EM
instruments.
• There were a number of recently
developed/improved crop sensing systems: the Hydro N-Sensor and the Greenseeker
from NTech.
• Much was presented on the use
of remotely sensed imagery in the detection of crop biomass, yield, fertiliser
response/requirement, management zone delineation. Some use in weed
detection.
• A greater increase over previous
years in interest/testing/application of guidance systems
• A whole afternoon spent on various
ways to define management zones, with many still trying to over complicate
the situation and also forgeting the yield/bottom-line aspect. (surprisingly
the latest industry survey suggests grid soil sampling is increasing despite
the progress on management zones and the obvious use for directed soil
sampling)
• Many presentations providing evidence
for the usefulness of real-time soil ECa measurements, however no real
indication that work is being conducted on calibration/interpretation of
the instruments to the soil factors of influence.
There was quite a few commercial companies
offering analysis and interpretation of PA data, with the majority concentrating
on remote sensed imagery. A number of new products will be available
here in Australia in the short-term that may help progress the cause of
PA.
But the most exciting and promising aspect
of the conference was the revelation by the Evans Property orchard company
of its vertically integrated system which combines excellent spatial data
with non-spatial production data at all levels of production. From
the soil/crop interactions through harvest, transport, marketing and business
management - it shows a company totally commited to the full philosophy
of PA. If one company can do it, then there is great hope for the
rest of us.
Commercial
PA Services
Andrew Smart
Precision Cropping Technologies
Precision Cropping Technologies P/L is
an outcome focused business specialising in Precision Agriculture. PCT
was founded by Colin Lye, Broughton Boydell and Andrew Smart to service,
in particular, the cotton industry of Australia. PCT's aim is to provide
outcomes through action plans that allow growers through agronomy and management
to make decisions using technology with a positive ROI.
PCT is a fully commercial P/A consultancy business
dealing in only data acquisition, processing and analysis. PCT fills the
missing link between data collection and its ability to help agronomist
and management make informed decisions. Even though the initial focus was
in cotton, innovative farmers in other industries have recognised the ability
of PCT to produce results. Therefore PCT is gaining a reputation in other
farming sectors interested in gaining value and an ROI from P/A.
This presentation is about the commercialisation
of PA and how important it is to the farming sector to have commercial
business focused on PA, and from a business perspective, how important
it is that research is being turned into commercial reality.
Precision
Ag. Are We Being One Dimensional?
Ian Yule
New Zealand Centre for Precision Agriculture,
Massey University, Palmerston North, NZ.
Roz Buick
Trimble Navigation Ltd, Agricultural Division,
7403 Church Range Blvd, Westminster, CO 80021. USA
Introduction
In the fifteen years that precision agriculture
has been going, what progress has been made? We started by suddenly discovering
that we could be two dimensional in our approach to describing yield.
We no longer had to work with averages, we saw the degree to which yield
varied and identified the clear financial opportunities that this created.
At first there was an anticipation that all we had to do was match our
yield variation to nutrient levels, alter our recipe and we would fix the
situation. Clearly this was somewhat naive.
In the late nineties the advent of affordable
RTK DGPS allowed us to become three dimensional, this offers a huge
range of possibilities. Alex McBratney (2000) “EM and topography,
a great deal of information”. The level of accuracy achievable with
RTK allows us to us GPS for a huge range of tasks in terms of creating
an efficient agricultural resource, describing that agricultural resource
and operating efficiently on it.
Longer term studies such as the one reported
by Craighead (2002) demonstrate the importance of temporal variation, the
fourth dimension. Clearly climate and climatic variation is a very important
driver. In the case of New Zealand, weather has a huge influence on crop
yields. In some years soil moisture deficit may be the most important factor,
while in others the crop protection programme will have a pivotal role
in determining final yield. We are only as strong as the weakest link in
the production chain, that chain will be tested at different points according
to the temporal variation. Management is a time dependant activity,
therefore our management system has to cover all four dimensions. Again
it is clear that the idea of timeliness of operations is hugely important.
The central argument in this paper is that
although the technology has changed significantly in the last 3 – 5 years,
our general perception of it has been largely very one-dimensional,
expressed in the original concept of yield mapping and variable rate treatments.
The way we try and quantify the benefits of the technology is also very
much agriculturally driven and passive in the sense that the gains from
improved performance have not been adequately examined. It is these gains
that will drive our production systems forward from being “mechanised”
to “mechatronic”-”ised”. Precision agriculture is simply the first
step in the development of agromechatronics and the wider use of information
technology in our land based industries.
What the first generation of precision
agriculture tools has taught us.
The first generation of P.A. applications gathered
momentum in the early nineties but these early adopters have not been followed
by an increasing wave of users. Two main reasons appear to be offered by
non-adopters, first, lack of conclusive evidence that the yield mapping
– variable rate control loop will produce a consistent payback. Second;
the level of complexity and the fact that many new and unfamiliar skills
are required.
Moving into unfamiliar areas was not restricted
to farmers, machinery manufacturers saw the opportunity for additional
sales by developing new products and creating their individually badged
P.A. system. In some cases this has appeared to be successful, others have
not. The level of design is improving from a fairly crude first generation.
Efficient and reliable data recording, handling and storage methods were
not adequately considered and this created many difficulties for users.
Systems quickly filled with unnecessary data and lack of reliability caused
user frustration and made others very wary about getting involved.
There does appear to have been a change in attitude in the last few years.
Many companies underestimated the cost of developing such systems and over
estimated the sales. This situation has encouraged them to utilise specialist
companies for their development.
Companies new to the agriculture sector that
had previously worked in other electronic and control application fields
also saw an opportunity. One of the main problems they encountered was
in achieving the level of sales and field support for their products. Farmers
traditionally enjoy a high level of service provided through extensive
dealer networks which are expensive to run. Servicing this dispersed market
with a very low level of service expertise, low user knowledge and an unfriendly
environment for electronic devices has also proved problematic for some.
A number of precision equipment manufacturers are now selling via
large agricultural equipment manufacturers to overcome these issues.
The technology relies on specialist GIS software,
these systems generally represent data in 2 or 3 dimensions and require
at least some level of training to operate them efficiently. They are generally
used to give a physical or statistical description of the data rather than
be totally geared towards management decision making. These systems have
some way to go before they can be used for seamless management decision
making. There are a number of issues surrounding who is the best person
to carry out this task, the manager is the decision maker but is that manager
necessarily the best data processor.
Second generation opportunities
Are we at the same stage with agro-mechatronics
that we were with agricultural mechanisation 70 – 80 years ago?
There has been a significant change in technological
development in the last 3 – 5 years. Many of the problems outlined previously
are being addressed and we are beginning to see an increasing level of
opportunity created through developments in the enabling technologies we
apply to our industries.
The increasing affordability of accurate DGPS
is a good example of where the enabling technology becomes affordable.
Units are now being sold to the land based industries in increasing numbers
for applications such as guidance and machine control. Driver assistance,
“ by lightbar” or other guidance device has been around for a number of
years, as has RTK and DGPS for auto-piloting vehicles. The sub-meter
DGPS receivers will now allow systems such as the Trimble AgGPS EZ-Guide
system to be used much more readily.
Buick (2002) gives an extensive explanation
of DGPS accuracy issues. The difference between static accuracy and dynamic
pass to pass accuracy is explored. A rule of thumb is presented that
“ a GPS receiver provides a relative pass-to-pass accuracy that is 2 to
3 times more accurate (i.e., a lower value) than the same GPS receiver’s
absolute accuracy. “ Buick presented that this rule of thumb is dependent
on a number of variables. Further data that indicated under favourable
GPS conditions using WAAS the factor of improvement from static to dynamic
pass-to-pass accuracy was a factor of five. Taking the Trimble AgGPS EZ-Guide
system from a static horizontal RMS error of 50 cm to a pass-to-pass
error of 9 cm under ideal GPS and WAAS conditions. Also presented
was a list of applications that can be achieved by varying classes of GPS
receiver (Buick 2002). The results are summarised in Table 1.
This second generation of receiver and controller
solutions will allow the greater use of control systems within harvesting
and other field operations. Yule (1999) demonstrated the significant savings
that could be achieved by something as simple as driving in the correct
gear and throttle setting. As saving in cultivating cost of 35% was demonstrated,
differences in cost while operating on slope and in compacted soil were
also observed from a fully instrumented tractor and implement combination.
Fully automated and optimised vehicles may be the third generation of precision
farming tools.
Economic examples
A number of payback examples can be given,
it should be remembered that one receiver can be used for a number of different
purposes as illustrated in table 1.
Potato mechanisation: $800 spent
per ha on mechanisation, (NSW Agriculture Web Page from http://www.agric.nsw.gov.au/reader/2897
) $232 on cultivation. $560 on harvesting. Improving the efficiency of
cultivation equipment will be of major advantage. Opportunities for steering
systems on larger rigs will also help to relieve width limitation while
extending working hours. Being able to reduce these costs by just 10% would
present a considerable opportunity. Reducing timeliness losses will
be an added advantage.
Wilson (2000) also illustrated an increase
in output from applying variable rate seeding, as well as savings
derived from variable rate lime and fertiliser application.
Spraying: Buick (2002) described
a payback period of just 4 months for a US$3,750 AgGPS EZ-Guide 110 system
used on a 1500 broad-acre situation – these results depend on individual
management and farm practices. But generally it is fairly easy to
demonstrate payback within a season or year for manual guidance applications.
Perry et al (2001) reported a saving in herbicide of 42% from spraying
certain species of weed which had a patchy habit. This gave an estimated
saving of in materials cost of between £2 and £18 ha-1 (AUS$6
to AUS$56). Advantages through increased accuracy due to use of guidance
have a direct financial benefit. This can become significant when crops
are repeated sprayed, such as the potato crop where fungicides have to
be regularly applied. Extending the working hours safely into the hours
of darkness is of major benefit to contractors and large farm operators.
Cereal harvesting: Experience with automatic
steering systems in maize harvesters motivated Claas to develop a
system for larger combine harvesters. The full width of larger machines
are not well utilised by drivers and steering the machine takes up
to 60% of the drivers time. High capacity machines now run at 10km/hr in
European conditions, (faster in lighter crops), this level of performance
is very difficult to maintain over long harvest days. Hieronymus
(2000) outlined the advantages as being; Reduced driver stress, better
utilization of the capacity of the machine through better use of the full
width of the head. The driver has more time to maintain machine settings
and best forward speed. Performance is maintained throughout the day. Increase
in accuracy of yield measuring and mapping.
Improved utilisation of the harvester not only
has advantages from the machine cost point of view but it also saves on
timeliness losses as the crop goes beyond its optimum harvest date. Yule
(1986) studied combine harvester fleet size for a range of case study farms.
It was found that the optimum level of crop loss for the largest farm in
the study (556 ha with an average wheat yield of over 10 tonnes ha-1) was
4.4 per cent, (total in-field loss and threshing loss), whereas most drivers
actually operated at a point where the harvester was loosing less than
1 per cent. Smaller farms had reduced losses due to their short season
and had the luxury of slower combine speeds.
Drainage example: The New Zealand Centre
for Precision Agriculture runs a drainage extension service which designs
and installs drainage systems. An RTKDGPS system is used for surveying,
this has a number of advantages over the old laser based system. Not least
is the increase in output, but the fact that the data is in a GIS ready
environment from the start of the process. This allows the design to be
georeferenced, other surveys can be incorporated and information
layers added. The next stage of controlling a trencher has been achieved
in Australia where the Trimble Site Vision ™ system has been utilised.
This allows the georeferenced design to loaded into the machine which is
then controlled from the drainage layout and plan in the Site Vision unit.
This saves time for the operator as well as reducing the risk of error.
One design focus has been to use this technology
combined with EM technology for soil surveying. This data along with accurate
topography gives a far more detailed level of information which takes
full account of spatial variability. Irrigation system requirements can
be calculated to match site variation. This has been taken up with some
enthusiasm by the New Zealand wine industry.
Conclusions
There have been many mistakes made over the
last ten to fifteen years in precision agriculture, but there has also
been considerable progress. There is a growing awareness that we are not
challenging the boundaries of the technology and we perhaps need to refocus
our efforts on areas which have received little attention in the first
phase of precision agriculture.
These second generation DGPS systems offer
greater financial advantage through improved performance and increased
efficiency, they also provide a valuable platform for additional control
applications which will have further financial benefit. This will allow
us to take systems mechanisation to a new level and continue a trend that
was started 70 – 80 years ago and has successfully managed to reduce our
food production costs in that era. New techniques and knowledge will be
integrated into these systems which will help to safeguard our environment,
farmers and the consumer.
The next increase in use of P.A. technology
is likely to come through greater machine control resulting in improved
utilisation of larger machines and thus cost reduction. This trend has
already started with increased sales of guidance systems and increased
interest in auto-pilots. These systems will have further potential for
information gathering and the incorporation of measurement sensors while
vehicles are on the land. Perhaps by Symposium number 10 we will be looking
at another step change in this technology.
References
Craighead, M. Yule, I.J. Variability, Crop
Rotation & Cultural Practises- The NZ Experience. In: Precision Agriculture
in Australasia, Proceedings of the 6th Annual Symposium on Australian Research
& Application. University of Sydney, August 2002
Hieronymus, P. Automatic Steering for Cereal
Harvesters, EurAgEng 2000 University of Warwick, England. July 2 – 7th
, 2000.
McBratney. A. Bishop, T. Digital Elevation
Models as a Crucial Layer for Precision Agriculture In: Precision Agriculture
in Australasia, Proceedings of the 4th Annual Symposium on Australian Research
& Application. University of Sydney, August 2000
Perry, N.H. Lutman, P.J.W. Miller, P.C.H. Wheeler,
H.c. A map based system for patch sprayings weeds –weed mapping. BCPC Conference
Weeds 2001.
Buick, R. GPS Guidance-Making an informed
decision. 6th International Conference on Precision Agriculture, Minneapolis,
MN, 14th – 17th July 2002.
Wilson, J. The application of precision agriculture
in cereal and potato production in Scotland. Proceedings of Precision Tools
for Improving Land Management. Fertiliser and Lime Research Centre. Massey
University, Palmerston North. New Zealand . Feb 14th – 15th , 2001
Yule, I.J. Machinery Utilisation on Arable
Farms, Masters Thesis. (Unpubl). University of Newcastle upon Tyne, UK.
1986
Yule, I.J. Kohnen, G. Nowak, M. A tractor performance
monitor with DGPS capability. Computers and Electronics in Agriculture
23: 155 – 174 (1999)
Class definitions from Buick(2002) for
Figure 1.
Class I: 1 to 2 meter or even 2 to 5
meter absolute accuracy. Differential corrections from radio
beacons or WAAS.
Class II: Sub meter or meter static accuracy.
Differential corrections received from Radio beacons, L- Band
satellite and/or WAAS. Fast update rate 5Hz or higher.
Class III: Decimeter static accuracy 0.1 –
0.3 meter. Typically dual frequency receivers. Some wide area
services are available otherwise local base station.
Class IV: Real Time Kinematic (RTK) Receivers.
Requires dedicated base station or series of base stations (to cover wider
area) or VRS (Virtual Reference System). Based on dual frequency
GPS receiver technology. Centimeter level of accuracy.
|
THE UNIVERSITY OF SYDNEY
