PLANT LEAF DISEASE

DETECTION USING
IMAGE PROCESSING
Presented by Gouri Chatterjee
2nd Semester
M.Tech Electronics and Communication Engineering

=========*******************************************=======

 Under the guidance of Dr. Satadal Saha (Professor) &

Ms. Susmita Sarkar (Assistant Professor)

Department of Electronics and Communication Engineering

MCKV INSTITUTE OF ENGINEERING

AGENDA / OVERVIEW

• PROBLEM STATEMENT

• INTRODUCTION

• LITERATURE SURVEY

• OBJECTIVE OF MY WORK

• DATASET PRESENTATION

• METHODOLOGY

• PROPOSED APPROACH

• REFERENCES
PROBLEM STATEMENT

An expected 70% to 80% Indian economy relies on agribusiness .there is developing Indian population . Which is increasingly

dependent on the agricultural yield. The end goal is kept in mind to develop progressively the diseases need to be examined in

earlier. Diseases are investigated utilizing different image processing techniques and diagnosed so that farmers can overcome

from yield and financial loss.

INTRODUCTION
Developed Technologies have provided the ability to produce sufficient food to meet the demand of society. But still, the safety and
security of the food or crops remained unattained. Factors like change in climate, the decline in pollinators, Plant disease, and others are
challenging to the farmers. An important foundation for these factors needs to be attained on a priority basis. Making use of analysis and
detection processes using present technology helps the farmers to get rid of such problems. The key issues and challenges are identified
by the researchers and the scientists, while analyzing the leaf diseases of plant. Some of them are as follows:

1. The quality of the leaf image must be high.

2. Publicly available Dataset requirement.

3. Noisy data affecting the leaf samples.

4. Through the process of segmentation, diseases may be identified but the samples must undergo training and testing.

5. Classification is one more challenge, in the stage of detecting the leaf diseases.

6. Color of the leaves may be varied due to environmental effect.

7. Variety of diseases can be seen in various kinds of plants, so detection of disease is quite difficult.

With the advancement of computing technology, machine learning and image processing can automatically detect and identify plant
diseases, playing a significant role in the automatic diagnosis of plant diseases8,9. Researchers have applied image processing and
machine learning to identify and categorize plant diseases.
LITERATURE SURVEY
 VARIOUS RESEARCHERS HAVE USED MACHINE LEARNING AND IMAGE PROCESSING TECHNIQUES FOR IDENTIFYING THE DISEASES
ON DIFFERENT TYPES OF PLANTS. THE AUTHORS OF PAPER [1] INVESTIGATED USING K-MEANS CLUSTERING METHOD FOR
BRINJAL LEAVES WITH IMAGE PROCESSING TECHNIQUES TO IDENTIFY PLANT LEAF DISEASE. THE AUTHORS PERFORMED
HISTOGRAM EQUALIZATION TO INCREASE IMAGE QUALITY PRIOR CLUSTERING PROCESS. COLOR CO-OCCURRENCE METHOD (CCM
METHOD) WAS USED TO EXTRACT THE COLOR AND TEXTURE FEATURES. THE FEATURES WERE TRAINED USING K-MEANS
CLUSTERING ALGORITHM WITH THREE CLUSTERS AS INFECTED OBJECT, INFECTED LEAF AND THE BLACK BACKGROUND OF LEAF.

 THE AUTHORS OF THE PAPER [2] PROPOSED CONVERTING RGB IMAGE INTO HSV AND PERFORM COLOR BASED SUBTRACTION OF
UNWANTED BACKGROUND BY RETAINING PIXEL HAVING G VALUE MORE THAN R AND B VALUES FOR PLANT LEAF DISEASE
CLASSIFICATION. THE CONNECTED ELEMENTS IN THE IMAGE ARE DISCOVERED OUT FROM THE CLUSTER BASED BACKGROUND
SUBTRACTION AND THE IMMENSE PART OF THE IMAGE IS KEPT AND OTHER PART IS REMOVED.

 THE AUTHORS OF THE PAPER [3] PROPOSED KNN AS AN EFFECTIVE METHOD IN IDENTIFYING LEAF DISEASES FOR AGRONOMICAL
CROP IMAGES. THEY USED LUMINANCE AND LINEAR CHARACTERISTICS IMAGE TO DETECT SKELETON OF LEAVES TO DETERMINE
WHETHER THE LEAF IS OF GRAPE OR NOT. THEN, GLCM (GRAY-LEVEL CO-OCCURRENCE MATRIX) FEATURES ARE EXTRACTED AND
DISEASES ARE CLASSIFIED BY USING THE OBTAINED GRAPE LEAF IMAGES.
LITERATURE SURVEY
 THE AUTHORS OF THE PAPER [4] PERFORMED CONVOLUTIONAL NEURAL NETWORK OPERATION FOR PLANT DISEASE DETECTION
USING PYTHON API THEY RESIZED IMAGE TO 96X96 RESOLUTION FOR IMAGE PROCESSING. DATA AUGMENTATION TECHNIQUE
WAS USED TO ROTATE, FLIP, SHIT IMAGES HORIZONTALLY AND VERTICALLY. THEY TRAINED THE IMAGE WITH 75 EPOCHS USING
32 BATCH SIZES FOR 35000 IMAGES.

 SIMILARLY, THE AUTHORS OF PAPER [5], PROPOSED FRAMEWORK RESNET50, RESNET101, DENSENET161, AND DENSENET169 AS
THEIR DEEP NEURAL NETWORK (DNN) FRAMEWORK TO DETECT DISEASE IN RICE PLANT. IMAGES WERE RESIZED AS 224 × 224
PIXELS, THE BATCH SIZE WAS SET TO 64 , EPOCH TO 15 AND THE LEARNING RATE WAS SET A CONSTANTLY OF 0.0001. THE
DENSENET161 PRODUCED THE BEST RESULTS WITH AN ACCURACY OF 95.74%.

 THE AUTHORS OF THE PAPER [6] INVESTIGATED ON USING K-MEANS CLUSTERING FOR THE IMAGE SEGMENTATION OF GRAPE
LEAF DISEASE. SHAPE, COLOR AND TEXTURE WERE EXTRACTED AS MAIN FEATURES. LINEAR SUPPORT VECTOR MACHINE (LSVM)
WAS USED FOR CLASSIFICATION PURPOSE. THE IMAGES WERE CLASSIFIED INTO TWO CLASSES DOWNY AND POWDERLY USING
THE EXTRACTED NINE TEXTURE FEATURES AND NINE COLOR FEATURES FOR ALL THREE SEGMENTED PARTS OF SINGLE LEAF
IMAGE.

 THE AUTHORS OF THE PAPER [7] PROPOSED USING THREE FEATURE DESCRIPTORS HU MOMENTS, HARALICK TEXTURE AND
COLOR HISTOGRAM FOR PLANT DISEASE CLASSIFICATION USING VARIOUS MACHINE LEARNING ALGORITHMS LOGISTIC
REGRESSION, SUPPORT VECTOR MACHINE, K-NEAREST NEIGHBOR, CART, RANDOM FORESTS AND NAIVE BAYES
OBJECTIVE OF MY WORK
My objective of work is to basically find & detect generic diseases of leaves commonly found in West Bengal through relevant technique or
methods.

West Bengal, one of the developing states in India, agriculture plays a noteworthy role. Agricultural intervention in the livelihood of rural
India indulges by about 58%. Among the agricultural products, crops such as these crops, rice, wheat, maize etc the most used crops.
Thus, preventing significant loss in quantity and yield of these crops is majorly dependent on recognition and classification of diseases
these plant might possess.

As per my findings some common diseases of West Bengal plant leaves are:

• Bacterial leaf spot

• Downy mildew

• Powdery mildew

• Mould

Mostly live plants are adversely affected by the diseases. This paper imparts representation of leaf disease detection employing a
technique that can identify drawbacks in these crops plant from images, based on colour, bound and texture to give the brisk and reliable
results to the farmer.
DATASET PRESENTATION
 I’m able to collect few samples of common leaves of West Bengal to present as pictured below along with the disease
detected:
 Curry plant leaf :

Disease: Powdery Mildew, which is a fungus that grows on leaves especially if the leaves are overcrowded
or if the sunlight falling on the plant is less or watering is more.
DATASET PRESENTATION
 Pumpkin plant leaf

Disease: There are three diseases that all pumpkin growers need to scout for weekly to prevent major losses. These
diseases are downy mildew, powdery mildew, and plectosporium blight (white blight)
DATASET PRESENTATION
 Bitter gourd plant leaf

Disease: Powdery mildew is one of the major diseases in bitter gourds in all parts of the India. Though it is a major
disease, now it is less common due to introduction of powdery mildew resistant. This disease predominantly prevalent
in the cooler spring and early summer months.
DATASET PRESENTATION
 Amaranth plant leaf :

Disease: Amaranth leaves can be affected by a number of diseases, including anthracnose, wet rot and stem canker were
observed in 91%, 40 and 47% respectively of farms visited.
METHODOLOGY
PROPOSED APPROACH

• The underlying approach for all of the existing techniques of image classification is almost the
same.

• First digit images are acquired from environmental around the sensor using a digital camera .

• The image processing techniques are applied to extract useful features that are necessary for
future analysis of these images.

• After that, several analytical discriminating techniques are used to classify the images
according to the specific problem at hand.
REFERENCES
[1] R. Anand, S. Veni, and J. Aravinth, “An application of image processing techniques for detection of diseases on brinjal leaves using k-means
clustering method,” in 2016 International Conference on Recent Trends in Information Technology (ICRTIT), Chennai, 2016, pp. 1–6, doi:
10.1109/ICRTIT.2016.7569531.

[2] C. G. Dhaware and K. H. Wanjale, “A modern approach for plant leaf disease classification which depends on leaf image processing,” in 2017
International Conference on Computer Communication and Informatics (ICCCI), Coimbatore, 2017, pp. 1–4, doi: 10.1109/ICCCI.2017.8117733.

[3] N. Krithika and A. G. Selvarani, “An individual grape leaf disease identification using leaf skeletons and KNN classification,” in 2017
International Conference on Innovations in Information, Embedded and Communication Systems (ICIIECS), Coimbatore, 2017, pp. 1–5, doi:
10.1109/ICIIECS.2017.8275951.

[4] S. V. Militante, B. D. Gerardo, and N. V. Dionisio, “Plant leaf detection and disease recognition using deep learning,” in 2019 IEEE Eurasia
Conference on IOT, Communication and Engineering (ECICE), Yunlin,
Taiwan, 2019, pp. 579–582, doi: 10.1109/ECICE47484.2019.8942686.

[5] S. Mathulaprangsan, K. Lanthong, D. Jetpipattanapong, S. Sateanpattanakul, and S. Patarapuwadol, “Rice diseases recognition using effective
deep learning models,” in 2020 Joint International Conference on Digital Arts, Media and Technology with ECTI Northern Section Conference
on Electrical, Electronics, Computer and Telecommunications Engineering (ECTI DAMT & NCON), Pattaya, Thailand, 2020, pp. 386–389, doi:
10.1109/ECTIDAMTNCON48261.2020.9090709.

[6] P. B. Padol and A. A. Yadav, “SVM classifier based grape leaf disease detection,” in 2016 Conference on Advances in Signal Processing
(CASP), Pune, 2016, pp. 175–179, doi: 10.1109/CASP.2016.7746160.

[7] S. Ramesh, R. Hebbar, M. Niveditha, R. Pooja, B. N. Prasad, N. Shashank, and P. Vinod, “Plant disease detection using machine learning,” in
2018 International Conference on Design Innovations for 3Cs Compute Communicate Control (ICDI3C), Bangalore, 2018, pp. 41–45, doi:
10.1109/ICDI3C.2018.00017.
END OF MY
PRESENTATION
THANK YOU !