IoT and Deep learning-based approach for an efficient land suitability prediction in smart farms

Abstract

"The earliest known way for humans to make a living is through farming. Smart farming is a new, improved vision of agriculture that incorporates the use of new technologies. In recent times, farmers have increasingly depended on technology to efficiently carry out their daily responsibilities and enhance the quality of their crops. In agriculture, land suitability is an important aspect, which describes how well area is conducive for plant growth. Experts in the field of land suitability can determine it or use mathematical tools to make accurate predictions. Artificial techniques have been proven to be efficient prediction tools for this purpose. Empowered by the Internet of Things and Big Data, Artificial Intelligence (AI) is capable of handling these kinds of tasks and easing the burden on farmers and experts. The devices, used to improve farming, generate data in several formats, which might lead to ambiguous data. This paper proposes an ontology-based solution to deal with the heterogeneity problem. Moreover, this paper uses a deep learning-based solution that uses streamed weather data generated from sensors. In fact, our system uses the long-short-term memory model to predict land suitability. The model exhibited encouraging outcomes that could potentially influence the field of agriculture."