Disease Detection of Apple Leaf Using Machine Learning Techniques

Abstract

Application of farming in sample and
identification of diseased leaves is another area
where farming can bring a revolution in terms
of disease control through the use of machine
learning. Building, in essence, works with
loopholes with the aid of state-of-the-art
computation to learn plant diseases through the
depiction of pictures. This process integrates
multiple stages, and one of the first is data
acquisition: the necessary set of images of leaves
of different access was collected, and the method
of support SVM’s would be a large space of
diseased and non-diseased samples in
comparison to the other classifiers. This implies
that before getting into the AI model, the data is
subjected to several processing steps that yield
features suitable for the model. increase the
extent of image detail of current objects of
interest and acquire new sources for the current
corpus. There are basically artificial neural
network approaches here, and among these,
convolutional neural networks (CNNs) are used
most in the current time. This comes from the
capacity to regulate and decide on the hierarchy
attributes of the distinct pictures and patterns.
Model selection reaches even the activity of
comparing features of some algorithms to make
the correct choice among the available ones,
using criteria of efficacy, for instance, including
accuracy and precision. In other words, there is
a significantly large phase of checking the
accuracy of the model in another dataset
entirely different from the input data set, whose
aim is to determine the generality and overall
robustness of the model in any other data set.