information systems for increasing productivity of agriculture · information & telecomm....
TRANSCRIPT
© Hitachi, Ltd. 2014. All rights reserved.
Information Systems for Increasing Productivity of Agriculture
Nov. 5th 2014 Takaomi Nishigaito Director & CTO Hitachi South America Ltda
© Hitachi, Ltd. 2014. All rights reserved.
1 Who am I?
CTO of Hitachi South America and General manager of R&D division.
Came to Brazil from R&D group, Hitachi, Japan to open R&D in Brazil in June 2013.
Focusing on Agriculture and Mining , which is originally strong in Brazil. Planning to make them much more strong using Hitachi`s technologies.
© Hitachi, Ltd. 2014. All rights reserved.
2 Who are we?
Power systems
Automotive systems
Digital media & Consumer products
Social infrastructure & Industrial Systems
Construction machinery
Financial services
Electronics systems & Equipment
High functional materials & Components
*Figures are on a consolidated basis for the FY ending 31 march 2014
JPY 9,616B (USD 93.36B)
14%
10% 7%
7%
13%
8%
8% 3% 12% 18% Information &
Telecommunication Systems
We are HITACHI , having IT × OT
© Hitachi, Ltd. 2014. All rights reserved.
3
Hatoyama-machi, Saitama
Hitachi-shi, Ibaraki
Hitachinaka-shi
Kokubunji-shi Tokyo
Yokohama-shi, Kanagawa
R&D Group
High Functional Materials & Components Group
Infrastructure Systems Group
Information & Telecomm. Systems Group
Power Systems Group
Construction Machinery Group
Automotive Systems Group
Healthcare Group
President
Akasaka, Minato-ku, Tokyo
*as at April 2014
Technology strategy planning
Technology Strategy Office
Energy, Societal, Industrial and Life infrastructure, Materials and Key devices
Hitachi Research Laboratory
Information platform technology, Monozukuri technology
Yokohama Research Laboratory
Contribute to expanding regional business Collaboration with key customers
Overseas research centers
Vision design, Experience design
Design Division
Expand business coverage, New areas in anticipation of future societal needs
Central Research Laboratory
R&D Group in HITACHI
© Hitachi, Ltd. 2014. All rights reserved.
4 R&D Organization
Cambridge
Sophia Antipolis
Farmington Hills
USA (Hitachi America, Ltd.) China (Hitachi (China) R&D Corporation) Europe (Hitachi Europe Ltd.)
Rail systems Automotive
systems Energy systems Healthcare
Service design Adv. Physics
Storage systems Automotive equipment Wireless comm. systems
Big data analytics Design
Platform studies on big data & Solution development
Customer-oriented data science
Asia (Hitachi Asia Ltd.)
Munich
Maidenhead
Social systems Information systems Production technology Home appliances
development Design
Singapore
India (Hitachi India Pvt. Ltd.)
Bangalore
Social infrastructure systems Information &
Telecommunication systems
Purpose: Facilitate collaboration with key customers
Beijing
São Paulo
System for agriculture and mining industry Social infrastructure systems
South America (Hitachi South America, Ltda.)
Santa Clara Japan Shanghai
© Hitachi, Ltd. 2014. All rights reserved.
5 Open innovation
Univ. of Campinas (HISA/Brazil R&D)
Study & preempt future market changes
through discussion with university
Pioneer new markets
Tsinghua Univ. (HCR&D)
Green ICT
Internet of Things
Next-generation cloud
Fudan Univ. (HCR&D)
Joint project with Chinese business
through partnership with university
HCR&D: Hitachi (China) R&D Corporation SIR: Semiconductor Innovation Research Project
EERC: Energy and Environmental Research Center HAS: Hitachi Asia Ltd.
DSI: Data Storage Institute A*STAR: Agency for Science, Technology and Research
HISA: Hitachi South America, Ltda.
IIT Hyderabad (HIL/R&D)
Identify India’s unique energy needs by analyzing energy demand curve
within universities Micro-grid
Diesel generator (2 units)
Genome analysis
A*STAR/DSI (HAS/R&D)
Practical verification of genome analytics
platform
Automotive
RWTH Aachen Univ. (HEU/ERD)
Map data linked chassis control
Technische Universität München
(HEU/ERD)
Engine combustion analysis & simulation Semiconductor
measurement
IBM (HAL/R&D SIR)
Joint research with leading consortium in
adv.semiconductor measurement
Univ. of Birmingham (HEU/ERD)
Carriage abnormality detection using
acoustics measurement
Railway maintenance
EERC (HAL/R&D)
Build-up solution portfolio for Oil&
Gas business
Energy
© Hitachi, Ltd. 2014. All rights reserved.
6 Global collaboration
North America
Europe
Japan
South America
Global Big Data Innovation Laboratory (GBDIL)
US Big Data Lab • Storage solutions • Big data analytics
Central Research Lab. Hitachi Research Lab. Yokohama Research Lab.
Design Division
Overseas research bases leading the expansion of big data business
Brazil R&D Division • Agriculture
European Big Data Lab •Healthcare
• Transportation
Denmark Big Data Lab (to be est. in 2014/4Q)
India
R&D Centre • Analytics workbench
© Hitachi, Ltd. 2014. All rights reserved.
7
≪Past≫
Decision making using historical data.
≪Current≫
Decision making using growth model + climate data
≪Future≫
Decision making using Combined model (Climate data Farm data + Satellite data )
IT in Agriculture fields
© Hitachi, Ltd. 2014. All rights reserved.
8 What can we know from Satellite ? A
mo
un
t o
f li
gh
t re
flecte
d
Wave length of Light (nm)
Wave length of Light (nm)
Abso
rbed L
ight
Chlorophyll A
Chlorophyll B
Healthy
NOT Healthy
Absorb Coastal Blue, Blue and Red for creating chlorophyll
NIR1, NIR2 indicate health
Able to know the growth (Chlorophyll), and health by analyzing spectrum of photos taken from space
© Hitachi, Ltd. 2014. All rights reserved.
9 Signal processing
Vegetation index is calculated using the relations between the intensity of specific frequency bands.
Vegetation Index
Calculation
Wave length (nm) Wave length (nm)
CB
B G
Y R
RE NIR1
NIR2
NDVI (Conventional) : Activities PNVI (HITACHI) : Activities according to Growth Stage * Absorption for chlorophyll gradually changes according to the growth stage. Our method considers this phenomena for increasing accuracy.
© Hitachi, Ltd. 2014. All rights reserved.
10 Harvest Scheduling
Able to know growth condition for Harvest scheduling.
PNVI : Hitachi NDVI : Conventional
Original
Field to be Harvested
Clear and Precise
Growth stage index ,which can be used to determine the harvesting, can be clearly obtained by our method.
© Hitachi, Ltd. 2014. All rights reserved.
11 Quality Control
Able to control quality by controlling fertilization.
Protein quantity (affect the quality of Rice. The less protein causes deliciousness) can be analyzed by using our classification method. Additional fertilization can be considered to control the quality before harvesting.
Original Classified Vegetation Index much
little
medium
medium low
©DigitalGlobe
© Hitachi, Ltd. 2014. All rights reserved.
12 Crop Classification
Able to classify crop variety
Original Estimated Actual
© Digital Globe
Wheat Rice
Carrot Other
Spectrum of crops
420 520 620 720 820 920
Pasture1 Pasture2 Barley Wheat1 Wheat2 Rice1 Rice2
Our spectrum analysis technology can be applied to find very tiny deference to distinguish varieties.
© Hitachi, Ltd. 2014. All rights reserved.
13 Fertility Estimation
Able to estimate soil condition
Actual Total Nitrogen[%]
Est
imate
d T
ota
l N
itro
gen[%
]
0
0,05
0,1
0,15
0,2
0,25
0,3
0,35
0,4
0 0,1 0,2 0,3 0,4
Total Nitrogen Map
Hi
Low ©DigitalGlobe
Good Correlation
Soil condition (total nitrogen) can also be estimated from satellites by observing corresponding bands.
© Hitachi, Ltd. 2014. All rights reserved.
14 Crop Yield Prediction
Able to PREDICT the Yield before harvesting.
1st Step : ≪Learning≫ Accumulate past harvested data, field data and modify growth curve. 2nd Step : ≪Prediction≫ Take one shot photo and predict yield.
Planting Growth Maturation Harvest
Cro
p G
row
th
time
Map update and farming monitoring
Images just before harvest
Satellite image acquisition
Yield data investigation of typical farm lands
Field data survey
Rice growth cycle
1st Step
2nd Step
© Hitachi, Ltd. 2014. All rights reserved.
15 Yield prediction flow
2 steps for yield prediction
Statistical analysis is executed in the predictor using Bayesian Model with the fusion of image and field data.
Map
Farming Datas
Crop map
Satellite image
Satellite Image
Location matching
Field extraction
Yield prediction
Correspond with survey data
Input Yield map
Predictor Learning
Prediction
Bayesian Model
© Hitachi, Ltd. 2014. All rights reserved.
16 Bayesian Model
Accumulated field data , including operation factor and climate factor, and satellite data are combined in Bayesian network for accurate estimation and prediction.
Spectral Signature
(PNVI, NDVI_XY)
Growth Status
Amount of Chlorophyll, etc
Moisture
Biomass (LAI、Height)
EAT: Effective Accumulated Temperature fPAR: Fraction of Photosynthetically Active Radiation LAI: Leaf Area Index
Satellite Data
Water Retaining Capacity
Years from Cultivation
Effective Rainfall
Rainfall fPAR
Growth Environment
Variety
Disease Weed
Fertilization
Soil Fertility
EAT
Operation Factor
Climate Factor
Accumulated Field Data
Planting Date
Yield
© Hitachi, Ltd. 2014. All rights reserved.
17 Estimated results for Soy beans
Very good agreement between actual and estimated yield had obtained in the experiments done in Brazil
Actual Estimated
High Low
© Hitachi, Ltd. 2014. All rights reserved.
18 Future in Agriculture fields
Soil Analysis
Plowing/ Fertilizer
Planting
Fertilizer/ Pesticide
Harvest
Agricultural
Cycle
Planting
Planning Harvest
Control
Cultivation
Planning
Shipping
1. Fertilizer Planning Support
・ Land improvement planning
・ Optimize fertilizer amount
4. Crop Quality Control
・ Crop yields, quality control
5. Harvest Control
・ Optimize harvest time
・ Judge crop loss
Production
Support Cycle
Quality
Control
2. Planting Management
・ Optimize the planting
inspection
・ Support planting plans
3. Additional Fertilizer Planning Support
・ Improve crop yields, stabilize
quality
・ Optimize additional fertilizer
Enhance IT × OT to all the phase in agriculture.