2019 5th International Conference on Advanced Computing & Communication Systems (ICACCS)

A Review on Machine Learning Classification

Techniques for Plant Disease Detection
Mrs. Shruthi U Dr. Nagaveni V Dr. Raghavendra B K
CSE Department CSE Department CSE Department
AcIT AcIT KSSEM
Bengaluru, India Bengaluru, India. Bengaluru, India.
[email protected] [email protected] [email protected]

Abstract— In India, Agriculture plays an essential role learning classification techniques used to recognize and
because of the rapid growth of population and increased in classify plant diseases.
demand for food. Therefore, it needs to increase in crop yield.
One major effect on low crop yield is disease caused by
bacteria, virus and fungus. It can be prevented by using plant
II. GENERAL PLANT DISEASE DETECTION SYSTEM
diseases detection techniques. Machine learning methods can be The plant diseases can be identified by observing leaf,
used for diseases identification because it mainly apply on data stem and root part of the plant. The digital image processing
themselves and gives priority to outcomes of certain task. This can be used to detect diseased leaf, stem, fruit and flower,
paper presents the stages of general plant diseases detection shape and colour of affected area. Image processing
system and comparative study on machine learning technique involves five basic steps and the data flow diagram
classification techniques for plant disease detection. In this is as shown below in Fig 1.
survey it observed that Convolutional Neural Network gives
high accuracy and detects more number of diseases of multiple
crops.

Keywords— Plant disease detection; Classification; Image

Acquisition Annoted
Machine Learning.
dataset

I. INTRODUCTION
In India, agriculture has become important source of the
economic development. Farmer selects the suitable crop
based on type of soil, weather condition of the location and
economic value. The agriculture industries started searching Image processing
new methods to increase production of food because of
increasing population, changes in weather and instability in
politics. This makes researchers to search new efficient and
precise technologies for high productivity. Feature extraction
Farmers can collect the information and data by use of
precision agriculture in in-formation technology to take best
decision on high output from the farm. Precision agriculture
is new technology, which provides advanced techniques to Classification
improve farm output. By utilizing these advanced
technologies, it is possible to achieve economic growth in
agriculture. Precision agriculture can be used for many Fig. 1. Plant disease detection system.
applications like pest detection in plants, weed detection,
yield production of crops and plant disease detection etc. A
farmer uses pesticides to control pest, prevent diseases and to • Image Acquisition: First stage of plant disease
in-crease crop yield. The diseases in crop are creating detection system is image acquisition. High quality plant
problem of low production and economic losses to farmers images can be acquired using digital cameras, scanners or
and agricultural industries. Therefore identification of disease drones.
and its severity based as become necessary.
• Annoted Dataset: Knowledge based dataset to be
Disease identification in plant is most important in created for captured images with different classes.
successful farming system. In general, a farmer recognizes
the symptoms of disease in plants by using naked eye • Image processing: Acquired images to be involved
observations and this requires continuous monitoring. in pre-processing stages to improve some image features
However, this process is more expensive in large plantations important for further processing. Segmentation pro-cess is
and sometimes this may be less accurate. In some countries used to partition the plant image in various segments. This
like India, farmers may have to show the specimen to experts, can be used for the extraction of diseased area in leaf, stem or
this makes time consuming and more expensive. The root of plant from background.
following sections in this paper contains general steps of • Feature extraction: Extraction of Colour, shape
plant diseases detection system and survey on machine and texture feature of disease part of the plant can be done

978-1-5386-9533-3/19/$31.00 ©2019 IEEE 281

Authorized licensed use limited to: University of Kashmir Srinagar. Downloaded on April 29,2024 at 10:51:38 UTC from IEEE Xplore. Restrictions apply.
 2019 5th International Conference on Advanced Computing & Communication Systems (ICACCS)

using grey level Co-occurrence Matrix(GLCM), Blend vision plant disease and observed using 40 images with accuracy of
and machine intelligence etc. 90%.
• Classification: Finally, any of the machine learning
techniques can be used to classify the various diseases in C. KNN Classifier
plants. K-Nearest Neighbors has been used for pattern
recognition, statistical estimation and classification in
Machine Learning. We made survey on plant disease
III. RELATED WORK
detection using KNN classifier as follows. [13] Proposed an
Extensive survey has been conducted to compare disease algorithm for identification of disease in sugarcane culture.
detection and classification techniques in Machine learning. Image processing algorithms are used for feature extraction.
We studied Support Vector Machine (SVM) Classification It secured an accuracy of 95% for Leaf scorch disease
Technique, Artificial Neural Network (ANN) Classification detection in sugarcane leaf. [14] Developed a method to
Technique, K- Nearest Neighbor Classification Technique, estimate severity and detection of cotton plant disease Grey
Fuzzy C-Means Classifier and Convolutional Neural Mildew disease achieved with accuracy of 82.5% using 40
Network Classification methods used in detection of plant images. [15] Proposed an algorithm for plant disease
diseases and its efficiency. detection using GLCM feature extraction method and KNN
classifier. The KNN classifier is proposed rather than SVM
A. SVM Classifier classifier to classify data in multiple classes. The
SVM Classifier is supervised learning method in Machine performance is tested in the terms of accuracy is in-creased
learning here analysed data is used for classification. The compared with SVM classifier.
following authors used SVM Classifier in disease detection of
different crops. [1] Detection of diseases on citrus trees which D. FUZZY Classifier
include grapefruit, lemons, lime and oranges leaf attack by Related work on Fuzzy Classifier in plant disease
canker and anthracnose diseases. The experimental result detection is an author [16] presented a method to identify the
obtained 95% of genuine acceptance rate. [2] Grape plant presence of infection in wheat crop images using Fuzzy
diseases Downy Mildew and Powdery Mildew detected and Classifier. This algorithm is tested with the dataset of
give 88.89% average accuracy for both the diseases. [3] oil healthy and unhelthy leaves. The classification of healthy
palm leaf diseases Chimaera and Anthracnose detection and unhelthy leaves found with accuracy of 88% and
achieves accuracy of 97% and 95% respectively. [4] Potato recognition of disease accuracy is 56%.
plant diseases are Late blight and Early blight detection over
300 publically available images with accuracy 95%. [5] E. Deep Learning
Grape leaf diseases Black Rot, Esca and Leaf Blight are
classified with accuracy using features from both LAB and Deep learning is a defferent learning method in ANN and
HSI colour model. [6] Developed a method to identify also a part of machine learning methods. [17] Proposed a
diseases in Tea plants. Three different types of diseases with model to detect healthy leaves and 13 different diseased
less in features are detected using SVM classifiers. The leaves of peach, cherry, pear, Apple and Grapevine using
developed method classified the diseases with accuracy of CNN classification technique. More than 30000 images used
90%. [7] Used for soybean culture to detect three different in dataset, achieved accuracy between 91% and 98% for
diseases Downy Mildew, Frog eye, and Septories leaf blgh. separate class test and average accuracy 96.3%. [18]
They reported with average classification accuracy Developed a method for plant disease detection using public
approximately 90% using big dataset. dataset 54306 images of 14 crops and 26 diseases and
performed with accuracy 99.35% using 20% of testing data
and 98.2% using 80% of testing data. [19] Designed a model
B. ANN Classifier to identify Soybean plant diseases Septoria, Frogeye and
Artificial Neural Network is computational model in Downy Mildew using CNN classifier. A dataset contains
machine learning and pattern recognition. Related work on 12673 leaf images with four classes and achieved 99.32%
Plant disease detection using ANN classifier as follows. [8] accuracy. [20] Developed CNN classification technique for
Evaluated a proposed work for recognition of plant diseases recognition of diseases in crops. The dataset containers
using feed forward back propagation algorithm and it 87848 images of 25 different plants in set of 58 disease and
performed well with a precision of around 93%. They tested achieved with accuracy 99.53%.
solution on early scorch, cottony mold, late scorch, and tiny
whiteness diseases which effect on plants. [9] Developed a The comparison of different type of Machine Learning
model to increase the accuracy in identification of two types classifiers used in plant disease detection is summarized and
of diseases caused by fungus are Downy Mildew and is given in Table 1.
Powdery Mildew in cucumber plant. [10] Introduced a Table 1. Comparison of classification techniques.
system to recognize and classify diseases like leaf spot,
bacterial blight, fruit spot and fruit rot diseases of Classification Culture No. of Result
pomegranate plant using back-propagation algorithm and the Technique Diseases
experimental result shows around 90% accuracy. [11] SVM Citrus [1] 2 95% of genuine
Proposed a work on identification of groundnut plant disease Classifier diseases acceptance rate.
is cercospora (leaf spot) using neural network Back
propagation method. The experimental results and Grape [2] 2 Average
observation shows out of 100 sample diseased leaf images diseases accuracy
they classified four types of diseases and secured 97.41% of 88.89%.
accuracy. [12] Proposed a method to detect pomegranate

282

Authorized licensed use limited to: University of Kashmir Srinagar. Downloaded on April 29,2024 at 10:51:38 UTC from IEEE Xplore. Restrictions apply.
 2019 5th International Conference on Advanced Computing & Communication Systems (ICACCS)

Oil palm 2 97% accuracy IV. CONCLUSION

[3] diseases for Chimaera A comparative study is carried out on five types of
and 95% machine learning classification techniques for recognition of
accuracy for plant disease is done in this review. SVM classifier is used
Anthracnose by many authors for classification of diseases when
disease. compared with other classifiers. The result shows that CNN
classifier detects more number of diseases with high
Potato [4] 2 Accuracy 90%. accuracy. In future, other classification techniques in machine
diseases learning like decision trees, Naïve Bayes classifier may be
used for disease detection in plants and in the sense of
Tea [5] 3 Accuracy 93%. helping farmer an automatic detection of all types of diseases
diseases in crop to be detected.
Reference
Soybean 3 Accuracy is [1] Kiran R. Gavhale, Ujwalla Gawande and Kamal O. Hajari,
[6] diseases approximately “Unhealthy region of citrus leaf detection using image processing
techniques”, IEEE International Conference on Convergence of
90% Technology (I2CT), Pune 2014, pp. 1-6.

ANN Not 5 Accuracy [2] Pranjali B. Padol; Anjali A. Yadav, “SVM Classifier Based Grape
Leaf Disease Detection”, IEEE Conference on Advances in Signal
Classifier Mentioned diseases around 93%. Processing (CASP), Pune 2016, pp. 175-179.
[8]
[3] Ahmad Nor Ikhwan Masazhar and Mahanijah Md Kamal, “Digital
Image Processing Technique for Palm Oil Leaf Disease Detection
Cucumber 2 Increased using Multiclass SVM”, IEEE 4th International Conference on Smart
[9] diseases accuracy Instrumentation, Measurement and Applications (ICSIMA), Malaysia
2017, pp. 1-6.

Pomegrana 4 Accuracy [4] Monzurul Islam, Anh Dinh and Khan Wahid, “Detection of potato
te[10] diseases around 90% Diseases Using Image Segmentation and Multiclass Support Vector
Machine”, IEEE 30th Canadian Conference on Electrical and
Computer Engineering (CCECE), Canada 2017, pp. 1-4.
Groundnut 4 Accuracy
[11] diseases 97.41%. [5] Nithesh Agarwal, Jyothi Singhai and Dheeraj K. Agarwal, “Grape
Leaf Disease Detection and Classification Using Multi- Class Support
Vector Machine”, proceeding of IEEE International conference on
KNN Sugarcane 1 Accuracy Recent Innovations in Signal Processing and Embedded Systems
(RISE), Bhopal 2017, pp. 238-244.
Classifier [13] disease 95%[13]
[6] Md. Selim Hossain, Rokeya Mumtahana Mou, Mohammed Mahedi
Cotton 1 Accuracy Hasan, Sajib Chakraborty and M. Abdur Razzak, “Recognition and
Detection of Tea Leaf’s Diseases Using Support Vector Machine”,
[14] disease 82.5%[14] IEEE 14th International Colloquium on Signal Processing & its
Applications (CSPA), Malaysia 2018, pp. 150-154.
Fuzzy Wheat [16] 1 Disease [7] Sukhvir Kaur , Shreelekha Pandey and Shivani Goel, “Semi-
Classifier disease detection automatic leaf disease detection and classification system for soybean
accuracy 88% culture”, journal on IET Image processing, Vol. 12, Issue 6, 2018, pp.
1038- 1048.
and recognition
of disease type [8] Dheeb Al Bashish, Malik Braik and Sulieman Bani- Ahmad, “A
accuracy56%. Framework for Detection and Classification of Plant Leaf and Stem
Diseases”, IEEE International Conference on Signal and Image
Processing (ICSIP), Chennai 2010, pp. 113-118.
CNN Peach, 13 Average
Classifier Cherry, diseases accuracy 96.3%. [9] Keyvan Asefpour Vakilian and Jafar Massah, “An artificial neural
network approach to identify fungal diseases of cucumber (Cucumis
Pear, sativus L.) Plants using digital image processing”, Archives of
Apple and Phytopathologyand Plant Protection, Vol. 46, Issue 13, Taylor &
Grapevine Francis 2013, pp. 1580-1588.
[17] [10] Mrunmayee Dhakate and Ingole A. B. , “Diagnosis of Pomegranate
Plant Diseases using Neural Network”, IEEE 5th National
14 crops 26 Conference on Computer Vision, Pattern Recognition, Image
Accuracy Processing and Graphics (NCVPRIPG), Patna 2015.
[18] diseases 99.35%.
[11] Ramakrishnan M. and Sahaya Anselin Nisha A., “Groundnut Leaf
Disease Detection and Classification by using Back Probagation
Soybean 3 Accuracy Algorithm”. IEEE International Conference on Communications and
[19] diseases 99.32%. Signal Processing (ICCSP), Melmaruvathur 2015, pp. 0964 – 0968.

[12] Rashmi Pawar and Ambaji Jadhav, “Pomegranate Disease Detection

25 plants 58 Accuracy and classification”, IEEE International Conference on Power,
[20] diseases 99.53%[20] Control, Signals and Instrumentation Engineering (ICPCSI), Chennai
2017, pp. 2475-2479.

283

Authorized licensed use limited to: University of Kashmir Srinagar. Downloaded on April 29,2024 at 10:51:38 UTC from IEEE Xplore. Restrictions apply.
 2019 5th International Conference on Advanced Computing & Communication Systems (ICACCS)

[13] Umapathy Eaganathan, Jothi Sophia, Vinukumar Lackose, Feroze

Jacob Benjamin, “Identification of Sugarcane Leaf Scorch Disease
using K-means Clustering Segmentation and KNN based
Classification”, International Journal of Advances in Computer
Science and Technology (IJACST), Vol. 3, No. 12, Special Issue of
ICCEeT, Dubai , 2014, pp. 11- 16.

[14] Aditya Parikh, Mehul S. Raval, Chandrasinh Parmar and Sanjay

Chaudhry, “Disease Detection and Severity Estimation in Cotton
Plant from Unconstrained Images”, IEEE International Conference
on Data Science and Advanced Analytic, Canada 2016, pp. 594- 601.

[15] Gautham Kaushal, Rajini Bala, “GLCM and KNN based Algorithm
for Plant Disease Detec-tion”, International Journal of Advanced
Research in Electrical, Electronics and Instrumentation Engineering,
Vol. 6, Issue 7, 2017, pp. 5845-5852.

[16] Diptesh Majumdar, Arya Ghosh, Dipak Kumar Kole, Aruna

Chakraborty and Dwijesh Dutta Majumder, “Application of Fuzzy C-
Means Clustering Method to Classify Wheat Leaf Images based on
the presence of rust disease”, Proceedings of the 3rd International
Conference on Frontiers of Intelligent Computing: Theory and
Applications, Vol. 327, 2015, pp. 277-284.

[17] Srdjan Sladojevic, Marko Arsenovic, Andras Anderla, Dubravko

Culibrk and Darko Stefa-novic, “Deep Neural Networks Based
Recognition of Plant Diseases by Leaf Image Classification”,
Computational Intelligence and Neuroscience, Article ID 3289801,
2016.

[18] Sharada P. Mohanty, David P. Hughes and Marcel Salathe, “ Using

Deep Learning for Image- Based Plant Disease Detection”, Frontiers
in Plant Science, Vol. 7, Article 1419, 2016.

[19] Serawork Wallelign, Mihai Polceanu and Cedric Buche, “Soybean

Plant Disease Identification Using Convolutional Neural Network”,
International Florida Artificial Intelligence Research Society
Conference (FLAIRS-31), Melbourne, United States 2018, pp. 146-
151.

[20] Konstantinos P. Ferentinos, “Deep learning models for

plant disease detection and Diagnosis”, Computers and
Electronics in Agriculture, Vol. 145, Elsevier 2018, pp.
311-318.

284

Authorized licensed use limited to: University of Kashmir Srinagar. Downloaded on April 29,2024 at 10:51:38 UTC from IEEE Xplore. Restrictions apply.