AN EXPERT SYSTEM FOR DIAGNOSIS OF DISEASES IN TOMATOES

Abstract
This paper discusses the design, implementation and testing of the
“diagnosis of diseases in tomatoes” expert system. This was created
using the “ES-Builder Web” program, which is a rule based browser
software. This concluded that the created expert system was a success
and had a good accuracy which has helped spot and differentiate
diseased and healthy tomatoes from one another.

Background
In recent years there have been many different kinds of advancements in agriculture,
one of those advancements are with the use of expert systems. All expert systems
being used are based on artificial intelligence and are designed to find new solutions
to technical problems. Created by a lot of experts’ knowledge and with the use of
computer systems. The goal of this project is to use rule and faces taken from
research and experts in the agriculture field to help solve problems and reduce
diseased tomatoes. This project will use a rule based expert system which makes it
possible to carry out analysis functions and reasoning. It allows the system to have
an identical outcome as an expert would have. All of current day’s commercial prop
production have potential to use Expert Systems. Each of these commercial prop
production’s expert system could bring improvement and raise performance of
workers to an expert’s level.

This study source was downloaded by 100000809861525 from CourseHero.com on 04-07-2022 05:29:35 GMT -05:00

https://www.coursehero.com/file/51563539/ICT319-Reportdocx/
 Introduction
In recent years there have been many different kinds of advancements in agriculture,
one of those advancements are with the use of expert systems. All expert systems
being used are based on artificial intelligence and are designed to find new solutions
to technical problems. Created by a lot of experts’ knowledge and with the use of
computer systems. The goal of this project is to use rule and faces taken from
research and experts in the agriculture field to help solve problems and reduce
diseased tomatoes. This project will use a rule based expert system which makes it
possible to carry out analysis functions and reasoning. It allows the system to have
an identical outcome as an expert would have. All of current day’s commercial prop
production have potential to use Expert Systems. Each of these commercial prop
production’s expert system could bring improvement and raise performance of
workers to a expert’s level.

AI Method and Tools

The expert system chosen to be used for the diagnosis of diseases in tomatoes
project is called: “ES-Builder”. A browser based expert system software. It is a rule
based which means the system applies to a set of human entered rule sets which it
shall follow. The reason of choosing this software instead of others is because of the
built in functions this expert system has to offer to its users. Function allowing to
display conclusions, expert system, display decision tree, decision table, list attribute
values and testing the build expert system to determine certainty factors.
Before starting the building process of the expert system, a list was made to ensure
no requirements were left out of the expert system. This list consisted of diseases
which could be found in tomato and tomato plants, Order of rule based questions
and the conclusions to each step. All these procedures were created by experts and
researchers.
Once all the knowledge had been collected the system was created. Each step
consisted of two conditions which are string values “Yes” and “No”. Each of these
two conditions either came to a conclusion or continued to the next condition if
available. Rules set in the system are IF-THEN statements.

This study source was downloaded by 100000809861525 from CourseHero.com on 04-07-2022 05:29:35 GMT -05:00

https://www.coursehero.com/file/51563539/ICT319-Reportdocx/
 Evaluation Method
Diseases found in tomatoes lacks a lot of online research and studies. These two
areas which were lacking a lot of expert information which created difficulties adding
more tomato diseased to the expert system.
To make the expert system easy to understand and read by the user more photos
were added to indicate what is meant by the expert system’s question. Each decision
step had photos linked to it which helps the used and makes them more comfortable
with the use of the software. A sample of 15 tests were done in the beginning which
ended up with negative results. Only 7 out of the 15 tests had the correct conclusion.
This became an issue due to having difficulties in finding the diseased tomatoes. To
solve this issue more research was done to remove certain knowledge from the
expert system and replace that knowledge with more important and relevant
knowledge to increase the accuracy of the results.

Project report schedule:

Tasks:
1. Research
Do more research to gain more knowledge and get more references to researchers
and information.
22 - 23 October
a. Introduction
i. Rewrite and adjust the current introduction to fit the report.
1. 24 - 26 October
b. Background
i. Write a background with the research done (1)
1. 27 - 28 October
b. Method / Tools
i. Use the method / tools
1. 29 October - 1 November
b. Discussion
1. 2 - 5 October
2. References

This study source was downloaded by 100000809861525 from CourseHero.com on 04-07-2022 05:29:35 GMT -05:00

https://www.coursehero.com/file/51563539/ICT319-Reportdocx/
 i. Noted at every step of the report to ensure all information is linked to the right
references.
1. Start -> End
b. User Guide
i. Create a readable user guide for the marker / reader.
1. 8 - 9 November
b. Finalize document
i. Finalize document by making any necessary adjustments in the report to fit
the requirements of the given project.
1. 10 - 13 November
b. Submit Report
1. 14 November

Results
The “ES-Builder Web” software has a very self explanatory design which makes it
easy to use and understand. When starting the expert system the user will be
brought to this figure 0.002 page where the user can start by clicking “Start tomato
check”.

The system was tested using diseased tomatoes and checking if they came through
with a state declared healthy or diseased and all the tested tomatoes which included:
non fresh tomatoes, healthy tomatoes, rotten tomatoes and early blight affect
tomatoes. All these tested tomatoes are in the table shown below.

The expert system brought back positive results after the conducted tests. A 100%
(8/8) of the diseased tomato came out as diseased and the healthy tomatoes came
in the healthy category. But due to leaf mold not being inside one of the conclusion,
the results had a 87.5% accuracy(7/8) which did reach a pass mark.

This study source was downloaded by 100000809861525 from CourseHero.com on 04-07-2022 05:29:35 GMT -05:00

https://www.coursehero.com/file/51563539/ICT319-Reportdocx/
 Discussion
The ES-Builder web based tomato disease checking expert system was designed to
assist farmers site with agriculture to decrease faulty tomatoes from being marked
healthy. The system performance was highly rated and had a accuracy of 87.5%.
Performance could increase by adding more conditions and types of tomatoes by
expert and this could also be adjusted to suit more fruits on farms and assist others.
The goal was reached by successfully building an expert system which could identify
and differentiate diseased tomatoes from healthy tomatoes. Performance and
accuracy could be improved by expanding the knowledge of the expert system by
implementing more rules, more conclusions and more conditions to the expert
system.

References
1. Expert system for decision support in Agriculture, N. Sriram 2014
http://agritech.tnau.ac.in/pdf/14.pdf
2. Agriculture Expert system, shodhganga, 2013
http://shodhganga.inflibnet.ac.in/bitstream/10603/71534/10/10_chapter%203.pdf
3. Common tomato plant problems, Saferbrand, 2018
http://www.saferbrand.com/articles/common-tomato-plant-problems-how-to-fix-them

This study source was downloaded by 100000809861525 from CourseHero.com on 04-07-2022 05:29:35 GMT -05:00

https://www.coursehero.com/file/51563539/ICT319-Reportdocx/
Powered by TCPDF (www.tcpdf.org)