Original papersAn intelligent mobile application for diagnosis of crop diseases in Pakistan using fuzzy inference system
Introduction
According to Food and Agriculture Organisation of The United Nations (Food, 2013), Pakistan is a rich country in agriculture sector since its inception and it has a vast arable land with the largest and oldest irrigation system in the entire region. Pakistan is the 4th largest cotton producing country and 7th largest wheat producing country besides many other crops like sugarcane (Food, 2013). But unfortunately, agriculture contribution in country’s gross domestic product (GDP) has declined from 54% to 24.6% over the last decade, and total crop production in Pakistan is almost 50% below its potential (Sindhu et al., 2010, Malik, 2012, G. of Pakistan, 2014). Despite of being one of the largest producers of crops, Pakistan is facing many difficulties and problems in the agriculture sector since last decade, which is mainly because of issues like lack of relevant research and lack of state of the art facilities. In agriculture extensions departments, there is lack of trained experts and modern equipment which do not help the farmers adequately to assist in solving their problems including efficient disease diagnosis. Disease diagnosis is inefficient and less accurate due to unavailability of experts in the field specifically the rural areas which encompass about 70% area of the country.
In this research work, an intelligent expert system for diagnosing the crop diseases using fuzzy logic based decision making algorithm has been proposed. This study focuses on two main crops in Pakistan: wheat and cotton. The system is capable to diagnose twenty-one common diseases of wheat and cotton. Proposed architecture has been developed using fuzzy logic as decision making engine at the backend and Android application has been designed for the front end using jFuzzylite library. jFuzzylite library (Rada-Vilela, 2013) is an open source library developed for building fuzzy inferences systems in Java and Android. In rural Pakistan, farmers are not much literate, so there is a provision of using the local language Urdu to interact with the mobile application along with default provision of English language. There are two main reasons of applying fuzzy logic for decision making in the proposed framework. First reason is that rules are derived from expert knowledge which is described in natural language and fuzzy logic is a powerful knowledge representation mechanism for linguistic knowledge. Second reason is that it handles the vagueness and uncertainty inherent in the problem domain which is not handled by classical set theory (Zadeh, 1994, Mamdani, 1974). It gives the farmer and other agriculture related people provision to provide vague inputs as intuitive guesses when they are not much clear. Different aggregation and defuzzification methods have been implemented for the evaluation purpose. Based on the empirical results, the best combination of these methods is recommended to be used at practical scale. It has been found that the optimal combination is to set max method for aggregation and middle of maximum (mom) method for defuzzification.
Rest of the paper is organised as follows: Section 2 discusses the problem domain and related work in literature, Section 3 describes the fundamental concepts of fuzzy logic and working mechanism of fuzzy inference system, Section 4 explains the proposed approach, Section 5 presents the results of the proposed system and their analysis and Section 6 concludes the paper highlighting some of the prominent future directions of this work.
Section snippets
Background and related work
Incorrect and late disease diagnosis is one of major problems for low yield of crops in the developing countries including Pakistan (Anwar et al., 1993, Mukhtar, 2009). Some of major cotton diseases are leaf curl virus, root rot and boll rot. Some of the major wheat diseases include black stem rust, yellow or stripe rust, loose smut and flag smut (ur Rehman Rattu et al., 2011).
In recent years, technological research in this domain has shown significant economical benefits (Evenson et al., 1979,
Fuzzy logic and fuzzy inference system
Fuzzy logic is a powerful logic based on fuzzy theory to incorporate the uncertainty and ambiguity of real world (Zadeh, 1965). In real world, there are many scenarios where boolean logic does not represent the real phenomenon in true spirit. Fuzzy logic takes care of partial truth and is capable enough to capture the degree of truthfulness of happening of an event in real world.
Fuzzy Inference System (FIS) is a rule-based system which exploits fuzzy logic in contrary to boolean logic. In many
Proposed architecture to diagnose crops diseases using fuzzy inference system
The literature review revealed that the problem domain of crop diseases significantly depends on the subjective and linguistic knowledge representing the key characteristics of the crops specifically their symptoms. This fact becomes more obvious if the context of farmer is considered as a priority factor. Farmer is one of the key stakeholders in the management of crops. At the same time, farmer has limited data available with him as far as the precision and clarity of the input is concerned.
Data collection
Raw data collection of symptoms and corresponding diseases is a crucial step in driving the flow of the entire system. Primarily, data has been collected for wheat and cotton from different sources including websites of agriculture departments of governments, websites of agriculture universities, online literature, farmers in local areas (field surveys) and agriculture experts in local areas. The raw data gathered from different sources has been stored in a plain database. It contains 160 rows
Conclusions and future work
This study focused on the development of the intelligent framework for the diagnosis of wheat and cotton crop diseases in Pakistan. A mobile application has been built for the disease diagnosis with English and Urdu language interfaces for the rural farmers. The collection and authentication of disease symptoms data has been done through different online resources, visits to crop fields, extension departments of agriculture universities and district agricultural offices in Pakistan. Fuzzy
References (53)
- et al.
Weather forecasting model using artificial neural network
Procedia Technol.
(2012) - et al.
Intelligent systems for the assessment of crop disorders
Comput. Electron. Agric.
(2012) - et al.
Evaluation of student performance in laboratory applications using fuzzy logic
Procedia Soc. Behav. Sci.
(2010) - et al.
A web-based intelligent disease-diagnosis system using a new fuzzy-logic based approach for drawing the inferences in crops
Comput. Electron. Agric.
(2011) - et al.
Stock market index prediction using artificial neural network
J. Econ. Financ. Admin. Sci.
(2016) - et al.
Using expert system as a training tool in the agriculture sector in Egypt
Expert Syst. Appl.
(1996) - et al.
Development and implementation of a acquisition methodology for crop management expert systems
Comput. Electron. Agric.
(1993) - et al.
Automation of agriculture support systems using Wisekar: case study of a crop-disease advisory service
Comput. Electron. Agric.
(2016) Fuzzy sets
Inf. Control
(1965)- A. Department, 2014. Agriculture Department, Government of Punjab, Pakistan. Technical Report. Ministry of Agriculture,...
Nematode diseases of rice in the Punjab, Pakistan
Pakistan J. Agric. Res.
Medical diagnosis with the aid of using fuzzy logic and intuitionistic fuzzy logic
Appl. Intell.
Expert system for diagnosis pest and disease in fruit plants
EPJ Web of Conf.
Mining the classification rules for egyptian rice diseases
Int. Arab J. Informat. Technol.
Economic benefits from research: an example from agriculture
Sci. Mag.
From data mining to knowledge discovery in databases
AI Mag.
The role of agriculture in development
Am. Econ. Rev.
Sustainable use of groundwater for irrigated agriculture: a case study of Punjab, Pakistan
Eur. Water
Assessing mobile technology usage for knowledge dissemination among farmers in Punjab
Informat. Technol. Develop.
Improved fuzzy rule promotion-based localized crop knowledge dissemination system for farmers of Punjab (India) using mobile phone technology
Cybernet. Syst.
Agro Genius: an emergent expert system for querying agricultural clarification using data mining technique
Int. J. Eng. Sci.
Cited by (46)
Ensemble of CNN models for classification of groundnut plant leaf disease detection
2023, Smart Agricultural TechnologyArtificial intelligence applications in the agrifood sectors
2023, Journal of Agriculture and Food ResearchDirected acyclic graphs-based diagnosis approach using small data sets for sustainability
2023, Computers and Industrial EngineeringCitation Excerpt :The disease diagnosis algorithm proposed in this paper can increase the ratio of solution existent from 4.00 % to 62.78 % through fuzzy inference without expanding the data set. Compared with the existing disease diagnosis algorithm based on fuzzy set theory (Toseef & Khan, 2018; Zadeh, 1965), the advantage of this algorithm is more lightweight, that is, it does not need to create fuzzy membership functions and other additional parameters for each indicator. Therefore, as long as the original disease data set is entered, the disease diagnosis system can help users diagnose possible diseases.
An app to assist farmers in the identification of diseases and pests of coffee leaves using deep learning
2022, Information Processing in AgricultureA rapid, low-cost deep learning system to classify strawberry disease based on cloud service
2022, Journal of Integrative AgricultureCitation Excerpt :To identify the main diseases of cotton and wheat, a crop disease identification method running on Android mobile devices using a fuzzy inference system as the primary back-end decision engine was developed. It takes the symptoms of the crop as input in the form of fuzzy input, and uses an inference engine to generate output in the form of disease identification (Toseef and Khan 2018). To deal with the detection of a variety of plant diseases under actual acquisition conditions, an adaptive algorithm based on deep residual neural network was proposed and applied.