from precision agriculture to smart farming capigi 2014/001 … · printscreens ‘making of task...
TRANSCRIPT
From Precision Agriculture to Smart
Farming
Corné Kempenaar Capigi conference Connecting the future, April 2, 2014
Content
Introduction to the topic of today
Some precision agriculture results
● Crop, soil and yield monitoring
● Variable rate application at grid level
● Individual plant treatment (the next level)
Conclusions and challenges
Discussion
Our challenge: More with less
Where we are now
Where we want to be
Ecological
Footprint
FoodDemand
What is precision agriculture ?
Precision agriculture (PA) is a farming management concept
based on observing, measuring and responding to variability
in crops / livestock
Spatial and temporal variation
Detection, decision making and implementation
Sensors, big data, DSS, GNSS, implements, ICT, smart phone,
FMS, Apps, Robotics, .........
Precision agriculture according to Univ. Wisconsin
Precision Agriculture 1.0
‘Controlled traffic farming’
Precision agriculture 2.0, the next step
Scales in precision agriculture
Grid
• In practice
Plant
•On station research
Leaf
•On institute research
Disease
•On institute and university research
National satellite data portal
Without: ● Geometric correction ● Mosaicing ● Cloud detection and filtering ● Atmosferic correction ● End products/indices
Sateliet Data Bands/polarisatie Spatiale Resolutie Temporele Resolutie
Formosat Panchromatisch Multispectraal
Black/white Blue, Green, Red, NIR
2 meter 8 meter
Every 10 days
DMC Multispectraal Green, Red, NIR 22 meter Ca. 3x per week
Radarsat Radar HH+HV polarisatie VV+VH polarisatie
25 meter Every 24 days
Beschikbare satellietbeelden binnen NSD vanaf Maart 2012
DMC versus Formosat
DMC (22 m) NDVI 4-Sep-2012
Formosat (2/8 m) NDVI 1-Sep-2012
Noordoostpolder
Geo-databank of an arable crop rotation Farm De Drieslag, Dronten
Satellite images of winter wheat field
50x winter wheat NDVI curves in Flevoland
NDVI groeicurve is een unieke kwantificering van de gewasontwikkeling per perceel
Ism teeltadviseurs verdere analyse van groeicurves
● Wat is de optimale groeicurve?
● Wat is de relatie met opbrengst?
Potato production and storage
Biomass sensors
Cropscan
Yara N Sensor (ALS and passive)
Terra (Aster sensor), Worldview-2 and TerraSAR-X satellites
90° sensed
area
v
90°sensed
area
Reglone
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
0 20 40 60
Reflection parameter CropScan
Min
imu
m e
ffe
cti
ve
do
se
(l/h
a)
Decision support model
90° sensed
area
v
90°sensed
area
Acknowledgement: Yara, Homburg Holland, PPO
Grid treatment: sensor op cabine of tractor, DSS, VRA on the go, as applied map
Reglone
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
0 20 40 60
Reflection parameter CropScan
Min
imu
m e
ffe
cti
ve
do
se
(l/h
a)
Use VRA in practice: Leipzig, 2009
Yara N-Sensor plus conventional sprayer
MLHD PHK dosing algorithm Reglone, normal, 1x
Ware potatoes BBCH GS 91, 7 august, sunny, hot (30 °C)
Pictures were taken four days after treatment
Fixed rate: 2,5 L Reglone/ha @ 300 L water per
ha
Variabele rate: av. 1,5 L Reglone/ha @ 200 L water
per ha
Figure x. Worldview-2 image, 15-08-2011, Flevoland. Test parcel for variable rate application in yellow outline (above). WDVI image of test parcel (lower-left). Reglone dose instruction map (lower-right)
Grid treatment: based on satellite image,
DSS and task map (2011)
Acknowledgement: TerraSphere, NSO en Mts Zondag
or sensing from UAS,
DSS en task map (2012)
Acknowledgement: TerraSphere, Vd Borne
Formosat image used for potato haulm killing (2013)
Task map potato haulm killing (2013) with Akkerweb
As-applied map Yara N-Sensor
Other precision agriculture applications in
development
Variable seeding/planting density
VRA fertilizer application
Soil pH management
VRA irrigation
Selective harvesting
Nematode management
Disease management
Weed management
......................
wikipedia.org
fieldcopter.eu
Veg. indices versus N-content biomass
Model for Nitrogen application in potato
Screendump of the N-advise model
Fuel use monitoring -> input soil management
(2013)
30
Printscreens
‘making of task
map clay content
-> VRA soil
herbicide weed
control)
Weed detection with near sensing camera’s
c
b
d 1.46 m
0.01 m
a
Conclusions and recommendations
Smart farming requires:
● Accurate data soil, climate, crop and livestock conditions
● Data infrastructure to analyse and apply the (big) data
● DSS that translate spatial and temporal variation in actions
● Implements for site or animal specific actions
● Integration of technology at farm and chain level
● Bottom up development
Cooperation between suppliers and end-users
Yield potential map concept will boost PA
Autonomous platforms in crops will change level of
precision in PA
Towards smart farming tech. (P.A. 2.0)