A Deep Learning Based Smart Agriculture Technique For An Iot Environment

Abstract

The application of deep learning strategies to the emerging field of smart agriculture is undergoing rapid development and is gaining increasing levels of interest. The application of deep learning algorithms to the topic of smart agriculture is something that is being explored, however it is quite early in its development. Utilizing Deep Learning in conjunction with Smart Agriculture in order to make the Internet of Things more accessible Deep learning may be able to assist in taking into consideration a variety of characteristics when developing a strategy for harvesting. These characteristics include the type, quality, and pH of the soil; the weather prediction (including temperature, precipitation, humidity, and hours of sunlight); and the schedule for applying fertilizers. For the objectives of this study, datasets of plant leaves, originating from both healthy and ill plants, were gathered for the training, validation, and testing of the CNN model. These datasets were acquired for the purposes of this study. When applied in agriculture for the purpose of identifying and classifying image of plants, the CNN model obtained the greatest attainable degree of accuracy.