agricultural business development with intelligent data analytics · 2018. 6. 14. · based on...
TRANSCRIPT
Agricultural business
development with intelligent
data analytics
M.Sc. Petri Linna
Tampere University of Technology
MIKÄ DATA
11.6.2018 2
Multilevel operational group
Farmers
6 different farms
Companies
Rural advisors
organizations
Aigiant
Research
University of Helsinki
University of Turku
Tampere university of technology
Pyhäjärvi institute
Administrative
organizations
Spiral includes
- Official
coordinating
meetings
- Workshops
- Seminars
- Weekly
project tasks
meetings
- Customer
meetingsMIKÄ
DATA
Acricultural challenges
Viability/profitEnvironmental requirements
Precision agriculture
Amount of manageable data
Closed softwares and services
Internal variations of the block
Weather phenomena
Data ownership
Access to MyData
Satellites
Fieldsensors
Harvestmapper Open data
Historydata
Farming is full of differentkinds of data
Drones
MIKÄ DATA-projectThe project will develop an intelligent data
analysis service that will highlight:
• Variations in soil types
• Nutrient levels (e.g. potassium and
phosphorus)
Project schedule 2017/2-2019/12 (3years)
DATA
DATA SOURCES
Drone Specs
11.6.2018 9
Drone: Airinov solo 3DR
Camera: Parrot SEQUOIA sensor
- 5 spectral bands (red, green, red edge,
near infrared and RGB)
Drone flights with multispectral camera
Parrot
https://www.parrot.com/eu/business-
solutions/parrot-disco-pro-ag
11.6.2018 11
Trimble
Combine harvest,
yield data sensor
Satellite data,
Digital globe, worldview 3
Satellite data, European
Space Agency, Sentinel-2
New data sources
11.6.2018 15
- Claas Isaria
- Davis Vantage pro2
- Drainage diagrams
Next steps
- Soil measurement with x-ray?
Service
Development environment
• Kuva palvelusta
11.6.2018 18
Based on GeoServer
11.6.2018 19
UI and Backends platform
from www.oskari.org
11.6.2018 21
Time series tool for WMS
11.6.2018 22
11.6.2018 23
Data-analytics &
Artificial Intelligence
AI – Neural networks
• Neural network model
– based on field data and yield data of combine
harvester
• PhD student, Petteri Nevavuori developes
this area
11.6.2018 24
Artificial Intelligent
• Why we need so much data?
– AI tells
• What data is important
• Predicts the field yield
– After AI model is ready, data is needed less
11.6.2018 25
Project steps
11.6.2018 26
Data
- Satellite
- Drone
- Combine harvester
- Yara Nsensors
- Soil data
New sources
- Weather stations
- Claas Isaria
- Drainage diagrams
- Precise soil
measurement?
Working of data:
- Correction
- Integration
- Analyses
- Correlations
- AI
ServiceIntelligent data Service:
- Architecture of service
- Use & education guides
- UI & Backend development
Data Sources:
Goal of intelligent service:
- Soil variation
- Nutrient levels (potassium,
phosphorus)
11.6.2018 27
http://drones2018.utu.fi