Abstract—In all of these tasks are processed

Abstract—In agriculture
research of automatic leaf disease detection is essential research topic as it
may prove benefits in monitoring large fields of crops, and thus automatically
detect symptoms of disease as soon as they appear on plant leaves. The term
disease is usually used only for destruction of live plants. This paper
provides various methods used to study of leaf disease detection using image
processing. The methods studies are for increasing throughput and reduction
subjectiveness arising from human experts in detecting the leaf
disease1.digital image processing is a technique used for enhancement of the
image. To improve agricultural products automatic detection of symptoms is

Leaf disease, Image processing, SVM, segmentation, morphological processing,
features extraction, neural networks, clustering, fuzzy logic

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!

order now


India is an agricultural country. Farmers have wide range
of diversity to select suitable fruit and vegetable crop. Research work
develops the advance computing system to identify the diseases using infected
images of various leaf spots. Images are captured by digital camera mobile and
processed using image growing, then the part of the leaf sport has been used
for the classification purpose of the train and test. The technique evolved
into the system is both Image processing techniques and advance computing

Image Analysis Can Be Applied For The Following Purposes:

1. To detect diseased leaf, stem, fruit.

2. To quantify affected area by disease.

3. To find the boundaries of the affected area.

4. To determine the color of the affected area.

5. To determine size & shape of leaf.

6. To identify the Object correctly.


Disease management is a challenging task. Mostly diseases
are seen on the leaves or stems of the plant. Precise quantification of these
visually observed diseases, pests, traits has not studied yet because of the
complexity of visual patterns. Hence there has been increasing demand for more
specific and sophisticated image pattern understanding 1.



Various Types Of Leaf Spot Diseases


– Fungal

– Viral




Most leaf diseases are caused by fungi, bacteria and
viruses. Fungi are identified primarily from their morphology, with emphasis
placed on their reproductive structures. Bacteria are considered more primitive
than fungi and generally have simpler life cycles. With few exceptions,
bacteria exist as single cells and increase in numbers by dividing into two
cells during a process called binary fission viruses are extremely tiny
particles consisting of protein and genetic material with no associated protein
9. In biological science, sometimes thousands of images are generated in a
single experiment. There images can be required for further studies like
classifying lesion, scoring quantitative traits, calculating area eaten by insects,
etc. Almost all of these tasks are processed manually or with distinct software
packages. It is not only tremendous amount of work but also suffers from two
major issues: excessive processing time and subjectiveness rising from
different individuals. Hence to conduct high throughput experiments, plant
biologist need efficient computer software to automatically extract and analyze
significant content. Here image processing plays important role 1. This paper
provides a survey to study in different image processing techniques used for
studding leaf diseases.





                      Various types of diseases


In paper 1 authors present image processing technique
for Rice disease identification and considered the two most common diseases in
the north east India, namely Leaf Blast (Magnaporthe Grisea) and Brown Spot
(Cochiobolus Miyabeanus). Image acquisition is basic step, after that author
use segmentation, boundary detection and spot detection method for feature
extraction of the infected parts of the leave. In this paper author introduces
zooming algorithm in which SOM (Self Organising Map) neural network is used for
classification diseased rice images. There are two methods to make input vector
in SOM. First method is the padding of zeros and the second method is the
interpolation of missing points. For fractional zooming to normalize the spots
size, interpolation method is applied. Image transformation in frequency domain
does not give better classification. For testing purposes, four different types
of images are applied; the zooming algorithm gives satisfactory results of
classification for test images.

In paper 2 authors present image-processing technique
for Leaf & stem disease detection. The author used a set of leaf images
from Jordan’s Al-Ghor area. The five plant diseases namely: Early scorch, Ashen
mold, Late scorch, Cottony mold and Tiny whiteness is tested by image
processing technique. In this technique at starting, image acquisition is
obtained and then K-Means clustering method is used for segmentation. After
that in feature extraction, CCM (Colour Co-occurrence Method) is used for
texture analysis of infected leaf and stem. Lastly paper presents Back
propagation algorithm for neural network in classification of plant diseases.
Result of this image processing technique shows accurate detection and
classification of plant diseases with high precision around 93%.

In paper 3 authors used both LABVIEW and MATLAB software
for image processing to detect chili plant disease. This combined technique
detects disease through leaf inspection in early stage. The Image is captured
using LABVIEW IMAQ Vision and MATLAB is used for further


operations of image processing. Image pre-processing
operations are Fourier filtering, edge detection and morphological operations.
In feature extractions, the color clustering is used to distinguish between
chili and non-chili leaves. Then image recognition and classification determine
the healthiness of each chili plant. This technique results in reducing use of
harmful chemicals for chili plant which reduces production cost and increases
high quality of chili.

In paper 4 authors present image processing technique
for detecting the Malus Domestica leaves disease. Intensity values of grayscale
images are obtained by histogram equalization method. In image segmentation,
Co-occurrence matrix method algorithm is used for texture analysis and K-means
clustering algorithm is used for color analysis. Texture analysis is
characterization of regions in an image by texture content. Color analysis
refers to minimizing the sum of squares of distance between objects and class
centroid or corresponding cluster. In threshold matching process individual
pixels value is compared with threshold value, if value is greater than threshold
then it is marked as object pixel. The texture and color analysis images are
compared with the previous images for detection of plant diseases. Author will
use Bayes and K-means clustering in future.

In paper 5 authors present image processing techniques
for detecting the Bacterial infection in plant. Common infection seen on plant
is Bacterial leaf scorch and early detection of this helps in improvement of
plant growth. The image processing starts with image acquisition which involves
basic steps such as capturing of image and converting it to computer readable
format. Then clustering is done to separate foreground and background image
with help of K-means clustering method in image segmentation. Clustering is
based on intensity mapping and leaf area highlighting is done by subtracting
the clustered leaf images from base images. Compared to Fuzzy logic, K-means
clustering algorithm is simple and effective in detecting the infected area
with reduced manual cluster selection requirement. With ADSP target boards and
FPGA tools, further implementation is possible.

In paper 6 authors present image processing technique
for detection of unhealthy region of Citrus leaf. There are four types of
citrus diseases namely: (i) Citrus canker, (ii) Anthracnose, (iii)
Overwatering, (iv) Citrus greening. Author proposed methodology in which image
acquisition is first step for capturing image by digital camera in high
resolution to create database. Color space conversion and image enhancement is
done in image pre-processing. Discrete cosine transform domain is used for
color image enhancement. YCbCr color system and L*a*b* color space are chosen
for color space conversion. In feature extraction author present statistical
method, using Gray-Level Co-Occurrence Matrix (GLCM) to see statistics such as
contrast, energy, homogeneity and entropy using graycoprops function. Two types
support vector machine (SVM) classifiers: SVMRBF and SVMPOLY are used for
differentiating citrus leaf diseases.

In paper 7 authors present image processing technique
for Orchid leaf disease detection. Black leaf spot and Sun scorch are two types
of orchid leaf diseases mostly found. The basic step of image processing is
image acquisition for capturing images and stores it in computer for further operation.
Image pre-processing involves histogram equalization, intensity adjustment and
filtering for enhancing or modifying the image. Three morphological processes
are used in border segmentation technique for remove small object and preserve
large object in image. Thresholding in segmentation is used for start and stop
point of line to trace edges. Author added ROI (region of interest) in GUI.
After the border segmentation process a classification is done by calculating
white pixels in image. This system gives high accuracy and low percentage of
error in result.

In paper 8 authors present image processing technique
for Tomato leaves diseases detection. In image acquisition phase, digital
images of infected tomato leaves are collected which include two types of
tomato diseases namely: Early blight and Powdery mildew. In pre-processing
phase some techniques are techniques are applied for image enhancement,
smoothness; remove noise, image resizing, image isolation, and background
removing. Author introduced Gabor wavelet transformation and Support vector
machine for identification and classification of tomato diseases. In feature
extraction phase with the help of Gabor wavelet transform feature vectors are
obtained for next classification phase. In classification phase, support vector
machine (SVM) is trained for identifying the category of tomato diseases. The
inputs of SVM are feature vectors and corresponding classes, whereas the
outputs are the decision that detect tomato’s leaf disease. SVM is employed using
Invmult Kernel, Cauchy Kernel and Laplacian Kernel functions. Grid search and
N-fold cross-validation techniques are used for performance evaluation.

In paper 9 authors described disease detection, in which
image processing is first step for obtaining image in digital form and
pre-processing to remove noise and other object from image. Pre-processing also
convert RGB images into grey images using equation f(x) = 0.2989*R + 0.5870*G +
0.114*B and makes histogram equalization. Image segmentation is done using
boundary and spot detection algorithms for finding infected part of leaf.
Classifications of objects are done using K-means clustering method. Otsu
threshold algorithm is used for thresholding which creates binary images from
grey images. With the help of feature extraction color, texture, morphology,
edges are used in plant disease detection. Leaf color extraction using H &
B components and Color co-occurrence method are feature extraction methods in
image processing. Classifications of diseases are done using artificial neural
network (ANN) and Back propagation network.

In paper 10 authors present image processing technique
to detect Scorch and Spot diseases of plant. First step is RGB image
acquisition of plant. Then in pre processing color transformation structure is
created and color values in RGB are converted to the space. The masking of
green-pixels is done after applying K-means clustering. This removes masked
cells inside the boundaries of infected clusters. Image segmentation is done to
obtain the useful segments in image. In feature extraction, color, texture and
edge features are computed using color co-occurrence methodology. Neural
Networks is configured for recognition and classification of diseases. Future
work will include analyzing citrus trees disease conditions in outdoor

In paper 11 authors present image processing technique
for Groundnut plant disease detection. Groundnut plant has two major diseases
namely: Early leaf spot (Cercospora) and Late leaf spot (Cercosporidium
personatum). After obtaining leaf images in RGB are converted to HSV color
images. Green colored pixels in image are found out to reduce processing time.
In color and texture feature extraction analysis, co-occurrence matrices
technique is used. In texture feature extraction there are two ways to analyze
the texture images. First method is structured approach and second method is
statistical approach. Author used statistical approach in this paper. Back
propagation algorithm is applied for classification and recognition of
groundnut diseases. In back propagation two type of phase are there namely: 1)
propagation and 2) weight update. Authors classified four different diseases
with 97 % of efficiency.

In paper 12 authors described plant disease recognition
technique, in which first phase is to create color transformation structure for
the RGB leaf image and convert color values from RGB to the space specified in
that structure. Then apply color space transformation and image is segmented
using the K-means technique. In the second phase called as Masking of green
pixels, the unnecessary part such as green area within leaf area is removed. In
third phase authors calculate the texture features for the segmented infected
object also remove masked cells inside the boundaries of the infected cluster.
Infected cluster are converted from RGB to HSI and SGDM matrix is generated for
H and S. In the fourth phase GLCM function is used to calculate the features
and compute of texture statistics. Finally, the extracted features are passed
through pre-trained neural network for disease recognition.

In paper 13 authors present image processing technique
for detecting disease of Sugarcane leaf. Authors choose 6 type of disease for
experiment, they are: Brown Spot, Downy mildew, Sugarcane Mosaic, Red stripe,
Red rot and Downy Fungal. In image acquisition, images are captured in better
quality resolutions with format such as TIF, PNG, JPEG and BMP for
image-analysis. In pre processing RGB images are converted to grayscale and
unwanted part of data from the images is removed. Segmentation locates healthy
area ofgiven image which contains green pixels and potentially infected area.
Three algorithms namely: Linear SVM, Non linear SVM and Multiclass SVM are used
in feature extraction for disease detection







Techniques Used

1 Rice Disease identification using Pattern Recognition Techniques

Zooming algorithm, SOM                 neural network

2 A Framework for Detection and Classification of Plant Leaf and Stem Diseases

K-Means clustering, Back propagation algorithm, CCM

3 Feasibility Study on Plant Chili Disease Detection Using Image Processing Techniques

Morphological processing,        Color clustering, LABVIEW          IMAQ

4   Remote   Area   Plant   Disease Detection Using Image

CCM,            K-Means clustering

5 A Novel
Algorithm for Detecting Bacterial Leaf Scorch (BLS)
of Shade Trees Using Image Processing

K-means clustering algorithm, Intensity mapping

6 Unhealthy Region of Citrus Leaf Detection Using Image
Processing Techniques

GLCM,            SF-CES,
SVMRBF                    &
SVMPOLY classifier

7 Orchid Leaf Disease Detection using Border Segmentation Techniques

Border segmentation, Pattern classification

8 Tomato leaves diseases detection approach based on support vector machines

SVM, Gabor wavelet transform

9  Plant
 Disease  Detection  Using Image Processing

Otsu  thresholding, ANN,    SVM,  
propagation network

10 Advance in Image Processing for Detection of Plant Diseases

CCM,                Neural network

11 Groundnut Leaf Disease Detection and Classification by using Back Probagation Algorithm

CCM,                    Back propagation algorithm

12 Plant Leaf Disease Detection and Classification Using Image Processing Techniques

RGB    to    HSI,  
means clustering, SGDM              Matrix, GLCM

13 Leaf Disease Detection and Prevention Using Image processing using Matlab

Linear SVM, Non Linear SVM and Multiclass SVM













This paper gives the survey on leaf disease detection and
classification techniques using image processing. Different authors used
different algorithms for accurate detection of diseases. Advantage of using
image processing method is that the leaf diseases can be identified at its
early stage. For improving recognition rate, most of researchers used artificial
neural networks and classifiers like ANN, SVM, etc. All methods in this paper
save time and provide efficient result.

Go Top

I'm Eleanor!

Would you like to get a custom essay? How about receiving a customized one?

Check it out