Plant Leaf Disease Detection Using Machine Learning

Abstract

Plants are now a significant source of energy and are essential to the solution to the global warming conundrum. Plant ailments, however, are endangering this vital source's survival. Convolutional neural networks (CNN) have proven to perform admirably (outperforming humans) in tasks involving object recognition and image classification. This study examines the viability of using CNN to identify plant diseases in leaf photos that were captured in their natural habitat.

Early maladies better crop quality and output, Spotting & pets are essential. Due to the decreased quality of their agricultural goods, unhealthy plants cause farmers to experience huge financial losses. In a country like India where the bulk of the population works in agriculture, it is crucial to detect the disease early on. Plant disease predictions that are quicker and more accurate could lower the losses. Deep learning has made significant improvements that have given rise to chances to enhance the accuracy and effectiveness of object recognition and detection systems. The major objectives of this paper are to identify plant diseases and to minimize financial losses. Deep learning has been proposed as a method for image recognition. We identified the three major forms of neural network architecture: faster region-based convolution neural network (Faster R-CNN), region-based fully CNN (R-CNN) (SSD), and single shot multibook detector. The research's suggested technique is capable of handling complex scenarios and is efficient at identifying different sickness types. An accuracy of 94.6% in the validation findings shows the convolutional neural network's potential and hints in the direction of a deep learning approach based on artificial intelligence for solving complicated problems.