IDENTIFICATION OF LEAF DISEASES IN
PEPPER PLANTS USING SOFT COMPUTING
TECHNIQUES
Jobin Francis Anto Sahaya Dhas D
Assistant Professor, Vimal Jyothi Professor, Vimal Jyothi Engineering
Engineering College, Kannur College, Kannur
jobinfrancis@vjec.ac.in dr.anto@vjec.ac.in
Abstract- Images convey relevant data and information in biological
sciences. Digital image processing and the image analysis technology
have a vital role in biology and agricultural sectors. Automatic detection
of plant diseases and cultivation of healthy plants is of great importance
and agricultural automation.[1] The case of a plant, the term disease is
defined as any impairment happening to the normal physiological
function , producing characteristic symptoms. The studies of plant
diseases refer to studying the visually observable patterns of a particular
plant. The identification of plants, leaves, stems and finding out the pests
or diseases, or its percentage is found very effective in the successful
cultivation of crops. The naked eye observation is the approach adopted
by many of the farmers for the detection and identification of plant
diseases. It requires continuous monitoring and found less useful on
large farms. Also, the farmers are unaware of non -native diseases. [1,2]
With the aid of imaging technology the plant disease detection systems
automatically detect the symptoms that appear on the leaves and stem of
a plant and helps in cultivating healthy plants in a farm. These systems
monitor the plant such as leaves and stem and any variation observed
from its characteristic features, variation will be automatically identified
and also will be informed to the user. This paper provides an evaluative
study on the existing disease detection systems in plants.
Keywords: Plant diseases, Image Segmentation, masking, disease
detection, feature extraction
INTRODUCTION
The external appearance is the most important quality
attribute of agriculture products. This outer appearance greatly
affects their point of - sale value and consumer’s buying
behavior. Therefore , the quality inspecting and grading systems
are essential in agricultural field to cultivate good healthy plants.
[1]
Plant diseases cause a major production and economic losses
in the agricultural industry. The disease management is a
challenging task.[3] Usually the diseases or its symptoms such as
colored spots or streaks are seen on the leaves or stem of a plant.
In plants. most of the leaf diseases are caused by fungi, bacteria,
and viruses. The diseases caused due to these organisms are
characterized by different visual symptoms that could be
observed in the leaves or stem of a plant[2,4]. Usually, these
symptoms are detected manually. With the help of image
processing, Automatic detection of various diseases can be
Anoop B K
Assistant Professor, Vimal Jyothi
Engineering College, Kannur
anoop013@vjec.ac.in
detected with the help of image processing. [4]Image processing
plays a crucial role in the detection of plant diseases since it
provides best results and reduces the human efforts.
The image processing could be used in the field of agriculture
for several applications. It includes detection of diseased leaf,
stem or fruit, to measure the affected area by disease, to
determine the color of the affected area. Pepper cultivation is one
of the most remunerative farming enterprises in India. Black
pepper is the most commonly used spice in the world. Its
successful growth was reported in areas where the temperature
ranges from 15-400C. The pepper plants give the better
cultivation if sufficient requirement is provided. [5,6, 7]Plant
disease is one of the main causes which degrade the quantity and
quality of the product. The naked eye observation by the experts
is approach usually taken in identification and detection of plants.
[4,8]This approach is time consuming in huge farms or land
areas. This paper discusses the importance of image processing
techniques in detection and identification of plant diseases in the
earlier stages and thereby the quality of the product could be
increased.
LITERATURE SURVEY
Different methodologies had been discussed by various
researchers in detecting and identifying plant diseases.Revathi et
al propossed a system to classify the Cotton Leaf Spot Diseases
Using Image Processing Edge Detection Techniques. The edge
detection based image segmentation had been done initially.
Hiary and Bani –Ahamed carried out the disease detection using a
three-fold mechanism. In this, at first, the infected region was
identified using k-means clustering. Next extracting the features
using color co-occurrence methodology and finally classification
is done by NN’s. Arivazhagan et al., proposed a four step process
for classification of the disease. First a color transformation
structure for the input RGB image is created then the green pixels
are masked and removed using a threshold value in the
segmentation process. Then compute the texture feature using
color co- occurrence method finally, the extracted features are
passed through the classifier. The image segmentation is done
using Otsu method in “Grading Method of Leaf Spot Disease
Based on Image Processing” by ShenWeizheng andWu Yachun.
The RGB image values were converted into HSI color model, and
the value of H plane are taken for the further analysis. Dhaygude
and Kumbhar proposed the detection algorithm consists mainly
four steps the RGB image are converted into HSV plane. The
masking process is done by using the values of mostly green
colored pixels. Next useful segments are extracted out using
segmentation techniques from the value of SGDM matrices
generated. Camargo and J.S. Smith describes an image processing
method that identifies visual symptoms of plant diseases by
analyzing the colored images.
METHODOLOGY
The methodology adopted in this work can be subdivided into
the following set of subsections as shown in the figure below.
Fig 1. Block Diagram Representation
First the leaf set images are taken with an image acquisition
module and then desired preprocessing steps starts. The images
under analysis were then segmented by using threshold based
methods. The segmented results were used for deciding whether
the pepper plant under observation are healthy or not.
Image acquisition
The test leaf images are taken using a digital camera. Then
different image processing techniques are applied to the leaf or
stem which is captured by the camera. Proper illumination should
be provided at the time of illumination. The images of leafs and
stem were taken from the same distance. Image acquisition
module should be placed at a constant distance with the plant
during the time of capturing the images as it affects the accuracy
of the system.
Color Transformation
The RGB images were converted into HSV color space
representation. In the RGB model each color appears in its
primary spectral components of red, green and blue and this
model is based on a Cartesian co-ordinate system. Though RGB
model matches to the human eye in such a way as strongly
perceptive to the primary colors, this model is not well suited for
describing colors in terms that are practical for human
interpretation. To avoid these limitations, the acquired RGB
images were converted into HSV format. After the conversion
from RGB to HSV, the hue and saturation components are taken
for further analysis since these two components contain most of
the information. Computations from RGB to HSV are carried out
on a per pixel basis.
Image Segmentation
The segmentation process is carried out in two phases,
(i)The masking process and (ii)threshold based segmentation.
(i)Masking of green pixels
Masking of pixels means the pixel value in an image set to
zero or some other value. Since the green color indicates the
healthy region of the plant leaves, to find the diseased portion it is
better to avoid green region to increase the speed of processing.
When the intensity of the green pixels is greater than the pre-
defined value, then all those values set to zero. After masking,
pixels with zeros values are discarded. In masking process
diseased portion of the leaf is identified using the H and S plane
values and a value of “1” is allocated to the particular portion.
Rest of the regions a value of “0” is given. As a result a binary
image containing only ones and zeros is obtained. Thus the
diseased area of the leaves can be extracted out.
(ii)Threshold based Segmentation
Similarity based segmentation is adopted to segment the input
images on the basis of similarity in the intensity or gray levels in
an image. This type of threshold based method is simple but very
powerful approach for segmenting images based on an image
space region. This method was usually used for light objects in a
dark back ground or dark objects in a light background.
Threshold based algorithm will choose a proper threshold value T
to divide image’s pixels into several classes and separate objects
from the background. A binary image simply means an image
contains only values of zeros and ones. Then the binary image is
multiplied with the original RGB color image. In this way, the
infected portion of the leaf is extracted.
In this work the leaf mask and the damaged area mask thus
generated are taken for the further analysis. The Damaged Area
Mask was multiplied with the resized RGB image obtained after
the pre-processing step. Since the mask contains only 1’s and 0’s
and in the mask the diseased portion of the leaves are having the
value of 1’s. When this image is multiplied with the RGB image,
only the diseased portions will come out as a result of
multiplication.
Detection of the diseased portion
The segmented output image shows the portion of the leaves
affected by a particular disease. Then the total affected area can
be calculated by counting the no of pixel values in the output
image. The Leaf Mask generated is an image which exactly
separates the leaf portion and the entire background using the
values of the Saturation plane. This Leaf Mask could be used for
finding the total leaf area of a particular leaf. Here a value is
calculated which is called as the Damage Ratio (DR). This ratio
is calculated by dividing the damaged area by the total leaf area.
( )
DR = ×100 %
( )

From the value of Damage ratio the level of disease affected

can be calculated.
Feature extraction
Feature extraction is the process of defining a set of features,
or image characteristics, which will most efficiently or
meaningfully represent the information that was important for
analysis and classification. Feature extraction involves reducing
the amount of resources required to describe a large set of
data.The extracted features are expected to contain the relevant
information from the input data, so that the desired task can be
performed by using this reduced representation instead of the
complete initial data.An important approach to region description
is to quantify its texture content. Texture descriptor provides
measures of properties such as smoothness, coarseness and
Fig. 2 RGB to HSV conversion
regularity. In this work statistical techniques are used to describe
the textures. In this step Grey Level Co-occurrence matrix of the Fig 2 shows the conversion of the input RGB image into the
leaf images are calculated.Grey-level co-occurrence matrix Hue, Saturation and Value planes. The HSV image is further
(GLCM) creates a matrix from image a given image I. This subjected to masking.
matrix creates the GLCM by calculating how often a pixel with
grey -level value i occurs horizontally adjacent to a pixel with the The fig 3 shows the masked output for the given test image.
value j. Each element (i, j) in GLCM specifies the number of
times that the pixel with value i occurred horizontally adjacent to
a pixel with value j. If I be an intensity image, GLCM scales the
image to eight grey-levels. Texture features can be extracted
using the GLCM values.
Disease classification
The classification process is done by adopting Neural
Network. Back propagation method is considered under the
supervised learning mechanism. The feed forward back
Propagation Neural Network is generally consisting of three
layers such as an input layer, a hidden layer, and an output layer.
The neural network is to be trained with the available data.
RESULTS AND DISCUSSIONS
The image analysis was carried out using MATLAB R2012A.
In this work, a set of twenty leaves were used for testing. Fig 1
shows the color conversion of a test input from RGB to HSV
image.

Fig3. Leaf mask and damaged area mask

 The fig 4 shows the output of the detected spot of disease.
Fig 4.Window showing the diseased portions
The fig 5 shows the results of the test images for a set of 16
leaves in that 10 leaves are unhealthy and 6 leaves are healthy.
Fig5. Leaf images taken for disease detection and identification
Table 1 Damage Ratio (DR) Results
If the DR value is set to a particular threshold value then a
comparison can be made on the amount of area affected by the
diseases. If the value of DR is less than some threshold, say five
percentages or less, then it is known that that leaf or that
particular plant is less affected by the disease and further
diagnosis can be given and the disease can be nullified earlier. If
the DR value is greater, then it is assured that leaf is above the
specified limit and the disease affected become severe. Then that
particular plant can be replaced from the set of plants. In this way
the healthy plants can be extracted out from a farm or from
garden with the help this algorithm.
The algorithm identifies two types of diseases with the help
of neural networks. Berry spot Disease and Quick wilt Disease.
Berry spot disease is a type of fungal disease found in pepper
plants when a plant is affected by this particular type of disease
the fertilizers can be added to rectify the problem. Fertilizers such
as Pseudo- monash can be added to this type of plant. Quick-wilt
disease is a type of disease due to the deficiency of minerals such
as nitrogen, magnesium and potassium. So when this type of
disease is identified the system will identify the type of disease
and the system will prescribe to add minerals to the particular
plant. The following table shows the results obtained after testing.
Twenty leaves were given for testing and the system correctly
detected the diseased leaves and identified the two types of
diseases. The results are summarized in table 2.
Table 2. disease identification results
The images are taken within a good light-up environment and
keeping a constant distance between the acquisition device and
the leaf. In future the images taken in varying lighting condition
is taken for the further modification of the algorithm. Since the
shape of the leaf determines the quality as well as the species of a
plant by the shape identification the type of the plant also can be
identified.
CONCLUSION
An image processing algorithm to find the disease
detection and identification is proposed. The pepper plant leaves
are taken as the set of leaves in detecting leaf diseases. The
algorithm produces better results and healthy and unhealthy
plants can be differentiated with the help of this algorithm. With
this image analysis technique good healthy pepper plants can be
extracted out from a cultivating farm which increases the
productivity and the quality of the pepper plants also can be
assured. This algorithm helps in identifying the presence of
diseases by observing the visual symptoms seen on the leaves of
the plant.
REFERNCES
[1] Sanjeev S Sannakki, Vijay S Rajpurohit, V B Nargund, Pallavi Kulkarni,
“Diagnosis and Classification of Grape Leaf Diseases using Neural
Networks”,IEEE proceedings of 4ICCCNT, 2013
[2] S. Arivazhagan, R. NewlinShebiah, S. Ananthi, S. Vishnu Varthini, “Detection
of unhealthy region of plant leaves and classification of plant leaf diseases using
texture features”, CIGR Journal ,Vol. 15, No.1, 2013.
[3] Arti N. Rathod, BhaveshTanawal, Vatsal Shah, “Image Processing Techniques
for Detection of Leaf Disease”, Vol 3, Issue 11, 2013.
[4] Jayamala K. PatilBharti, “Advances in image processing for detection of plant
diseases”, Journal of Advanced Bioinformatics Applications and Research, Vol 2,
Issue 2, pp 135-141, 2011.
[5] Camargoa, J.S. Smith, “An image-processing based algorithm to automatically
identify plant disease visual symptoms”, Biosyst Eng., Vol 102:9–21, 2009.
[6] SindhujaSankarana, AshishMishraa, Reza Ehsania, Cristina Davisb, “A review
of advanced techniques for detecting plant diseases”, Computers and Electronics
in Agriculture, Vol 72, pp. 1–13, 2010.
[7] Niketamoda, Bharat jadhav, Smeetanaikwadi, “Detection and classification of
plantdiseases by image processing”, International Journal of Innovative Science,
Engineering & Technology, Vol. 1 Issue 2, 2014.
[8] P.Revathi, M.Hemalatha, “Classification of Cotton Leaf Spot Diseases
Using Image Processing Edge Detection Techniques” IEEE proceedings of
INCOSET, pp 169-173, 2012.
[9]H. Al-Hiary, S. Bani-Ahmad, M. Reyalat, M. Braik and Z. ALRahamneh,
“Fast and Accurate Detection and Classification of Plant Diseases” IJCA, vol.
17(1), pp. 31-38, 2011.
[10] ShenWeizheng, Wu Yachun, Chen zhanliang, Wei Hongda, “Grading
Method of Leaf Spot Disease Based on Image Processing” ICCSS, pp. 491-494,
2008.
[11] Mrunalini R. Badnakhe, Prashant R. Deshmukh“Infected Leaf Analysis and
Comparison by Otsu Threshold and k-Means Clustering” IJARCSSE, vol. 2(3),
pp. 449-452, 2012.
[12] Joanna Sekulska-Nalewajko, JaroslawGoclawski, “A semi-automatic method
for the discrimination of diseased regions in detached leaf images using fuzzy c-
means clustering”. Polyana-Svalyava (zakarpattya), Ukraine: pp.172-175, 2011
[13] Stephen Gang Wu, Forrest Sheng Bao, Eric YouXu, Yu – Xuan Wang Yi –
Fan Chang, “A LeafRecognition Algorithm for Plant Classification Using
Probabilistic NeuralNetwork”, IEEE 7th InternationalSymposium on Signal
Processing and InformationTechnology, 2007
[14] S.S. Abu-Naser, K. A. Kashkash and M.Fayyad, “Developing an Expert
System for Plant Disease Diagnosis”, Journal of Artificial Intelligence1(2):78-85,
2008.
[15] K.Attakitmongcol, A.Srikaew, “Grape leaf disease detection from color
imagery system using hybrid intelligent system”, IEEE proceedings of ECTICON,
PP-513-516, 2008.

[16] Rakesh Kaundal, Amar S Kapoor, Gajendra PS Raghava, “Machine learning

techniques indisease forecasting: a case study on rice blastprediction”, BMC
Bioinformatics, Vol. 7:485, 2006.