This ppt contained four articles review in one. based on onion leaf and bulb disease detection.
Reviewed by Abebe Tora Helana, MSc student @wolaita sodo university. (2024).
if you need articles, contact me on abebetora79@gmail.com
AI+A11Y 11MAY2024 HYDERBAD GAAD 2024 - HelloA11Y (11 May 2024)
onion bulb and leaf disease detection by using different methods.
1. WOLAITA SODO UNIVERSITY
SHOOL OF INFORMATICS
DEPARTMENT OF INFORMATION TECHNOLOGY
IT MSc Regular
Course:- IMS
Article review By:-
Abebe Tora Pgr/82835/15
Submitted To: - Dr. siraj
Sub. Date: - Jan 4/2014
2. 2
Title and authors
1. Detection of internal defects in onion bulbs by means of single-point
and scanning laser Doppler vibrometry. By Sandra Landahl, Leon A.
Terry
2. Image-based Onion Disease (Purple Blotch) Detection using Deep
Convolutional Neural Network. By Muhammad Ahmed Zaki1,
Sanam Narejo2* Muhammad Ahsan3, Sammer Zai4, Muhammad
Rizwan Anjum5, Naseer u Din6 .
3. Layout Classification of Red Onion Disease on Onion Leaf Image
Using Artificial Neural Network.by U N Mubarokhah, R Dijaya* and
M I Maulana
4. Plant leaf disease identification using image Processing and svm, ann
classifier methods.by Mrs. S.Sivasakthi, MCA, M.Phil.,
3. 3
Introduction
• Based on their article the authors introduced the onion, An onion also
known as the bulb onion or common onion, is a vegetable that is the
most widely cultivated class of the onion.
• Onion is an economically valuable crop and is the second-largest
vegetable crop in the world. The spread of various diseases highly
affected the production of the onion crop. E.g. purple blotch, botrytis
(bulb disease), stem phylium blight(stem disease) and downy mildew.
4. 4
Introduction
• the problem of rejected or downgraded onion lots due to internal
defects, which result in significant financial losses for wholesalers and
growers. The need for non-invasive assessment methods to identify
these defects and introduces Laser Doppler vibrometry (LDV) as a
potential solution.
• And method for identifying onion diseases through the classification of
leaf images using artificial neural networks and deep convolutional
neural network.
• In the next slide authors with their titles, methods, findings and also
limitations of articles are described.
5. 5
No. Authors titles Methods Findings limitations
1. Sandra
Landahl
Leon A.
Terry
Detection of
internal defects
in onion bulbs
by means of
single-point
and scanning
laser Doppler
vibrometry.
Laser Doppler
vibrometry
LDV can detect
internal defects
in onion bulbs
without
damaging them
Inability to
differentiate
defects like
sprouting
and double-
hearted
bulbs.
focus on
specific
onion
cultivars
2. Muhammad
Ahmed Zaki,
Sanam
Narejo2*
Image-based
Onion Disease
(Purple Blotch)
Detection using
Deep
Convolutional
Neural Network.
Deep
Convolutional
Neural Network
Pre-trained
enhanced
InceptionV3
model.
deep learning
method can detect
purple blotch
disease in onion
crops with an
accuracy of
85.47%
a limited
training
dataset
lack of
detailed
information
on the dataset
and validation
process
5
6. 6
6
3. U N
Mubarokha
h, R Dijaya*
Layout
Classification of
Red Onion
Disease on
Onion Leaf
Image Using
Artificial
Neural
Network
Artificial Neural
Network
Accurately
classifies red
onion diseases
into leaf rot and
purple spots
using an ANN
algorithm,
demonstrating its
effectiveness in
identifying leaf
images.
Generalizabili
ty to different
onion
varieties or
disease types.
4. Mrs.
S.Sivasakthi,
MCA,
M.Phil.,
Plant leaf
disease
identification
using image
Processing and
svm, ann
classifier
methods
Image Processing
and machine
learning
(SVM,(support
vector machine)
ANN Classifier
Methods)
utilizing color and
texture features
for accurate
detection.
It
acknowledges
manual disease
detection
7. 7
comparison
7
No. Method Finding Evaluation metrics
1. Laser Doppler
vibrometry
LDV can detect internal defects in
onion bulbs without damaging
them
Accuracy
Sensitivity
Specificity
Precision
F1 Score
2. Deep Convolutional
Neural Network
Pre-trained enhanced
InceptionV3 model.
deep learning method can detect
purple blotch disease in onion
crops with an accuracy of 85.47%
Classification
Accuracy
Detection Rate
Precision and
Recall
F1 Score
3. Artificial Neural
Network
Accurately classifies red onion
diseases into leaf rot and purple
spots using an ANN algorithm,
demonstrating its effectiveness in
identifying leaf images.
Accuracy
4. Image Processing and
SVM, (support vector
machine)
ANN Classifier
Methods
Utilizing color and texture features
for accurate detection.
Accuracy
Precision
Recall/Sensitivity
Specificity
F1 Score
8. Evaluation metrics discuss
• Accuracy: This metric measures the overall correctness of the
detection method in identifying internal defects in onion bulbs.
• Sensitivity: Sensitivity, also known as recall or true positive rate,
measures the ability of the detection method to correctly identify
true positive cases,
• Specificity: Specificity measures the ability of the detection method
to correctly identify true negative cases,
• Precision: Precision measures the proportion of correctly identified
positive cases (true positives) out of all the cases identified as
positive by the detection method.
• F1 Score: The F1 score is the harmonic mean of precision and
sensitivity.
9. Future work
1. Future research should enhance LDV (laser Doppler vibrometry)
technology's commercial sorting line applicability by developing
calibration models for onion cultivars, distinguishing developing
and double-hearted bulbs, and exploring cost-effectiveness and
feasibility.
2. Future work should expand the training dataset, incorporate
diverse onion diseases, and provide detailed information on the
dataset, validation process, and experimental setup for improved
model generalization and reproducibility.
10. Future work. Cont…
3. The paper suggests further development of a method for onion
farmers to classify diseases using leaf images, aiming to minimize
losses during produce and extend its application to other plant classes.
4. The authors suggest future work on feature extraction,
segmentation, machine learning algorithms, larger datasets, and deep
learning techniques for improved disease detection accuracy.
11. Final result
1. The 1st study demonstrated the effectiveness of LDV in detecting
internal defects in onion bulbs, albeit with limitations that require
further research and calibration for commercial use.
2. The 2nd paper presents a deep convolutional neural network
method for detecting purple blotch disease in onion crops,
showing promising accuracy but highlighting limitations in dataset
and detailed information.
3. The paper presents a method for classifying red onion diseases
using Artificial Neural Networks, showing promising results in
identifying leaf rot and purple spots, but requires further research
and improvement.
4. The proposed method for identifying plant leaf diseases using
image processing and SVM and ANN classifiers shows promise,
but further research is needed to improve performance.
12. Critics
Weak side
The main weak side of all articles there was no total number of data
they used for training and testing
the study did not address the cost-effectiveness of implementing
LDV on a commercial scale compared to other techniques, which
could be a potential drawback.
One potential criticism of the study is the lack of discussion on the
computational resources required for training and deploying the
proposed system.
Strong side
Based on their dataset their result is sufficient.