30-08-2014, 11:20 AM
Automatic Fungal Disease Detection based on
Wavelet Feature Extraction and PCA Analysis in
Commercial Crops
Automatic Fungal.pdf (Size: 265.64 KB / Downloads: 25)
Abstract
— This paper describes automatic detection and
classification of visual symptoms affected by fungal
disease. Algorithms are developed to acquire and process
color images of fungal disease affected on commercial
crops like chili, cotton and sugarcane. The developed
algorithms are used to preprocess, segment, extract and
reduce features from fungal affected parts of a crop. The
feature extraction is done with discrete wavelet transform
(DWT) and features are further reduced by using
Principal component analysis (PCA). Reduced features
are then used as inputs to classifiers and tests are
performed to classify image samples. We have used
statistical based Mahalanobis distance and Probabilistic
neural network (PNN) classifiers. The average
classification accuracies using Mahalanobis distance
classifier are 83.17% and using PNN classifier are
86.48%.
INTRODUCTION
India ranked within the world's five largest producers
of over 80% of agricultural produce items, including
many commercial crops. Agriculture is still the largest
economic sector and plays a major role in socioeconomic
development of India. Agriculture in India is the means
of livelihood of almost two thirds of the workforce in
India. India has over 210 million acres of farm land.
Jowar, wheat, sunflower, cereals are the major crops.
Apple, banana, sapota, grapes, oranges are the most
common fruits. Sugarcane, cotton, chili, groundnuts are
the major commercial crops.
I. PROPOSED METHDOLOGY
In the present work tasks like image acquisition,
preprocessing, segmentation, feature extraction, feature
reduction and classification are carried out. The detailed
block diagram of adopted methodology is shown in Fig.2.
The classification tree is given in Fig.3.
RESULTS AND DISCUSSIONS
All the algorithms used in this work are implemented
using MATLAB 7.0. The image samples are divided into
two halves and one half is used for training and other is
for testing. Around 15% image samples are used for
validation of the designed classifier model.
The percentage accuracy is defined as the ratio of
correctly recognized image samples to the total number of
test image samples. The Percentage accuracy is given by
equation. (2).
CONCLUSIONS
The technology leverage farmers can take up to asses
the crop, look at the possibility of diseases at early stages,
take decision on possible treatment, and the like. The
identification of the symptoms of fungal diseases, by
means of a machine vision system may support farmers
in proper assessment of crops. Here we used image
samples of commercial crops that showed visual
symptoms of a fungal disease. Features were extracted
from affected region and used as inputs to a Mahalanobis
distance and PNN classifiers. The classification accuracy
and computational speed up for training taken by
classifiers shows that PNN classifier works better for our
approach.
The work carried out has relevance to the real world
classification of commercial crop disease and it involves
both image processing and pattern recognition techniques.
For future study, different neural network architectures,
Support Vector machine (SVM), Statistical methods can
be used for classification. We can extend this project to
classify disease affected on fruits, vegetables, cereals etc.