GEOSENSE: AN INFORMATION, COMMUNICATION AND DISSEMINATION ...insait.in/AIPA2012/articles/037.pdf · GeoSense: An Information, Communication and Dissemination System for Decision Support

194 Agro-Informatics and Precision Agriculture 2012 (AIPA 2012)

GEOSENSE: AN INFORMATION, COMMUNICATION AND DISSEMINATION SYSTEM FOR DECISION SUPPORT IN PRECISION FARMING

J. Adinarayana1, D. Sudharsan1*, A.K. Tripathy1, S. Sawant1, S.N. Merchant1, U.B. Desai2, S. Devasekhar1, K. Karandikar1, S. Ninomiya3, M. Hirafuji4 and T. Kiura5

1Indian Institute of Technology Bombay, Mumbai, India 2Indian Institute of Technology Hyderabad, Andhra Pradesh, India

3The University of Tokyo, Nishi Tokyo, Japan 4National Agricultural Research Centre for Hokkaido Region, Kasai-gun, Japan

5National Agricultural Research Centre, Tsukuba, Japan E-mail: *a [email protected]

ABSTRACT

The use of Information, Communication and Dissemination Systems (ICDS) in agricultural decision making processes by the farming community in India are inviting attention to bring the improvement in traditional practices with innovative strategies in an effort to consider recent economic, environmental and social crisis in the rural sector. These innovative strategies gravitate towards a technology-based micro-management precision farming and they require data/information in real-time. Geographical Information Communication Technology (Geo-ICT) for Location Based Services (LBS) and multi-range communication based distributed sensing (proximal) devices (FieldServer and Agrisens) and its associated systems (Flux Towers and FieldTwitter) were used to help the precision farming community with real-time weather/soil/ crop/environmental data/information. This integrated ICDS with open-source tools/techniques, cloud computing and crop models, has been extended to a real-time DSS for information and modeling services christened as GeoSense.

With modified/improved agriculture models and data mining techniques, currently, GeoSense supports the rural extension community in precision irrigation, precision protection and crop yield modeling aspects. In addition, GeoSense has demonstrated its capabilities in climate change, weather profiles/energy flux aspects to calculate and project water needs irrigation scheduling and pest/disease management activities that help in making better farming strategies and to overcome the weather aberrations. With the available FieldServer web-camera, an image processing algorithms has been developed to observe and predict pest incidences in a real-time manner. GeoSense researches are in progress to utilize Open Geospatial Consortium (OGC) standards through Service Oriented Architecture approach for sensory data visualization.

Keywords: ICDS, Wireless Sensor Network, Geo-ICT, Precision Farming.

1. INTRODUCTION

With inclement climatic conditions, and the much aware global climatic changes, the rural community is facing uncertainty in their livelihoods. The situation is more worrisome in fragile semi-arid tropics of India. Some of the dynamic uncertainties that the farming community facing are: crop water requirement, managing pest/diseases, yield aberrations, etc. within the same agro-climatic zones. Consequences of green revolution lead to an evergreen revolution, which will be triggered by farming systems approach that can help to produce more from the available land, water and labour resources, without either ecological or social harm (Kesavan and Swaminathan, 2006). This could be achieved through precision farming, as it proposes to prescribe tailor made management practices. Rural community needs timely dynamic information/assistance to combat the situation. Geographical Information and Communication Technologies (Geo-ICT) (combination of GIS and ICT for location based services and geo-computations) and Sensor Network (SN) (a proximal distributed sensing units pertaining to weather, crop and soil parameters under micro-climatic conditions) are promising real-time information gathering and dissemination technologies towards developing solutions for majority of the agricultural processes on a real-time basis. These tools could be used for sensor web enablement and determining dynamic weather, agriculture and its related information in real time and utilize these parameters for decision making in agricultural processes on a real-time basis. Interlinking these technologies promises to be an interesting combination for generating host of useful information for various applications such as: agriculture, disaster management/mitigation, early warning systems, real time weather and environmental information systems, etc.

Realizing the importance of location based services and dynamic information in precision farming, an attempt has been made to integrate Geo-ICT and sensor network to develop a real time decision support system, called GeoSense.

Proceedings of AIPA 2012, INDIA

GeoSense: An Information, Communication and Dissemination System for Decision Support in Precision Farming 195

2. TOOLS AND SYSTEMS USED

GeoSense is an Indo-Japan initiative on integrating Geo-ICT and WSN for Precision Farming. Two sensor systems (AgriSens and FieldServer) were implemented and real time data was collected and processed at remote stations for agricultural advisory and research based on location based services and sensor data mining.

2.1 FieldServer

FieldServers (FS) is evolved out of many dynamic experiments on agriculture/environmental aspects in 90s. Currently, with 3rd generation FieldServers available, it is a Wi-Fi (long range communication) based self-organizing distributed sensing device (Figure 1) with 24 bit and 24 channels. Embedded board in FS could accommodate the sensors to sense weather, agricultural and environmental parameters such as air-temperature, humidity, relative humidity, CO2, etc. FS transfers sensory data directly to the gateway, a central/server FS in the field, then it is further transmitted over remote server on web (FieldServer, 2011).

Fig. 1: GeoSense Architecture

2.2 Agrisens

AgiSens consist of Stargate (base station) communicating to the various sensor hubs, called Mote, which are placed in different positions and distributed over field (SPANN Lab, 2011). Stargate plays an important role in receiving the data from mote network and transmitting the same to remote server through mobile GPRS network. Each mote has an array of sensors placed at various locations in the experimental plot (Figure 1). Different sensors used in AgriSens are Temperature, Humidity, Leaf Wetness, Soil Moisture, etc. Mote transfers the collected sensor data wirelessly in Zigbee mode (receiving and transmitting) to the base station (Stargate).

2.3 Flux Tower

Two Flux Towers were deployed in maize field (Figure 2) to study the weather profiles and partitioning of energy into different fluxes (Latent Heat Flux, Sensible heat Flux, Ground Heat Flux). Each Flux Tower consists of three sensor

Fig. 2: Flux Tower Deployed in Maize Field Fig. 3: FieldTwitter Deployed in Groundnut Crop


modules with temperature, relative humidity and CO2 concentration sensors at 03 different heights (1 m, 2 m and 3 m). Real time knowledge of weather profiles and energy fluxes allow farming community to calculate water requirement (ET), irrigation scheduling, etc. Flux tower sensors were embedded with FieldServer Engine (FSE) board (one of the components of FS) and are in parallel connection with FS with registered jack (RJ) 45 (RJ45) connectors. The connected FS collects and transmits sensory data to the designated server in the same manner as FS.

2.4 Filed Twitter

FieldTwitter (FT) (Hirafuji et al., 2011) comprises Ardunio (Ardunio, 2012) then transmits sensory data to the Internet clouds through Fon (Fonnera, 2012). Ardunio is attached with an external handmade soil moisture sensor (probe) at a depth of 15 cm. This is a cost effective sensing system, particularly useful in developing countries where WSN is still a novice and costly technology (Figure 3). Ardunio transmits/tweets sensory data either through gateway or in twitter environment. Anyone can follow the FieldTwitter sensory data in Twitter social network (Twitter, 2012) in the name of “Hydbot01”. FT sensory data was stored in twitter database (Twillog, 2012) in the form of webpage, with XML syntax, which could be useful to maintain FT database. In FT web interface, the sensory data is available in raw format (analog to digital conversion—ADC).

PHP-based algorithms have been developed for converting raw data into usable format (units) and store the data in GeoSense database for all sensory devices.

3. GEOSENSE DEPLOYMENT

3.1 Test Bed and Experiment

Long term field experiments (with different dates of sowing and different levels of nitrogen) are being carried out since 1980 at Agricultural Research Institute (ARI), Acharya N.G. Ranga Agricultural University (ANGRAU), Rajedranagar, Hyderabad, Andhra Pradesh, India to study climate change effects on various regional important crops (rice, maize and groundnut) to observe its effect on vulnerability and planning adoption strategies. The test bed is situated at 17°19′ 00″ Latitude and 78°23′ 00″ Longitude and at an altitude of 543.3 meter above Mean Sea Level (MSL). The test bed falls under semi-arid tropics, and belongs to Southern Telangana agro-climatic region with an average annual rainfall of 531.5 mm and temperatures ranging from 15°C to 41.6°C. The physio-chemical analysis of experimental site indicate that the soil is of clay loam nature with high organic carbon content, medium in available nitrogen, phosphorus and low in available potassium.

Fig. 4: Experimental Crops with GeoSense Systems

GeoSense research work with augment weather based precision farming equipment, particularly on precision irrigation and crop water requirement with maize (monsoon) and groundnut crops, crop yield modeling on rice crop as well as pest/ disease management were carried out with standard experiment design (Figure 4).

3.2 GeoSense Private Cloud Services

The strategy is to utilize private cloud computing in the agent system placed in the test bed. Fit2-PC associated with Opera-Unite (Opera, 2012) tool act as a cloud server for virtualization/storage/process/web models for online applications/ utilization (Figure 5).

4. GEOSENSE APPLICATIONS

Currently, the dynamic GeoSense is developed to provide assistance in precision farming aspects (irrigation scheduling and crop yield modeling). The Hargreaves EvapoTranspiration (ET) estimation (Hargreaves and Allen, 2003) model (scripted with Java, PHP, HTML and CSS) was used for predicting crop water requirement of maize and groundnut


crops, and SImulation Model for a RIce-Weather relation (SIMRIW) (Hoire et al., 1987) was implemented for the rice crop yield prediction at various growth stages. Also, the dynamic sensory information was used for energy balance based weather profile and crop pest/disease studies.

4.1 Crop Water Requirement (CWR)

Many techniques are available to determine the crop water requirement, common being Penman-Monteith (Allen et al., 1998) and Hargreaves methods. Presently, in Geosense, CWR interface was developed with sample, yet robust Hargreaves method. The web based interface enables the user community to obtain reference crop ET (ET0) with minimum data set (temperature and solar radiation) dynamically from WSN system (Sudharsan et al., 2010, 2011, 2012). The obtained ET0 value, after multiplying with the standard Food and Agricultural Organization (FAO, 2011) crop co-efficient (kc) values for respective crop, provides CWR for maize or groundnut (Figure 6).

Fig. 5: GeoSense Cloud Computing Flow Chart Fig. 6: GeoSense Interface for Crop Water Requirement

of Maize Crop

4.2 Crop (Rice) Yield Simulation

A simplified process model (SIMRIW) was used for simulating growth and yield of irrigated rice crop (Horie et al., 1987). Initial studies were carried on BPT 5204 and MTU 1010 cultivars of rice to compare actual yield with simulated yield (with DSSAT and SIMRIW). The yield data, used for the comparison, was obtained from long term experiment (1994–1997) (Rao and Reddy, 1998) and weather parameters from the adjacent (100 meters away) weather station. The three yields (DSSAT, SIMRIW and actual) were compared using statistical techniques (e.g. correlation–coefficient, linear regression and T-test) to gain confidence levels to use SIMRIW dynamically. The confidence levels on SIMRIW helped developing Java based interface to (1) use dynamic FieldServer based sensory information (CO2 concentration, Temperature and Solar Radiation) (2) run model ubiquitously and (3) simulate daily/weekly/monthly/seasonal leaf area index, dry weight, grain yield and potential yield. The SIMRIW model was run with 2010 Kharif (from June to September) sensory information for different dates of planting. Subsequently, SIMRIW and CERES-Rice simulated final yields with sensory information of Kharif 2010 were correlated with actual/observed yields (Sudharsan et al., 2012).

4.3 Energy Balance and Weather Profile Studies

Flux Tower studies on energy balance and weather profile for Kharif 2010 season indicate that CO2 concentration during early hours of the day was higher. As the day progresses, photosynthesis process get activated and thus reduces CO2 concentration as compared to early hours as well as late hours of the day. In case of temperature profile, temperature at middle height (2 m) is higher than the other two heights while temperature of other two heights is nearly same. Relative humidity profile, unlike the temperature profile, shows decreasing trend during daytime and ascending sides in the night. Bowen ratio ranges from 0.35 to 0.49 in the season. The average latent heat flux for the season was 3.63 W/m2 while sensible heat flux was found to be 1.56 W/m2. The per cent energy utilized in latent heat flux (ET) and sensible heat flux is about 70 per cent and 30 per cent, respectively. The results have shown that this energy balance study can be used in precision agriculture to improve water management strategies (Zeggaf et al., 2007; Karandikar et al., 2011).

4.4 Crop Pest/Disease Prediction

Data driven precision agriculture aspects, particularly the pest/disease management, require a dynamic crop-weather data. Therefore, an attempt has been made to understand the crop-weather-pest/disease relations using wireless sensor


and field-level surveillance data, in conjunction with data mining techniques wherein, the closely related and interdependent pest (Thrips)—disease (Bud Necrosis) as well as Leaf Spot and Rust disease dynamics of peanut (groundnut) crop were studied (Tripathy et al., 2011). Data mining techniques were used to turn the data into useful information/knowledge/ relations/trends and correlation of crop-weather-pest/disease continuum (Figure 7). These dynamics obtained from the data mining techniques and trained through mathematical models were validated with corresponding surveillance data. Results obtained from 2009 and 2010 Kharif seasons and 2009–10 and 2010–11 Rabi seasons (post monsoon) data have been used to develop decision support system for pest/disease predictions in peanut crop. In an another attempt, image processing algorithms were used for automatic identification and count of pests such as aphids, whiteflies (Adult Jassids), predators (Coccinellide Beetle), etc. which are stuck to a yellow sticky trap placed in front of the FieldServer camera (Figure 8). The size and color components of object were selected as the features for automatic identification.

Fig. 7: DM Processing Flow for Pest/Disease Dynamics Fig. 8: Deployment of Yellow Sticky Trap and Different

Insects in Trap and Field

4.5 Real-time Decision Support System

The push-based GeoSense DSS (Figure 9) is designed to cater the rural/farming community for precision agriculture decision-making, i.e., how much to irrigate according to the plant and soil conditions and the likely yield the farming community can expect on a day/month/season basis. This system also provides local agricultural market information, which helps the farming community make necessary and ubiquitous strategic decisions. These systems also preserve the flow in information dissemination and provide real-time location-specific distributed sensory information, which can also be used with user-defined models/techniques. FS-based Flux Tower and FieldTwitter were used to establish efficiency of FS outreach on social environment platforms, respectively.

4.6 Towards Interoperable Service Oriented Architecture

To avoid heterogeneity in data collection and storage, an attempt has been made to implement the Open Geospatial Consortium (OGC) specified Sensor Web Enablement standards to the pre-existing distributed sensor systems (Field Server and Agrisens) through Sensor Model Language and Sensor Observation Service (Arthur and Priest, 2007; Botts et al., 2006; Botts et al., 2007). An interface has been developed to visualize the sensor observations and measurements. The Layered architecture of service consists of Distributed Application Clients, Sensor Observation Service (SOS) and WSN GeoSense, respectively (Figure 10). The database architecture for SOS is obtained from open source project of 52North SOS (Walkowski et al., 2012).

The data from two different WSN systems is collected together in SOS database by using SOS wrapper and it is accessed subsequently by the geographically distributed application clients through standard XML-HTTP requests. The SOS wrapper helps to convert raw data from different formats (text format in AS and XML format in FS) to real data in the SOS database. It processes the raw voltage data and converts into the real values at fixed intervals by using the calibration equations specified in the SensorML, which is stored in required relations of SOS database by executing Structured Query Language (SQL) insert statements. The SOS wrapper facilitates transactional data insertion, which help in the real time observation of the data. Standardization has provided an ability to operate in interoperable manner with any other OGC standardized and geographically separated sensor services. The resulting system can facilitate the rural extension community and environmental researchers for precision agriculture, environmental applications, etc. Open source platforms have been used for a cost effective and standardized interoperable service oriented WSN in agriculture.


Fig. 9: GeoSense Homepage Fig. 10: Service Oriented Architecture for GeoSense

5. CONCLUSIONS

Realizing the need for development of a real-time decision support system in precision farming aspects, an attempt has been made to utilize/integrate Geo-ICT and WSN (together with cloud computing functionality) to provide information and modeling services ubiquitously. Experiments were laid out in semi-arid tropic area on crop yield modeling, crop water requirement and pest/disease modeling by utilizing mote-based Agrisens and Wi-Fi based FieldServer distributed sensing devices. Thus, this system can empower the precision farming/extension community with information, modeling and decision making processes. This integrated real time GeoSense decision support system is a part of Indo-Japan bilateral initiative.

REFERENCES Allen, R.G., Pereira, L.S., Raes, D. and Smith, M., 1998, Crop evapotranspiration: guidelines for computing crop water requirements. FAO Irrigation and Drainage, 56. Ardunio, 2012, Open-source hardware and software. http://www.arduino.cc/, accessed on 2 Jan, 2012. Arthur, N. and Priest, M., 2007, Sensor Observation Service (SOS). Open Geospatial Consortium Implementation Standard,

http://www.opengeospatial.org/standards, accessed on 26 Oct., 2011. Botts, M., Robin, A., Davidson, J. and Simonis, I., 2006, OpenGIS Sensor Web Enablement (SWE) Architecture Document. Open

Geospatial Consortium, http://www.opengeospatial.org/standards, accessed on 26 Oct., 2011. Botts, M. and Robin, A., 2007, OpenGIS Sensor Model Language (SensorML) Implementation Specification. Open Geospatial

Consortium, http://www.opengeospatial.org/standards, accessed on 26 Oct., 2011. DSSAT, 2011, Crop Modeling Tool. Decision Support System for Agrotechnology Transfer, http://www.icasa.net/dssat/, accessed on 10

June, 2011. FAO, 2011, Food and Agriculture Organization. Crop coefficient values. http://texaset.tamu.edu/cropcoe.php, accessed on 06 June, 2011. FieldServer, 2011, NARC (National Agricultural Research Center), Japan. http://model.job.affrc.go.jp/FieldServer/ FieldServerEn/,

accessed on 06 June, 2011. GeoSense, 2011, Geo-ICT and Sensor Network based Decision Support Systems in Agriculture and Environment Assessment

India. http://www.csre.iitb.ac.in/geosense/, accessed on 06 June, 2011. Hargreaves, G..H. and Allen R.G., 2003, History and Evaluation of Hargreaves Evapor-transpiration Equation. Journal of Irrigation

and Drainage Engineering, 129, 53. Horie, T., Nakagawa, H., Ceneno, H.G.S. and Kropff, M.J., 1987, The Rice Crop Simulation Model SIMRIW and its Testing.

Modeling the Impact of Climate Change on Rice Production in Asia, 1995, pp. 51. Karandikar, K., Adinarayana, J., Tripathy, A.K., Sudharsan, D., et al., 2011, Energy Balance and Weather Profile Studies using

Sensor Network based Flux Tower—A preliminary Approach, Geomatrix 11, 14–25. Kesavan, P.C. and Swaminathan, M.S., 2006, From green revolution to evergreen revolution. Current Science, 91, 145–146. Opera, 2012, Opera-Unite. Available on www.opera.com/unite/, accessed on 11 June, 2011. SPANN Lab, 2011, Agrisens SPANN Lab IIT Bombay. http://www.ee.iitb.ac.in/spann/index.html, accessed on 3 Nov., 2011. Sudharsan, D., Adinarayana, J., Tripathy, A.K., et al., 2010, Dynamic simulation and comparison of rice growth and yield

prediction by wireless sensor network technology. 3rd International Rice Congress, pp. 10–15, Hanoi, Vietnam.


Sudharsan, D., Adinarayana, J., Tripathy, A.K., 2011, GeoSense: Geo-ICT and wireless sensor network based dynamic real-time decision support system for precision agriculture. International Symposium on Sustainable Bioproduction: Water, Energy and Food (CIGR ‘11), Tokyo, Japan, September 2011.

Sudharsan, D., Adinarayana, J., Reddy, D.R., et al., 2012, Evaluation of weather-based rice yield models in India. International Journal of Biometeorology, DOI 10.1007/s00484-012-0538-6.

Sudharsan, D., Adinarayana, J., Tripathy, A.K., et al., 2012, GeoSense: a Multi-mode Information and Communication System. International Scholarly Research Network-Sensor Network (ISRN-SN), Article ID 215103, 1–13.

Tripathy, A.K., Adinarayana, J., Sudharsan, D., et al., 2011, Data mining and wireless sensor network for agriculture pest/disease predictions, Information and Communication Technologies (WICT), 1229–1234.

Twillog, 2012, Twitter database. http://twitter.com/HydBot01, accessed on 2 Jan., 2012. Twitter, 2012, Instant social updating network. http://twitter.com/HydBot01, accessed on 2 Jan. 2012. Walkowski, A.C., Stasch, C., Künster, S., Kuhlmann, J., Hollmann, C. and Strotmann, A., 2012, Sensor Observation Service version

3.1.1, 52° North Initiative for Geospatial Open Source Software GmbH. https://wiki.52north.org/bin/view/Sensornet/Sensor ObservationService, accessed on 25 Jan., 2012.

Yogeswara Rao, A. and Reddy, R.D., 1998, Dynamic simulation of low land rice growth and yield in Andhra Pradesh. DST Project Completion Report. ARI, ANGRAU, RajendraNagar, Hyderabad.

Zeggaf, T.A., Anyoji, H., Takeuchi, S. and Yano, T., 2007, A Bowen Ratio Technique for Partitioning Energy Fluxes between Maize Transpiration and Soil Surface Evaporation. American Society of Agronomy, 100 (4), 988–996.

GEOSENSE: AN INFORMATION, COMMUNICATION AND DISSEMINATION ...insait.in/AIPA2012/articles/037.pdf · GeoSense: An Information, Communication and Dissemination System for Decision Support

Documents