IAETSD JOURNAL FOR ADVANCED RESEARCH IN APPLIED SCIENCES ISSN NO: 2394-8442

A Survey on Detection and Computing the Amount of Plant Diseases using

Image Processing

Shravankumar Arjunagi*, Dr. Nagaraj.B.Patil**, Sharanbasappa Honnashetty***

*Department of Computer Science and Engineering, VTU University, Belgaum, Kalaburagi
** Department of Computer Science and Engineering, VTU University, Belgaum, Raichur
*** Department of Computer Science, Gulbarga University, Kalaburagi
* sarjunagi@gmail.com, ** nagarajbpatil1974@gmail.com, *** sharan.sjce@yahoo.com

ABSTRACT : Agriculture has become a means to feed ever growing population. Plant diseases have turned into a big
problem as it can cause significant reduction in both quality and quantity of agricultural products. However, the cultivation
of essential crops for optimum yield and quality can be improved by the aid of technological support. Automatic detection
of plant diseases is an important research topic as it may prove benefits in monitoring large fields of crops, and thus
automatically detect the diseases from the symptoms that appear on the various parts of plant. This paper presents a survey
on automated detection of various diseases associated with crops and also given a proposed methodology for computing
amount of diseases in various crops.

Keywords - crop disease, preprocessing, segmentation, classification

I. INTRODUCTION
Agriculture is the main source of income in India. Almost 70 percent of the population in India depends on farming for
most of the rural people it is an essential mean of livelihood. Sometimes farmers are unable to pay attention to the diseases
or face difficulty in identifying the diseases, which lead to loss of the crop. Each plant disease has different stages of growth,
whenever the disease occurs on a plant, farmers have to keep eyes on the infection. This manual approach of disease
detection is time-consuming and requires some precaution during the selection of pesticides. To overcome by this manual
detection, an automated detection helps formers to quickly identify disease and insist them for selection of pesticides. This
can be achieved by Capturing images of infected part of crop and getting them processed by an automated system could
become an attractive solution for farmers. This paper focuses on how image processing and machine learning is utilized in
detection of diseases in rice plants. The general process of plant disease identification is shown in Fig.1. Major steps of the
processing are as follows. The images are captured from the field, the images are preprocessed, the infected parts are
extracted from the leaf, then features are extracted from the segmented images, and finally classification of the disease is
performed using machine learning techniques later amount of disease can be computed.

Fig1: The general process of plant disease identification and Computing Amount of Disease

VOLUME 4, ISSUE 7, DEC/2017 344 http://iaetsdjaras.org/

IAETSD JOURNAL FOR ADVANCED RESEARCH IN APPLIED SCIENCES ISSN NO: 2394-8442

The performance of a plant disease detection system can be evaluated by measuring the precision of the machine learning
algorithms (classification). To get comparative understanding of various works, we carry out a survey on different techniques
and approaches used for the detection of the plant diseases. We survey both image processing techniques and machine
learning techniques applied on Plant disease classification and presented our methodology. This paper is divided into eight
sections. Section II presents study on different types of plant diseases. Section III presents a survey of different image
processing operations applied for crop disease identification and identifies important findings. Section IV presents related
work carried out and section V presents our proposal about computing the amount of disease in plant. Finally, Section VI
summarizes work in the form of conclusion.

II. DIFFERENT TYPES OF PLANT DISEASES

This section briefly explains different types of diseases that occur on crops and also insist readers to understand about what
types of image processing operations would be needed and what type of features need to be considered to prepare such
disease detection system. The images of frequently occurring diseases in few crops are shown below

Cotton

Cercospora leaf spot: found rapidly in humid, warm and wet conditions, typically after cover closing down. The disease is
caused by the fungal pathogen “Cercospora beticola” The fungus overwinters in infected cotton crop in the field.

Red Spot: on leaf develop by ambient temperature of 68-86 F degree high relatively humidity. Infections occur early season
on the cotyledons. Mature spot usually have concentric ring that are irregular circular with straw colored or reddish brown
centers.

Bacterial blight: the cotton plant resulting from infection by Xanthomonas citri pathovar malvacearum bacteria. Can affect
the cotton plant during all growth stages, infecting stems, leaves, bracts and bolls. It causes seedling blight, leaf spot, blackarm
(on stem and petioles) and boll rot. On cotyledons small, green, water-soaked rounded (or irregular) spots form which turn
brown. Cotyledons can be distorted if the infection is intense. Black and elongated lesions can girdle the hypocotyls and kill
seedlings. On the leaves, scattered small dark-green, water-soaked, areolate spots, form measuring 1–2 mm on the lower
surface.

Myrothecium leaf spot: in cotton manifests as small, circular, tan coloured spots with broad violet to brown margin
surrounded by zones of translucent areas forming concentric rings. After a few days, dark green sporodochia surrounded by
a rim of white hair like mycelia are formed particularly in the region where rings are formed. He further observed that the
spots may coalesce resulting in blighting of the affected tissues

Potato

Late blight: The pathogen is favored by moist, cool environments sporulation is optimal at 12-18°C in water-saturated or
nearly saturated environments, and zoospore production is favored at temperatures below 15° C. Lesion growth rates are

VOLUME 4, ISSUE 7, DEC/2017 345 http://iaetsdjaras.org/

IAETSD JOURNAL FOR ADVANCED RESEARCH IN APPLIED SCIENCES ISSN NO: 2394-8442

typically optimal at a slightly warmer temperature range of 20 to 24° C .late blight is most destructive fungal diseases
Phytophthora infestans is an specifically an oomycete which causes the serious infection to potato and tomato.

Tomato

Septoria leaf spot: It is one of the most destructive diseases of tomato foliage and is particularly severe in areas where wet,
humid weather persists for extended periods. It’s usually appears on the lower leaves after the first fruit sets.

Fungus: Despite tomato plants' popularity, they can be tricky to grow. The finicky plants are susceptible to fungi that can
cause damage from shriveled foliage to rotted fruit. The most common tomato fungi include early blight, late blight, and
anthracnose and septoria leaf spot.

Wheat

Stem rust and leaf rust: The stem, black and cereal rusts are caused by the fungus Puccinia graminis and are significant
diseases affecting cereal crops. Crop species which are affected by the disease include bread wheat, durum, barley these
diseases have affected cereal farming.

Stripe rust: develops in early spring favored by cool temperatures and high humidity. Primary symptoms consist of narrow
orange-yellow stripes on leaves, sheaths, awns and glumes severe infections affect yield by reducing kernel numbers, weight
and overall quality. The fungus Puccinia striiformis causes stripe rust.

Cucumber

Angular leaf: Leaf spot is a serious disease of the cucumber family that results in reduced yields and fruit of poor quality.
You may first notice the disease after a period of warm weather accompanied by rain. Treat angular leaf spot with a
combination of good garden sanitation, fungicides and keeping the plants as dry as possible

Downy mildew: of cucumber is caused by Pseudoperonospora cubensis, a member of the water mold group of fungal-like
organisms. The downy mildew pathogen is an obligate parasite that needs living cucurbit plants in order to grow and survive.
It does not overwinter on plant debris in most areas. Powdery mildew: is a fungus that forms as a white or gray powder on
the leaves of plants causing the leaves to become deformed and die.

Rice

VOLUME 4, ISSUE 7, DEC/2017 346 http://iaetsdjaras.org/

IAETSD JOURNAL FOR ADVANCED RESEARCH IN APPLIED SCIENCES ISSN NO: 2394-8442

Brown spot: is a fungal disease that infects the coleoptile, leaves, leaf sheath, panicle branches, glumes, and spikelets. Its
most observable damage is the numerous big spots on the leaves which can kill the whole leaf. When infection occurs in the
seed, unfilled grains or spotted or discolored seeds are formed.

Leaf blast: Blast is caused by the fungus Magnaporthe oryzae. It can affect all above ground parts of a rice plant: leaf, collar,
node, neck, parts of panicle, and sometimes leaf sheath.

Sheath blight: this leads to the formation of lesions and production of empty grains. Also, initial lesions are water-soaked
to greenish gray and later become grayish.

Leaf scald: is a fungal disease caused by Microdochium oryzae, which causes the scalded appearance of leaves

III. RELATED WORK

The work [18], describes the classification of rice bacterial leaf blight, rice sheath blight and rice Blast disease using support
vector machine classifier. the images are taken from field CCD color camera and these are preprocessed because of
unavoidable that dewdrop, insects’ excrement and dust might appear in the images so median filter is used to remove these
noises. Here images were transformed from a red, green, blue (RGB) color representation to y1=2g-r-b and y2=2r-g-b
representation for segmentation process, later edge of disease spot was removed using Otsu method. For feature extraction
color, shape and texture were considered and area (A) and perimeter (P) of disease spots were calculated from the binary
image of disease spot and adopted classifying rice disease using texture features from the gray level co-occurrence matrix
(GLCM).support vector machine classifier is used and achieved with 97.2% of accuracy.

The work [4], describes the classification of Cotton Leaf Spot images and categorizes the diseases using support vector
machine. The images are taken from digital camera and preprocessed, which includes median filter for noise removal and
conversion. Here the RGB image is converted to a grayscale image, next, the image segmentation based on gray-level
threshold segmentation is adapted and the binary image is gained. Image analysis is done based on color and shape of the
leaves. Blob Analysis is used for shape feature extraction and for color spaces like the HSI and CIE color model is also used.
SVM is used for cotton leaf spot disease classification.

The work [1], describes automated detection of cotton leave disease using K-means clustering algorithm. The images are
taken from database created by Dr. Punjabrao Krishi Vidyapith, Akola. The images are preprocessed to remove noises
generated by light intensity and shadow by contrast enhancement using Histogram equalization. Later Instance-based
classifier such as the k-mean classifier used for database comparison and k-mean clustering algorithm is used for classification
and achieved accuracy of 80.56 %. The work [3], describes detection of cotton leaf disease where features are extracted using
Active Contour Model. The images are acquired by using Cannon A460 digital camera and preprocessed using Low pass
filter for smoothing the images and noise reduction. In this work they used active contour model for segmentation.

The work [17], describes recognition of paddy diseases. Here Image processing starts with the digitized a color image of
paddy disease leaf. Then a method of mathematics morphology is used for segmentation. Then texture, shape and color
features of color image of disease spot on leaf were extracted, and a classification method of membership function was used
to discriminate between the three types of diseases. The analysis of the results showed over 70% classification accuracy.

VOLUME 4, ISSUE 7, DEC/2017 347 http://iaetsdjaras.org/

IAETSD JOURNAL FOR ADVANCED RESEARCH IN APPLIED SCIENCES ISSN NO: 2394-8442

The work [19], describes automated system to classify the leaf brown spot and the leaf blast diseases of rice plant. Here the
images are grabbed using Nikon COOLPIX P4 digital camera in macro mode and preprocessed using mean filtering
technique to remove noise later Otsu’s threshold based segmentation algorithm has been applied to segment the images. For
classification they used two different classifier, first one is Bayes’ Classifier and second one is Support vector machine by
achieving 79.5% and 68.1% of accuracy.

IV. SURVEY AND ANALYSIS OF IMAGE PROCESSING OPERATIONS

APPLIED FOR DISEASE IDENTIFICATION
This section describes image processing techniques used in various works on crop disease detection. This section presents a
survey of 19 papers of crop disease detection including criteria such as type of image, image dataset, Preprocessing,
Segmentation, Feature Extraction, classification and accuracy.

Table: Survey Table

Feature
N Disease/ Classifica Accur
Crop Ref Data Set Preprocessing Segmentation Extractio
o Characteristic tion acy
n

3 disease Histogram
equalization,
Digital 1) Alternaria k-means Color, 89.56
1 Resizing, NN
Camera 2) Cercospora clustering Shape %
Feature
3) Red spot extraction

Cannon

A460 digital 3 disease

camera and 1) Bacterial

Segmentation, Otsu’s
Leica blight Color,
2 thresholding SVM: 90%
Feature
1 Cotton Wild M3C 2) Alternaria Shape
extraction algorithm
stereo 3)
zoom Myrothecium
microscope

3 disease Image
Acquisition,
Cannon 1) Bacterial
Not
A460 Blight Smoothing, Snake Color, Not
3 specifi
digital 2) Segmentation, segmentation Shape specified
ed
camera Myrothecium
Feature
3) Alternaria extraction

VOLUME 4, ISSUE 7, DEC/2017 348 http://iaetsdjaras.org/

IAETSD JOURNAL FOR ADVANCED RESEARCH IN APPLIED SCIENCES ISSN NO: 2394-8442

Image Color,
Mobile Disease restoration k-means
4 Filtering Shape, SVM: 89%
camera 1)Leaf spot, clustering
Segmentation Texture

3 disease

1) Bacterial Enhancement ,
Not
Digital blight Segmentation, k-means Color, Not
5 specifi
Camera 2)Myrotheciu clustering Shape specified
Feature ed
m extraction
3)Alternaria

4 disease Image
Boundary
preprocessing,
1) Foliar leaf ,
spot Segmentation, Otsu Shape, SVM:
digital thresholding
6 2)Alternaria Feature ext. 97.2%
camera Color,
3) Bacterial Statistical
Texture
Blight analysis of
(GLCM)
4) Cercospora disease

fungal diseases

1)Strip Rust,

A CCD 2)Stem Rust, radial

camera thresholding basis
3) Leaf Rust, method and color, function
(Canon 4)Tan Spot,
2 Wheat 7 median filter morphological shape and (RBF) 98.3%.
A3500, 16 5)Pink Snow texture neural
Mega operators
pixels) Mold, network
6) Septoria and
7)Powdery
Mildew

Canon 6D Disease:
digital
camera 1)target spot,
2)angular leaf RGB, YCbCr,
LDASM, K- Not
Cucum 24-bit RGB spot, HSV, LDA,
3 8 L a b . means, Otsu color specifi
ber diseased NN
3) downy and Graph cut ed
images of
4928×3264 mildew and
4)powdery
pixels mildew

VOLUME 4, ISSUE 7, DEC/2017 349 http://iaetsdjaras.org/

IAETSD JOURNAL FOR ADVANCED RESEARCH IN APPLIED SCIENCES ISSN NO: 2394-8442

OTSU

Algorithm,
Digital shape,
camera of cucumber gray level co- color and BP neural
9 median filtering occurrence 80%
Olympus disease network
FE5030 matrix texture
(GLCM),

Radial
Basis
growth rate Function
Digital (RBF)
and Not
Not Not Not
10 Wireless neural
development specifi
Monitoring specified specified specified network,
of sea ed
System
cucumbers nearest
neighbor
clustering

Radial
Basis
Function
(RBF),

polynomia
common color, l and Not
cucumber Not Not
11 digital texture Sigmoid specifi
camera disease leaf specified specified kernel
and shape ed
function,

Support
vector
machine
(SVM)

camera of
maturity and
Euclidean texture, Not
resolution 8 Not KNN-
12 quality for distance, color
mega pixels SVM specifi
cherry specified
at a distance and shape ed
tomatoes
of 20cm

Otsu's
4 Tomato Disease Algorithm,

common 1)tomato late mean and color, binary

Not
13 digital blight, standard shape and decision 97.3%
camera specified deviation, texture tree
2)Septoria

spot, gray-level co
occurrence

VOLUME 4, ISSUE 7, DEC/2017 350 http://iaetsdjaras.org/

IAETSD JOURNAL FOR ADVANCED RESEARCH IN APPLIED SCIENCES ISSN NO: 2394-8442

3)bacterial matrix
spot,

4)bacterial
canker,

5)tomato leaf
curl

Thresholding
Disease Algorithm, Not
common color, Not
RGB, YCbCr,
14 digital 1)Septoria Leaf L a b . k-means shape and specifi
camera Clustering texture specified
Spot ed
Algorithm

histogram Size,
digital nature of equalization Not
Otsu color,
15 camera of technique, 91%
fungus algorithm, location, specified
6MPixel
locus

HIS color
model,

Sony DSC Adaptive color,

Potato Late blight leaf Adaptive Not
5 16 Rx100/13 Thresholding, shape and 96%
disease Thresholding,
Camera texture specified
Holes filling,

color,
Disease
common Digitization, Thresholding, area, Nearest
17 digital 1)Rice Blast , 80%
Quantization Sobel Shape, Neighbor
camera
2)Brown Spot
Length

6 Rice Resolution Area,

Disease reduction, perimeter,
common Support
18 digital 1)Rice Blast , remove noise Otsu’s method contrast, Vector 88%
camera Machine
2)Brown Spot using median roundnes
filter s

VOLUME 4, ISSUE 7, DEC/2017 351 http://iaetsdjaras.org/

IAETSD JOURNAL FOR ADVANCED RESEARCH IN APPLIED SCIENCES ISSN NO: 2394-8442

Radial
distributi Bayes
Disease mean filter for on of hue Classifier 68.1%
common
19 digital 1)Rice Blast , image Otsu’s method from and
Support
camera center to Vector 79.5%
2)Brown Spot enhancement boundary Machine
of spot

V. PROPOSED METHODOLOGY
The proposed work aims at the major issues envisaged in identification of the crop and amount of disease in the crop.
Following are the stages proposed to identification of the crop and amount of disease in the crop.

The images of the crops are captured from the different types of soil from the southern region of India. Usually the image
will be in high resolution, and computation of each image takes lot of time, so it will be resized to lower resolution. The
RGB images of leaves are converted into the space specified in the color transformation structure of machine independent
type. And then Apply fuzzy c-means clustering we identify the mostly green colored pixels. After that, based on specified
threshold value that is computed for these pixels, the mostly green pixels are masked. The pixels with zeros red, green, blue
values were completely removed. Convert the infected (cluster / clusters) from RGB to HSI Translation performing SGDM
Matrix Generation for H and S and Calling the GLCM function to calculate the features. In the feature extraction, the crops
texture and color feature will be extracted of each of the database images and input image. In identification, Support vector
machines (SVMs) is proposed. This involves analyzing a given set of labeled observations (the training set) so as to predict
the labels of unlabelled future data (the test set).By using FCM value we identifying amount of disease in the infected part.

VI. CONCLUSION
Plant diseases can incur tremendous amount of loss in agriculture if enough attention is not given. Using computer and
communication technologies, an automated system can be built which can provide early notification of disease. In the same
direction, we tried to provide our contributions in image processing aspects of such system. We have studied that various
alternatives exist for various operations in image processing. This paper reviewed and summarized techniques of the image
processing that have been used in disease identification. We found that extraction of disease region from the leaf image is
the driving step, for which we have studied and compare various segmentation techniques. We utilized our survey and study,
presented in this paper, to propose our work in the same direction. The paper presented detailed schematic diagram of the
proposed work and discussed important steps. An image processing techniques can give opportunities to researchers to
address problems in various domains that affect to society directly or indirectly.

VOLUME 4, ISSUE 7, DEC/2017 352 http://iaetsdjaras.org/

IAETSD JOURNAL FOR ADVANCED RESEARCH IN APPLIED SCIENCES ISSN NO: 2394-8442

REFERENCES
[1] P. Warne and S.Ganorkar, “Detection Of Diseases On Cotton Leaves Using K-Mean Clustering Method,” International Research Journal
of Engineering and Technology (IRJET), Pp. 425–431, 2015.
[2] P. Rothe and R Kshirsagar, “Svm-Based Classifier System For Recognition Of Cotton Leaf Diseases,” International Journal Of Emerging
Technologies In Computational And Applied Sciences (Ijetcas), Pp. 427–432, 2014.
[3] P. Rothe and R.Kshirsagar “A Study and Implementation of Active Contour Model For Feature Extraction : With Diseased Cotton Leaf
as Example,” International Journal of Current Engineering and Technology, Vol. 4, No. 2, Pp. 812–816, 2014.
[4] S. Patil and R. Zambre, “Classification Of Cotton Leaf Spot Disease Using Support Vector Machine,” International Journal of Engineering
Research, vol. 3, no. 4, pp. 1511–1514, 2014.
[5] D. Abhang, P. Autade, S. Bhokare, and P. Dawange, “A Survey On, Analysis and Detection of Disease on Cotton Leaves,” International
Journal of Advance Research and Innovative Ideas In Education, No. 4, Pp. 236–243, 2015.
[6] “Cotton Distress” [online]. Available: http://cottonmarketnews.com/2015/05/27/india-record-rs-2-5kcr- loss-seen-on-cotton-buying/
[7] Zahra Sarayloo,Davud Asemani, “Designing a classifier for automatic detection of fungal diseases in wheat plant by pattern recognition
techniques” 23rd Iranian Conference on Electrical Engineering (ICEE), 978-1-4799-1972-7/15/$31.00 c_ 2015 IEEE 1193
[8] Na Wu1,2, Miao Li1, Lei Chen1, Yuan Yuan1, Shide Song “A LDA-based segmentation model for classifying pixels in crop diseased
images” Proceedings of the 36th Chinese Control Conference July 26-28, 2017, Dalian, China. This work is supported by National Natural
Science Foundation (NNSF) of China under Grant 31501223. Page no 11499
[9] Yong Wei, Ruokui Chang, Yuanhong Wang, Hua Liu, Yanhong Du, Jianfeng Xu, and Ling Yang, “A Study of Image Processing on
Identifying Cucumber Disease” International Federation for Information Processing 2012, pp. 201–209
[10] Min Sun, Ji Chen,Daoliang Li, “ Water Temperature Prediction in Sea Cucumber Aquaculture Ponds by RBF Neural Network model”
, 2012 International Conference on Systems and Informatics (ICSAI 2012), 978-1-4673-0199-2/12/$31.00 ©2012 IEEE, page no:
1154
[11] Zhang Jian, Zhang Wei. “Support Vector Machine for Recognition of Cucumber Leaf Diseases”, 978-1-4244-5848-6/10/$26.00
©2010 IEEE page no: 264
[12] V.Pavithra, R.Pounroja, Dr.B.SathyaBama, “Machine Vision Based Automatic Sorting of Cherry Tomatoes”, IEEE sponsored 2nd
international conference on eletronics and communication system (icecs 2015), 978-1-4788-7225-8/15/$31.00 ©2015 IEEE page No 271
[13] H. Sabrol and K. Satish, “Tomato Plant Disease Classification in Digital Images using Classification Tree”, International Conference on
Communication and Signal Processing, April 6-8, 2016, India, 978-1-5090-0396-9/16/$31.00 ©2016 IEEE, Page No 1242.
[14] Tanvi Mehra, Vinay Kumar, Pragya Gupta. “Maturity and Disease Detection in Tomato Using Computer Vision”, 2016 fourth
international conference on Parallel, distributed and grid computing (PDGC), 978-1-5090-3669-1/16/$31.00, 2016 IEEE, page no: 399
[15] Shruti and Nidhi Seth, “Fungus/Disease Analysis in Tomato Crop using Image Processing Techniques”, International Journal of Computer
Trends and Technology (IJCTT) – volume 13 number 2 – Jul 2014, ISSN: 2231-2803 Page61
[16] Aparanjita, rudransh Sharma, anushikha singh, malay kishore dutta, kamil riha, petr kriz,
” image processing based automated identification of late blight disease from leaf image of potato crops” 978-1-5090-3982-1/17/$31.00 ©2017
IEEE page no: 758
[17] G. Anthonys, and N. Wickramarachchi. "An image recognition system for crop disease identification of paddy fields in Sri Lanka." In
International Conference on Industrial and Information Systems (ICIIS), 2009, pp. 403-407. IEEE, 2009.
[18] Q. Yao, Z. Guan, and Y. Zhou, "Application of support vector machine for detecting rice diseases using shape and color texture features,"
in International Conference on Engineering Computation, pp. 79-83 200
[19] S. Phadikar, J. Sil, and A. K. Das, "Classification of Rice Leaf Diseases Based on Morphological Changes," International Journal of
Information and Electronics Engineering, vol. 2, pp. 460-463, May 2012.

VOLUME 4, ISSUE 7, DEC/2017 353 http://iaetsdjaras.org/