HPC and Big Data technologies for agricultural information and sensor systems Dr. Gábor ÉLŐ, associate professor Péter SZÁRMES, doctoral student Széchenyi István University, Győr, Hungary
Contents Modern technologies in agriculture and AgroDat.hu project in Hungary Data collection from sensors, surveys Analytics and visualization 2
Big Data technologies Very large amount of data from 3
Hungary in Europe 4
Technologies in agriculture Agriculture and the Hungarian economy Precision farming Agricultural sensors and information systems Cultivated land: 5,86 million ha (63%) 5
AgroDat.hu project Integrated information system Data collections (sensors and surveys) Correlations and forecasts Suggestions (eg. irrigation, spraying) Information portal Collecting information from different databases (FAO database, etc.) Displaying information according to client s needs (semantic search) 6
Project tasks Expert system and IT services (HP) High-volume, continuous agricultural data collection from sensor networks (SZE) HPC data processing and storage (SZTAKI) Data analysis and visualization, decision support system (enet) 7
DATA COLLECTION 8
Soil sensors soil humidity and temperature, electric conductivity Electric conductivity salt content Soil humidity biochemical activity Soil temperature - germination 9
Soil sensors water potential Water potential - water intake and water stream in plants 10
Environmental sensors sunlight Intensity of the Photosyntetically Active Radiation (PAR, 400-700 nm) 11
Environmental sensors spectral reflectance Normalized Difference Vegetation Index (NDVI) leaf area Photochemical Reflectance Index (PRI) photosynthetic performance 12
Environmental sensors relative humidity, air temperature, precipitation, wind Relative humidity Air temperature Vapor pressure Precipitation Wind speed and direction 13
Environmental sensors leaf wetness Leaf wetness and ice formation 14
Sensor system communication Sensors data logger SDI-12 standard Data loggers servers GSM network, M2M SIM management 15
Data collection from surveys Additional data from farmers Import from official farming log? Several shorter data forms about important events 16
Factors of data growth 17
ANALYTICS, VISUALIZATION 18
Agricultural decision support system 1. Identification of important environmental and biological states and processes 2. Selecting sensors and supporting equipment to record data about these states, processes 3. Collecting, storing and communicating field data to computers 4. Data processing and manipulation in order to create useful information and knowledge 5. Displaying information and knowledge in appropriate forms to support decision making 19
Hardware elements Special computer system with tailor-made architecture Fast storage, high amount of data Processor-intensive calculations 20
Hardware elements Special computer system with tailor-made architecture Parallelized calculations, web servers Microserver clusters 21
Planned system configuration 22
Software elements Analytics and visualization Pattern recognition with various algorithms Building an agricultural knowledge base Creating a decision support system Examples from www.fruitionsciences.com: 23
Software elements Search engine Searching textual databases (eg. FAO studies) Semantic search (eg. HP Autonomy) Web portal Sharing information (sensor data, analyses, forecasts, etc.) Search options (news, reports, studies) 24
Conclusion R&D objectives Integrated sensor station Energy-efficient, weather-resistant, scalable (communication) Big Data hardware and software systems for agricultural use Effective processing of high-volume, diverse-type data Large-scale, long-term correlation study Between agricultural results and different environmental parameters 25
QUESTIONS? 26
THANK YOU FOR YOUR ATTENTION! Gábor ÉLŐ E: elo@sze.hu Péter SZÁRMES E: peter.szarmes@sze.hu Széchenyi István University, Győr, Hungary 27