PROYECTO: JUGO DE ARÁNDANO

Transcription

PROYECTO: JUGO DE ARÁNDANO
Advances in Precision Viticulture
Integrating Chilean Agri-Food Sector to the new
millennium
Stanley Best S.
Director Programa
Agricultura de Precisión
INIA-Chile
sbest@inia.cl
F: 56 42 206761
Our vision is associated with consumer requirements
Nutritious foods (functional characteristics)
High in antioxidants and other phytonutrients.
Healthy Food
Chemical waste (reduction -> Effectiveness - Efficiency).
Microbial Health
Good Taste
Taste, maturity, visual characteristics, life, etc.
The technology will influence the way of selling in the short term !!
Whole Chain Assurance:
information transmission with product?
REQUIREMENTS
PRE-
POST
FARM
FARM
GATE
GATE
Growers
Farmers
Food
Packing and
Processing
Retail
Stores
Consumers
Key components
Pre-Farm and Post Farm Gate Standards
o
ISO Guide 65
o
Traceability
o
o
Risk Assessment
o
Residue Monitoring
Actual Problems
Actual
Projected
Seasonal changes and their effect on yields and quality
Inversiones en tecnología en el sector agrícola en EEUU
Outline of topics for new technologies applicable to Agriculture
Systems
Evaluation for
spatial and
temporal
variability of
vineyard.
Variable
application
systems and
logistics
control.
Optical
detection and
estimation.
Continuous
monitoring
systems of
agricultural
variables.
Robotics in
Agriculture
TRANSFORMING DATA INTO INFORMATION FOR FARM MANAGEMENT
DATA
Informationand in agriculture
knowledge
Action are
Problems
in viticulture
in general
associated with multiple variables which acting
dynamically with their environment to adapt every
Knowing how and
Field Data
Temporal
context
existing
change
according
to
a
genetically
Whenpreset
to implement a
Field Prediction
GPS Data
Geographic context
Recommendation.
Computer Vision
patterns.
Raw Sensor
Data
Integrate and
Contextualizing Data
Information Application
Associated with the
Experience
EC Maps and point
Monitoring definition
3D Compaction and soil Maps
Agronomiocal Knowledge
Space - Time Monitoring (few or none)
Usere Integrates
Knowledge:
Variable irrigation
Evaluation (Monitoring)
Diagnostics
Yield and Quality
Management Zones Generation
Redefine management as required in the different zones
Veris
Em-38
Soil Texture
Plant Vigor
Segmentation
Soil Variability Maps
Vigor Maps
Segmentation
Vigor Maps
An environment is an area that has similar
characteristics and deserves to be
handling with homogeneous inputs.
Irrigation Zones
INIA Canopy Analyzer System (ICAS).
High Imagines Resolution
|
Computer
vision
technology
Acquire
Process
Understand
Analyze
Segmentation for yield estimation
Color Space
Decision Tree, neural network
Movil Application (APP)
gr /capture
2012 - 2013 - 2014
6.000
5.000
4.000
3.000
2.000
1.000
0
R² = 0.7991
0
500.000
1.000.000
1.500.000
Píxels with grapes
RMSE = ± 0,206 kg
Kg estimados
3,5 Kg
Vitis Vinifera Case
Distribution curves
•
•
Distribution of Berries Radio
Weight functions for each segment
Xi
Xi
Xi
-3s
-2s
Xi
-1s
m
1s
Berries Radio
2s
3s
CieLab color space
Development of Fruit Growth Charts
Tracking the evolution of fruit?
Projected Growth (final weight)
TIME LAPSE
CAMERA
But Know Yield it is Not Enough We
Need to KnowAUTOMATIC
Yield of What !!!
CAPTURES
ADVANTAJES
HIGHER TEMPORARY (DIARY)
HIGHER RESOLUTION OF BIOLOGICAL CHANGES
LOG AUTOMATED time-lapse.
DIC
ENE
FEB
MAR
Which Quality of Fruit We Have?
Multiplex (Fluorescence)
VIS-NIR Spectrometer
Spectral signatures
Results
Wine A : Premium
Wine B+C : Varietal
Price Difference US$ 10/bottle
Grapes Quality Monitoring
Ferari Index
Enology Evaluation
18
Comparative of Grapes Chemical
Quality characteristics and NDVI
map
Association of soil and plant variables with grapes chemical
quality
characteristics
Topography
NDVI Maps
Exposure Maps
CE Maps (EM38)
Model Integration
Figure. Distribution of probability density of Ferrari
index (Quality Index) in Cabernet Sauvignon variety,
season 2012, 2013 and 201x
Fruit Quality Distribution from Model Developed
3 kinds of grape quality according to the experience gained in the field and developed
defining wines High Quality, Low Quality and Medium Quality which cover classifiers
developed above is then generated.
Figure . Comparison in the estimation of the Model and real Ferrari Index obtained in the
field for two seasons of work (right) and maps of estimation error for both blocks.
Variety
High Quality
Medium Quality
LowQuality
Cabernet Sauvignon
7,01%
10,81%
2,21%
Fruit Quality Distribution from Model Developed
To evaluate the incidence of variables on the fruit quality was a test of significance
proposed by Weiss-Indurkhya
Figure. Significance of Patterns for the Cabernet Sauvignon.
Model Validation
Real Ferari Index maps
Modelled Ferari Index Maps
Main Problem : Equipment Cost (30.000E)
Figure. Ferari Index calibration: Cabernet Sauvignon, training data 2012 and 2013, and comparison of the results
obtained with Ferari Index map. Left graphics: Real Ferari Index maps; Right graphics: Modelled Ferrari Index Maps.
Ense, the classification results were 90 % of well classified areas (R2>0.9 and Mean Absolute Error < 0.1) for
Var. Cabernet Sauvignon.
Future Development
Lab.
analysis
JAZ
Challenges for
viticulture
JAZ
QSI
Lab.
Analisys
To know the fruit quality characteristics could reduce the export
rejection rate.
Chemo metric Association
NIR Spectromete
r
Jaz Modular Optical Sensing Suite
QUALITY GRAPE MODEL
Multiplex (Fluorescence)
QUALITY DEFINITION FOR HARVEST
RGB
Luminancia
Laboratory analyze
(under controlled lighting conditions)
Rojo
Azul
Total Anthocyanins Content Evaluation
Reflactance of Filter 1
6000
R² = 0.7
5900
Total Anthocyanins Content Evaluation
Reflactance of Filter 2
R² = 0.81
Luminicencia (ND)
Reflactancia 850 nm (ND)
5500
5450
5400
5350
5300
5250
5200
5150
5100
5050
5000
4950
1120
Verde
5800
5700
Filtro 5
Filtro 6
Filtro
7
5600
Filtro 8
5500
1140
1160
1180
1200
1220
1240
1260
1280
Contenido Antocianos totales (mgr/lt)
antocianinas tot (BpH1)
Linear (antocianinas tot (BpH1))
1300
5400
1120
1140
1160
1180
1200
1220
1240
1260
1280
Contenido Antocianos totales (mgr/lt)
antocianinas tot (BpH1)
Exponencial (antocianinas tot (BpH1))
1300
Filtro 5
Frecuencia r
30%
25%
20%
15%
10%
5%
0%
1610
1896
2182
2468 2754 3040
Marca de clase
3326
3611
Results
QSI-Imagine
Lab. Results
Antocianos Totales, Filtro 5
R² = 0.8516
Antocianos Totales, Filtro 6
1400
1400
1200
1200
1000
1000
800
800
600
600
400
400
200
200
0
R² = 0.873
0
0
1000
2000
3000
4000
5000
6000
7000
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
Factors associated with grapes chemical characteristics and Stem
Water Potential (plant water stress, Bars).
CLUSTER OF GRAPES CHEMICAL CHARACTERISTICS IN RELATION TO SEASONAL WEIGHTED STEM WATER POTENTIA.
Factors associated with grapes chemical characteristics and Stem
Water Potential (plant water stress, Bars).
Plant Water Stress Evolution
Plano de zonas de riego
IMPORTANT:
More samples.
Reduces the overall error
in the block
water deficit assessment
CWSI 2010-12
18
16
14
Ψ [bar]
12
10
8
R² = 0.7
p<0.007
n= 380
6
4
2
0
0,00
0,20
0,40
0,60
0,80
1,00
CWSI 2010--11
CWSI 2011
CWSI 2010
CWSI 2012
Linear (CWSI 2011)
1,20
Irrigation
Variabilidad
espacial de
suelo
MODELO DE ESTRÉS HÍDRICO
ESPACIAL Y TEMPORAL
Mapas de NDVI e
IAF
Mapas de
potencial
xilemático
(metrics)
SYSTEM DELIVERY
Necesidad
espacial
diaria de
riego
estimadas
en los
próximos
días
Necesidad espacial diaria de riego
estimadas en los próximos días
TAW Suelo y
localización de
sensores de
Humedad del
Suelo
Web Platform integration (user visualization), water deficit spatial and temporal model
Uso de Tecnología Satelital a nivel de
macro escala para seguimiento
Uso Modelo Etr para seguimiento de cultivos
en california
Movable sensor platforms
Effect on labor shortage
Future Expectations in Agriculture
Growth in Robotics
Multi-robot system
Intelligent robot integrating
with PF
Autonomous vehicle
Automatic Guidance system
(Deere: AutoTrack, etc.)
Navigation system
(Deere: Pararel track, etc.)
2000 2007 2014 2021
Irrigation
Future of Crops Monitoring
and Management
(Vineyard Example)
Integrated Vision
Remote Sensing
Variation of management
based monitoring
(alert)/prediction. All in a
interactive platform
Quiality
Yield
Spatial
Predictive
Models
Pre-Harvest Map
Harvest Map
Soil
Winemaking and Viticulturist
diagnosis
wheater
Close
Monitoring
The age of new technological convergences
New technological, Economical and organizational paradigm :
• We need to change from homogenizer logic into diversity logic.
• Mutual empowerment. Advances in some technology drastically accelerate
others.
• The different technologies must be "enabled" to work with others.
• "Synergistic combination" of two or more generic technologies in the search
for common goals.
• It clear that ICT integrated into the internet cloud will be the way to pull
down the different technologies in the hand of the final user but, those
development must integrate a social and technology issues, in order to full fit
the goal of real introduction.
Thank you very much for
your attention