DETECTION OF PLANT LEAVE DISEASE USING HYBRID TECHNIQUES

Plant illnesses are not good for you elements that reduce crop quality and quantity significantly. Biologists or farmers with a lot of experience frequently inspect plants for illness with their naked eyes, however this procedure is typically inaccurate and takes a long time. We apply artificial intelligence and computer vision techniques in this work to create and construct an intelligent classification system for leaf diseases. This study compares the simulation results of two approaches for performance evaluation. In the first section, photos data from Plant Village collection (for apple, maize, Plants for potatoes, tomatoes, and rice) are augmented with data, and deep features are retrieved using a neural network using convolutions (CNN). A Bayesian optimal machine classifier using support vectors is used to classify these characteristics, and the results are measured in terms of precision, sensitivity, f-score, and accuracy. The approaches described above will allow farmers all around the globe to take early action to avoid their crops from becoming irrevocably damaged, so sparing the world and themselves from a possible economic disaster. The second stage of the approach begins with the preparation of data set photos, after which texture and colour characteristics are retrieved using histograms of oriented gradients (HoG), GLCM, and colour moments. In this case, the three categories of features, namely colour, texture, and depth, are merged to generate hybrid features. To get simulation results, binary particle swarm optimization is a technique for reducing the number of particles in a used to pick these hybrid features, followed by classification usinga random forest classifier. Binary optimization of particle swarms is a technique for lowering the total number of particles in a technique for lowering the total number of particles in a critical in hybrid feature selection; the reason behind this Algorithm is to provide the best output with the fewest features. The above-mentioned assessment factors are used to conduct a comparative study of both methodologies.