Spatial Analysis For Generating
Recommendations For
Agricultural Crop Production
Yash Jain, Amita Sharma, Sanjay Chaudhary,
Vikas Kumar Tyagi
Dhirubhai Ambani Institute of Information And
Communication Technology
Presented At
ICGTA 2012
IIT Bombay
Outline
Introduction
Motivation
Problem Statement
Methodology
Spatial Data
Analysis
Generating Recommendations
Ontologies and SWRL
Current Work
Future Work Plan
Introduction
Wide uncertainty in weather pattern all over the country.
Crop production has been affected by the changing
climatic conditions.
There is a need to monitor and foresee the agricultural
production based on the changing climate.
Moreover there is a need to generate recommendations
based on the climate and soil conditions.
We want to generate recommendations for cotton crop in
the North Gujarat region.
Motivation
Cotton – an important non-food crop provides lint to
textile industry, as high protein feed to livestock, oil for
human consumption, byproducts used as fertilizer,
produce paper, cardboard, etc. [1]
India is the 2nd largest cotton producer and consumer
[2]
Gujarat is second in cotton production after
Maharashtra.
North Gujarat contributes maximum in cotton
production in Gujarat.
Contd..
Different climate variable affect cotton production in
different ways and by different amount
We need to analyze the weather parameters with respect
to cotton yield to get impact of these parameters on
cotton yield or cotton production
Other variables affecting cotton are soil moisture,
relative humidity, bright sunshine hours etc.
Recommendations/ Alerts can be generated by analyzing the weather parameters and shifts in climate.
Rainfall Variations in Banaskantha, 1991-
2008
Yield Variations in Banaskantha, 1991-2008
Methodology
Obtain spatial data related to climatic parameters as well
as past data related to agriculture and integrate the data
obtained from different sources.
Develop algorithm for generation of alerts/
recommendations based on spatial data processing and
knowledge base.
Define interfaces to distribute the information to the end
users.
Proposed Architecture of the System
[2][
5]
Experiments and Results
S
No
Variable Standard
Error
Coefficient t-
Statistics
p-
value
R-
Square
F-test
1 Intercept 173.25 811.73 4.69 0.001
87.5
0.0003
2 RF_June 1.43 -3.14 -2.20 0.053
3 RF_July 0.31 -1.72 -5.40 0.000
4 RF_Aug 0.31 0.83 2.67 0.024
5 RF_Sep 0.60 1.36 2.28 0.046
6 RF_Oct 1.27 -4.03 -3.18 0.010
Creation of Spatial Database
GIS is playing an increasing role in agriculture
production throughout the world by helping farmers
increase production, reduce costs, and manage their land
more efficiently[2].
Combination of ES and GIS can make the geography
information to bring into the ES decision process[3].
An attribute database is created for the districts of state
which stores the data regarding the meteorological
parameters and soil condition.
The climatic parameters such as rainfall, surface
temperature, humidity and soil qualities such as soil
nutrient contents will be stored in the attribute database..
Analysis of the Spatial Data
The attribute database shall be queried to find out areas
with some threshold parameters.
We shall execute spatial queries to identify specific
scenarios. For example:
◦ Identify areas where rainfall is less:
◦ Identify areas where irrigation facilities are available.
Spatial overlay analysis will be done to identify areas
with typical combinations of temperature, rainfall and
soil quality.
Analysis Of Spatial Data(Contd..)
The growth patterns of cotton in North Gujarat can be
analyzed by performing an overlay analysis of the crop
yield along with meteorological parameters and soil
data.
The overlay analysis reveals relationship of crop yield
with meteorological parameters. Moreover it also helps
to classify data based on climatic parameters and soil
conditions.
The classified data will be passed to the reasoning
engine for generation of recommendations.
Ontologies In the context of computer and information sciences, an
ontology defines a set of representational primitives with
which to model a domain of knowledge or discourse.
The representational primitives are typically classes (or
sets), attributes (or properties), and relationships (or
relations among class members).
We have developed a cotton ontology by extending
AGROVOC.
Ontologies provide a conceptual knowledge base for the
application.
An ontology for rice has been developed for helping the
researchers retrieve documents with greater precision.[4]
Semantic Annotations can be added using domain
ontologies[5].
Source: http://keet.wordpress.com/2009/11/20/72010-semwebtech-lectures-34-ontology-engineering-top-down-
and-bottom-up/
Semantic Web Rule Language
SWRL allows reasoning about OWL individuals. As
defined by W3C, SWRL extends the set of OWL axioms
with Horn-like rules.
It combines the sublanguages of OWL-DL, OWL Lite
and RuleML. Rules are of the form of an implication
between an antecedent (body) and consequent (head).
For example:
Cotton(?x1) ^ hasSoilTemperature(?x1,?x2) ^
swrlb:greaterThan(?x2, 25) ^ hasHumidity(?x1,?x3)
^ swrlb:greaterThan(?x3, 80) ->
hasPossibilityOfOccurence(?x1, RootRot)
Generation Of Alerts/Recommendations
The processing of spatial data based on spatio temporal
modeling will provide useful information regarding the
possible crop outcome in a particular region.
Based on the outcomes, alerts/recommendations can be
generated for a specific region depending on its climatic
conditions.
The recommendations will guide the farmers in their
choice for fertilizers, crop and pesticides.
Interfacing Rule Engine With GIS
We have to interface the reasoning engine with the GIS
server.
We are currently exploring the ways of accessing
ArcObjects with Java API.
An interconnection component will be developed which
will serve to transfer the data obtained from GIS
analysis to the rule engine in an XML format.
Finally a service oriented architecture will be developed
that will allow access to our services in a platform
independent manner.
Information Distribution To End Users
We need to define interfaces through which end users
can request for services and receive the required
information.
One possible way is to have a web based access to the
information. The web services will be exposed and can
be invoked by a web application which takes as input
the location of the user and provides alerts and
recommendations
Information Distribution through SMS.
Current Work
Gathering Spatial Data from multiple sources.
Working on a GIS software to add data and perform
analysis.
Building an attribute database for the districts of Gujarat
which comprise of meteorological data as well as soil
parameters.
Adding concepts and rules to the Ontology.
Identifying ways to interface GIS and rule engine.
Future Work
Integrating spatial data available for different climate
parameters.
We shall use LISS3 as well as AWiFS data for
identification of areas growing cotton.
Developing algorithms for processing of spatial data and
generating alerts.
Defining interfaces to distribute information.
Improving Cotton Ontology for by adding new concepts
and rules targeted towards specific cotton varieties.
Developing a Service Oriented Architecture to expose
the functionality as Web Services.
Acknowledgement
This work is a part of on-going R&D project “Title of
the Project: "Service-Oriented Architecture for Spatial
Data Integration and Spatial Reasoning" funded by
NRDMS, Department of Science and Technology,
Government of India.
We would like to acknowledge Krishi bhavan,
Gandhinagar and Krishi Vigyan Kendra for their support
in terms of meteorological data as well as data related to
cotton crop.
References
[1]Freeland Jr., Thomas B. , Pettigrew, Bill, Thaxton, Peggy, Andrew,
Gordon L., “Agrometeorology and Cotton Production” Guide of
Agricultural Meteorological Practices (GAMP), 2010, edition
(WMO-No.134), Chapter 13.
[2] Osakwe, Emeka, “Cotton Fact Sheet India”, International Cotton
Advisory Committee, May 19, 2009.
[3] Zhiqing Zhu, Rongmei Zhang, and Jieli Sun. “Research on gis-
based agriculture expert system.” In Proceedings of the 2009 WRI
World Congress on Software Engineering – Volume 03, WCSE ’09,
pages 252–255, Washington, DC, USA, 2009. IEEE Computer
Society
[4] Aree Thunkijjanukij,” Ontology Development For Agricultural
Research Knowledge Management: A Case Study For Thai Rice”,
Phd Dissertation, Kasetsart University, 2009.
Contd..
[5] N. Agarwal, M. Rao, S.S. Mantha, and J.A. Gokhale. Annotation of
geospatial data based on semantics for agriculture: Case study for
india. In Computer Research and Development (ICCRD), 2011 3rd
International Conference on, volume 1, pages 139 –142, march
2011.
[6] M. Shoaib and A. Basharat. Semantic web based integrated
agriculture information framework. In Computer Research and
Development, 2010 Second International Conference on, pages
285 –289, may 2010.