A study on real time plant disease diagonsis system

Thakre Gaurav et al, International Journal of Advance Research, Ideas and Innovations in Technology.
© 2017, www.IJARIIT.com All Rights Reserved Page | 1118
ISSN: 2454-132X
Impact factor: 4.295
(Volume 3, Issue 6)
Available online at www.ijariit.com
A Study on Real Time Plant Disease Diagnosis System
Gaurav Thakre
Sinhgad Institute of Technology and
Science, Pune, Maharashtra
gaurav1thakre@gmail.com
Amol R. More
amolmore91@hotmail.com
Kishor S. Gajakosh
kishorsgajakosh@gmail.com
Muktanand O. Yewale
Sinhgad Institute of Technology and Science,
Pune, Maharashtra
yewalemuktanand@gmail.com
D. O. Shamkuwar
Sinhgad Institute of Technology and Science,
Pune, Maharashtra
vah@gmail.com
Abstract: We aim to develop a real time application to the farmers for managing crop diseases. However, disease detection requires
continuous monitoring of experts which might be prohibitively expensive in large farms area. Automatic detection of plant diseases
is an essential research topic as it may prove benefits in monitoring large fields of crops and thus automatically detect the symptoms
of diseases as soon as they appear on plant leaves. Regarding plant disease diagnosis methodologies to detect diseases on crops,
image processing in disease diagnosis and eAGROBOT was studied. This paper is aiming to all are collectively used and formed
semi real time system for a disease diagnosis which uses image processing and data mining concepts to give pesticide
recommendation and pesticide cost estimation system. Thus the android application makes a good foundation for following effective
characteristic parameters for the disease diagnoses and setting up recommender system. The system is to be designed and developed
using Android studio as front-end software and SQLite as back-end software. The pictures and remedial measures of the diseases
were stored in the database and can be retrieved whenever necessary. The challenge is to make the farmers listen to the crop disease
diagnosis system and to get the advice related to the crop diseases. The constraint here is to develop the expert in local languages so
that farmers can operate the ES by themselves and get expert advice from the system.
Keywords: Pesticide, Real Time Plant Disease, Pesticide Cost Estimation System.
1. INTRODUCTION
Diseases are important factors to restrict the growth of crops in agriculture producing, which may reduce yields of crops greatly and
quality of products. At present, the diagnosis of crops diseases mostly depends on manual recognition, but some problem occurs, on the
one hand, it can be mistakenly diagnosed by farmers because they usually judge the symptoms by their experiences. On the other hand,
the disease treatment may be dallied over because the technician or expert can’t go to the locale to diagnose in good time. Relative to
the person’s vision, computer image processing technique take on some characteristics such as speediness, huge information and
distinguish small diversity which can’t be distinguished by person’s eyes, so image processing technique can help farmers to judge the
reasons and severity of crop diseases, and it takes on important theoretical and practical significance for improving the automatic
management of crop. Common symptoms include abnormal leaf growth, color distortion, stunted growth, shriveled and damaged pods.
Although diseases and insect pests can cause considerable yield losses or bring death to plants and it also directly affects human health.

These require careful diagnosis and timely handling to protect the crops from heavy loses. In the plant, diseases can be found in various
parts such as fruit, stems, and leaves. Leaf presents several advantages over flowers and fruits at all seasons worldwide. This enables
machine vision to provide image based automatic detection and guidance. The idea behind creating an application is to enable many
people to benefit from the knowledge of one person - the expert. A proposed application typically has three components viz., knowledge
base, inference engine and user interface. The knowledge base is the component that contains the knowledge obtained from the domain
expert. Normally, the way of representing Knowledge is using rules. The inference engine is the component that manipulates the
knowledge found in the knowledge base as needed to arrive at a result or solution. The user interface is the component that allows the
user to give input to the system and receive the results of those inputs.
In this paper, we propose a novel framework that focuses on the disease detection for plants. The farmers can capture the images of
plant leaves using any mobile camera having a resolution greater than 5 mega pixels. The farmer needs to just capture the image of the
plant leaf through the mobile camera and send the image to a central system. After acquiring images, through the proposed work, normal
and abnormal leaves are classified based upon the extracted features from leaf images. The extracted features, class and the percentage
area of abnormality will be given as input to the DSS. The proposed algorithm will automatically recognize the plant based on the color,
texture, and shape of the leaf. The purpose of the proposed project and research work is to provide inputs for the DSS, developed for
providing advice to the farmers as for when they require over mobile internet.
A. Challenges Faced By Farmers
Plant diseases can cause a significant reduction in both quality and quantity of agricultural products [4]. Typically different types of
diseases are seen at different stages of development of the crops. The rate of spread differs and so does the type of pesticide. Visual
inspection is the main approach adopted in practice for detection and identification of plant diseases [4]. However, this requires
continuous monitoring by experts which might be prohibitively expensive. Further, in some cases, farmers may have to travel a long
distance to contact experts, this makes consulting experts too expensive and time consuming [5][6][7]. Therefore, quick, automatic, less
expensive but accurate method to detect plant disease cases is of great practical significance [5] [6].
B. Current Practices
During a survey conducted among the farmers, they are aware of the more common diseases and seek the help of nearby knowledgeable
farmers or approach dealers with a sample of the infected crop if in case of doubt. Even then, once in three years, some farmers face
extensive crop damage, the main reason being novel diseases and delay in getting critical information about controlling the spread. Over
the years farmers have been using helplines setup by the government, such as the Kisan Call Centre, and are also exploring new solutions
that leverage emerging imaging and M2M technologies The satellite based solutions can indicate the presence of disease but detection
of disease type is more challenging if the spread is limited to 1 acre. Agriculture departments under Government of India on the other
hand plan to take corrective action when a considerable mass of land up to 5 to 8 acres is found infected. In contrast, small-holder
farmers use pesticides when less than 1/3rd of an acre is infected.
C. Opportunity for Image Based Solutions
Both visible and infrared regions of the electromagnetic spectra are used extensively to detect and diagnose defects in the crop as well
as soil. Studies show that machine learning methods can successfully be applied as an effective early disease detection mechanism [4]
[5] [6] [7]. Examples of such machine learning methods that have been applied in agricultural researches include Artificial Neural
Networks (ANNs), Decision Trees, K-means, k-nearest neighbors, and Support Vector Machines (SVMs). Proposed method based on
the application of K-means as a clustering procedure and ANNs as a classifier tool [6] in terms of speed of computation and accuracy.
Two steps are added; one in which green colored pixels are identified for masking and second where pixels with zeros red, green and
blue values together with the pixels on the boundaries of the infected cluster are completely removed. Some solutions are also available
in the microscopic to the telescopic range, such as the ColorPro software for estimation of infected leaf area, chlorophyll, protein, and
bacterial colonies count; CytoPro for chromosome analysis and others by BARC [2]. This has manifested confidence in image based
solutions similar to the satellite solutions.
The rest of the paper is organized as follows: Section 1 gives an introductory part and importance of leaf disease detection. Also describes
various types of leaf diseases and its symptoms. Section 2 describes the literature survey which presents a detailed discussion on recent
work carried out in this area. The proposed approach is illustrated in section 3. The pre-processing steps used in the algorithm are
introduced in section 4 which includes a proposed methodology for leaves disease extraction and classification on various image
processing techniques. Section 5 describes the details regarding system architecture. Finally, section VI concludes the paper.

2. LITERATURE SURVEY
Although professional agriculture engineers are responsible for the recognition of plant diseases, intelligent systems can be used for
their diagnosis in early stages. The expert systems that have been proposed in the literature for this purpose, are often based on facts
described by the user or image processing of plant photos in visible, infrared, light etc. The recognition of a disease can often be based
on symptoms like lesions or spots in various parts of a plant. The color, area and the number of these spots can determine to a great
extent the disease that has mortified a plant. Higher cost molecular analyses and tests can follow if necessary. A Windows Phone
application is described here capable of recognizing vineyard diseases through photos of the leaves with an accuracy higher than 90%.
This application can easily be extended to different plant diseases and different smart phone platforms. But it is using a small training
set for the image detection and also has the limitation of an application platform that is windows based application everyone cannot have
a windows phone so it is not the feasible platform as much concerned with farmer’s use [1].
A Study on the Method of Image Pre-Processing for Recognition of Crop Diseases Focuses on image pre-processing which
can make a good foundation for the effective characteristic parameters for the disease diagnoses and more for setting up pattern
recognition system. It only focuses on the image pre- processing for better segmentation effect. We are aiming to join all functionalities
in which image processing is a mechanism only. Leaves with spots must be pre-processed firstly in order to carry out the intelligent
diagnosis to crop diseases based on image processing and appropriate features should be extracted on the basic of this. And only in this
way, crop diseases can be recognized accurately. So, the image pre-processing is very important. The image pre-processing can make
following extracting of characteristic parameters not to be affected by background, shape, size of leaf, light and camera and it also make
a good foundation for following effective characteristic parameters for the disease diagnoses, as well as setting up pattern recognition
system [2].
eAGROBOT- A Robot for Early Crop Disease Detection using Image Processing is a Real time testing system results are obtained from
cotton and groundnut plantations. The accuracy levels for disease identification for groundnut and cotton plantations are found to be
satisfactory. To achieve better detection accuracy with different plant species we have to use efficient methodology [3].
Detection and Classification of Diseases of Grape Plant Using Opposite Color Local Binary Pattern Feature and Machine
Learning for Automated Decision Support System Identification of plant disease through the leaf texture analysis and pattern recognition
system can be further improved by improving the training ratio [4].
Image segmentation is the key component of identifying plant leaf diseases. Most of the available techniques for leaf disease
segmentation use grayscale values, an automatic seeded region growing (SRG) algorithm for color. The color difference between
adjacent regions is computed using Euclidean distance metric in the algorithm. The look up table is created by traversing the image
vertically and horizontally and any change in the labels of the pixel is noted in the table. The incorporation of the table helps in the better
organization in region merging step and helps in the further segmentation of the image. It must be noted that the performance of colored
image segmentation largely depends on the color space chosen. The algorithm is first implemented in the YCbCr color space and then
implemented in other color spaces like YCgCr, CIELAB, and RGB to check for the best performance of the segmentation algorithm.
Experimental results show that the SRG algorithm along with the proposed modification for region merging gives good results in the
YCbCr comp Check for the best performance of the segmentation algorithm. In some cases of plant disease, some disease spots are
missed and also boundary is not segmented [5].
Unhealthy Region of Citrus Leaf Detection Using Image Processing Techniques to determine the defect and severity areas of
plant leaves. Need further research on test equipment, image processing and analysis method [6].
Development of Expert System to Diagnose Rice Diseases in Meghalaya State is to develop an expert system to the farmers in
Meghalaya state for managing rice disease. Need expertise in generating the result. The decision making process may be delayed in
some cases hence it unfollow the time constraints [7]

3. PROPOSED APPROACH
1. Platform
An image processing technique that can be implemented as a smart phone application is presented in this paper for the recognition of
plant diseases. The described image processing technique can be used either as a standalone application for more accurate diagnoses [2]
[4] [5]. The system isolates the lesions (or spots) that can appear at various parts of a plant like the leaves, or the fruit. The diagnosis is
based on the number of spots, their area, and their color features. These features are compared with predetermined limits in order to
select the matching disease.All this functionality is performed with the help of android application which has latest android platform
compatibility.It uses android SQLite database for database operations. The library information is provided explicitly and some android
libraries are used by the application.
2. Image Processing
It is the key process done after platform concern management, it performs all image processing operations with the help of segmentation
techniques like segmentation and image feature extraction methods.
3. Database Management and Connectivity
Databases are primarily divided into two categories based on the type of information storage one is for image type data storage which
has two type of images stored that is healthy images of crop and defected images of the crop. The second database is for information
storage of crops their diseases, pesticides and their cost database records. Both data interlinked with processors for generating the end
result.
4. Experimentation and Results
The experimentation process includes comparing the extracted image with the image in the database to check the disease after computing
the image is sent to compare with First database images. The generated row then compared with information records to generate the
result report with the pesticide recommendation and their cost estimation system.
ALGORITHM STEPS AND METHODOLOGIES
1. Distance Matrix
Distance Matrix is used to calculate the distance between each pair of species, and then find a tree that predicts the observed set of
distances as closely as possible. The distance matrix is a two dimensional array containing the distances, taken pairwise, between the
elements of a set. The distance matrix, introduced by CavalliSforza and Edwards and by Fitch and Margoliash. They were influenced
by the clustering algorithms of Sokal and Sneath. Distance matrix is used in hierarchical clustering and phylogenetic analysis to carry
out their distance and to calculate exact measures between the points. In general, the merges and splits are determined in a greedy manner.
The results of hierarchical clustering are usually presented in a dendrogram. Manhattan Distance considered by Hermann Minkowski in
19th
Century Germany is a form of geometry in which the usual distance function or metric of Euclidean geometry is replaced by a new
metric. Manhattan Distance is also known as Taxicab metric/rectilinear distance/Snake distance.

Block row distance is defined as
The formula for this distance between a point X=(X1, X2,….Xn) and a point Y= (Y1, Y2,…..Yn) is 𝑑 = ∑ |𝑥𝑖 − 𝑦𝑖|𝑛
𝑖=1
Where n is the number of variables, and Xi and Yi are the values of the ith
variable, at points X and Y respectively.
2. Image Segmentation
Segmentation partitions an image into distinct regions containing each pixel with similar attributes. Segmentation is the first step from
low-level image processing transforming a greyscale or color image into one or more other images to high-level image description in
terms of features, objects, and scenes. Segmentation techniques are either contextual or non-contextual.
3. Non-Contextual Thresholding
Thresholding is the simplest non-contextual segmentation technique. With a single threshold, it transforms a greyscale or color image
into a binary image considered as a binary region map. Non-contextual thresholding groups pixels with no account of their relative
locations in the image plane. Contextual segmentation can be more successful in separating individual objects because it accounts for
the closeness of pixels that belong to an individual object.
4. Region Growing
The bottom-up region growing algorithm starts from a set of seed pixels defined by the user and sequentially adds a pixel to a region
provided that the pixel has not been assigned to any other region, is a neighbor of that region, and its addition preserves uniformity of
the growing region.
Generally, a "good" complete segmentation must satisfy the following criteria:
1. All pixels have to be assigned to regions.
2. Each pixel has to belong to a single region only.
3. Each region is a connected set of pixels.
4. Each region has to be uniform with respect to a given predicate.
5. Any merged pair of adjacent regions has to be non-uniform.
5. K-Means Clustering
The k-means clustering algorithm attempts to split a given individual data set which is the set containing no information as to class
identity into a fixed number (k) of clusters. Initially, k number of so called centroids are chosen. A centroid is a data point of imaginary
or real number at the center of a cluster.
Clustering is the process of partitioning a group of data points into a small number of clusters. For instance, the items in a
supermarket are clustered in categories (butter, cheese, and milk are grouped in dairy products). Of course, this is a qualitative kind of
partitioning. A quantitative approach would be to measure certain features of the products, say the percentage of milk and others, and
products with high percentage of milk would be grouped together. In general, we have n data points 𝑋𝐼,i=1...n that have to be partitioned
in k clusters. The goal is to assign a cluster to each data point. K-means is a clustering method that aims to find the positions 𝑢𝑖,i=1...k of
the clusters that minimize the distance from the data points to the cluster. K-means clustering solves:
𝑎𝑟𝑔 𝑚𝑖𝑛 ∑
𝑘
𝑖=0
∑ 𝑑(𝑥, 𝜇𝑖) = arg 𝑚𝑖𝑛
𝑛
𝑥∈𝑐 𝑖
∑
𝑘
𝑖=1
∑
𝑛
𝑥∈𝑐 𝑖
||𝑥 − 𝜇𝑖||2
2
Where ci is the set of points that belong to cluster i. The K-means clustering uses the square of the Euclidean distance 𝑑(𝑥, 𝜇𝑖)=∥x−μi∥2
2
.
This problem is not trivial (in fact it is NP-hard), so the K-means algorithm only hopes to find the global minimum, possibly getting
stuck in a different solution.
6. Color Transformation
Color can be described by its red (R), green (G) and blue (B) coordinates (the well-known RGB system), or by some its linear
transformation as XYZ, CMY, YUV, IQ, among others. The CIE adopted systems CIELAB and CIELUV, in which, to a good
approximation, equal changes in the coordinates result in equal changes in perception of the color. Nevertheless, sometimes it is useful
to describe the colors in an image by some type of cylindrical-like coordinate system, it means by its hue, saturation and some value
representing brightness. If the RGB coordinates are in the interval from 0 to 1, each color can be represented by the point in the cube in
the RGB space.

Let us imagine the attitude of the cube, where the body diagonal linking “black” vertex and “white” vertex is vertical. Then the height
of each point in the cube corresponds to the brightness of the color, the angle or azimuth corresponds to the hue and the relative distance
from the vertical diagonal corresponds to the saturation of the color.
The present color models have some disadvantages in practical use. E.g. we convert an image in some image processing application into
some brightness-hue-saturation model and we would like to work with individual components (coordinates) as with separate images.
There is desirable regarding the back conversion to have all combinations of the values. It means we need such model, where the range
of values of saturation is identical for all hues. From this point of view, the GLHS color model is probably the best from the current
ones, particularly for wmin = wmid = wmax = 1/3. The good model should satisfy some demands as:
1. The brightness should be a linear combination of all three RGB components. At least, it must be a continuous growing function
of all of them.
2. The hue differences between the basic colors (red, green and blue) should be 120◦
and similarly between the complement colors
(yellow, purple and cyan). The Hue difference between a basic color and an adjacent complement one (e.g. red and yellow)
should be 60◦
.
3. The saturation should be 1 for the colors on the surface of the RGB color cube, it means in case of one of the RGB components
is 0 or 1 except black and white vertices and it is 0 in case of R=G=B.
4. SYSTEM ARCHITECTURE
The architecture is categorized into three sections i.e. real time input section, application processing, and result.Real time input goes
through these three sections and provides us a result. Our aim behind providing this system architecture is that with the help of leaf
image we need to identify disease and recommend pesticides for remove that disease. For that here we need the image processing
because here we are finding disease with the help of crop leaf. For clicking the reliable image user need a good quality camera. After
clicking the image there might be the possibility of noisy data as well as unwanted data, to remove this we are using image processing
technique. After that for identifying the disease we need to extract the features in the activity comes in picture conversion of a colored
image into gray scale image due to gray scale image there are only two colors i.e. white and black. Using that we can identify the disease
on the basis of points present on the leaf and patterns present on the leaf.
Data mining come into the picture for identifying the disease we need to create training data set. In training data set there are multiple
images which have the disease. Which leaf have already occurred disease this all images are stored in training data set. Then we can
compare the extracted image with training data set. Then we need to compare this image with training data set, if the extracted image is
matched with particular training data set image then we can identify the disease. After that, we got pesticides recommendation from the
image. Based on recommending pesticides the cost of pesticides is estimated for particular disease and report is generated as the Disease
diagnosis report.
1. Real Time Input
To identify the disease our system provides the user interface. In that user interface, there is one section which manages input image of
a leaf. This is mobile application the user can capture the image using android mobiles, which have a good camera and sufficient ram
memory. After clicking this image then image goes to in image processing section. Here we assume that users are not familiar with
capturing an accurate image that’s why we are using image processing here, in that we can remove the noisy data which is present in
the leaf image.

2. Application Processing
Now actual working application starts from here. In that section extract the feature of the image that means firstly convert this image
into gray scale image then identify the pattern of points which is present on the leaf. We are working here on the distance of points on a
leaf. We are finding here this distance using city block algorithm. Then we need to find out edges on the leaf this can find out using
segmentation algorithm. After extracting the image it goes into data mining section. In that, we need to maintain the training data set
and data set are stored in the form of a cluster.
3. Result
In training data set we have maintained the different pesticides as per the disease. After knowing the disease our task is how to remove
this disease and crop makes disease free. There are multiple pesticides for one disease, these different pesticides have different cost then
we will recommend pesticides with the cost estimation. The user should select any one pesticides from here and apply for his crop.
5. CONCLUSION
The classification and recognition of crop diseases are of the major technical and economic importance in the agricultural Industry. To
automate these activities like texture, color and shape disease recognition system is feasible. A smart phone application for plant disease
recognition is presented. It is based on image processing that analyzes the color features of the spots in plant parts. Images were acquired
under laboratory condition using a digital camera. The management of plants requires close monitoring especially for the management
of a disease that can affect production significantly and subsequently the postharvest life. The naked eye observation of experts is the
main approach adopted in practice for detection of plant diseases. The system allows the expert to evaluate the analysis results and
provide feedbacks to the famers as a result of their mobile phones. The goal of this application is to develop an image recognition system
that can recognize crop diseases. It can be evaluated on various common crop diseases with an accuracy that exceeds 90% using a small
training set.
REFERENCES
1. Nikos Petrellis,"A Smart Phone Image Processing Application for Plant Disease Diagnosis ", 2017 6th International Conference on
Modern Circuits and Systems Technologies (MOCAST).
2. Geng Ying1, 2 Li Miao Yuan YuanHu Zelin "A Study on the Method of Image Pre-Processing for Recognition of Crop Diseases ",
International Conference on Advanced Computer Control.
3. Sai Kirthi Pilli1, Bharathiraja Nallathambi, Smith Jessy George, Vivek Diwanji,"eAGROBOT- A Robot for Early Crop Disease
Detection using Image Processing", Ieee Sponsored 2nd International Conference On Electronics And Communication System (ICECS
2015).
4. Harshal Waghmare, Radha Kokare,” Detection and Classification of Diseases of Grape Plant Using Opposite Colour Local Binary
Pattern Feature and Machine Learning for Automated Decision Support System”, 2016 3rd International Conference on Signal
Processing and Integrated Networks (SPIN).
5. Rajat Kanti Sarkar, Ankita Pramanik,” Segmentation of Plant Disease Spots Using Automatic SRG Algorithm: A Look Up Table
Approach”, 2015 International Conference on Advances in Computer Engineering and Applications (ICACEA) IMS Engineering
College, Ghaziabad, India.
6. Prof. Ujwalla Gawande , Mr. Kamal O. Hajari, Ms. Kiran R. Gavhale, “Unhealthy Region of Citrus Leaf Detection Using Image
Processing Techniques”, International Conference for Convergence of Technology – 2014.
7. P. Mercy Nesa Ranil, T. Rajesh2 and R. Saravanan3,” Development of Expert System to Diagnose Rice Diseases in Meghalaya State”,
2013 Fifth International Conference on Advanced Computing (ICoAC)

A study on real time plant disease diagonsis system

More Related Content

What's hot (20)

Similar to A study on real time plant disease diagonsis system (20)

Recently uploaded (20)

A study on real time plant disease diagonsis system