An Intelligent Agriculture Application Based on Deep Learning

Mu-Yen Chen,Hsin-Te Wu,Wen-Yu Chiu

DOI: https://doi.org/10.1109/icsse.2018.8520209

2018-06-01

Abstract:Ever since the Internet of Things was developed, countries around the world have been applying the technology to solve current issues in agriculture, such as combating climate change and reducing labor costs. In the moment, intelligent agriculture has focused mainly on greenhouse cultivation; however, it takes significant amount of material and labor resources to build greenhouses, and their products are mostly high-valued cash crops. This study employs an outdoor farm as the field of experimentation and uses self-made model vehicles that patrols regularly to obtain information on the farm's crops. The study first establishes deep learning in intelligent agriculture to understand the farmer's farming skills so that such information may in turn help in further establishing an Internet of Things system that realizes automated farming systems. This study utilizes small, self-made model vehicles to gather crop information every day. The vehicles' designed functions include the following: (1) automated patrol along preprogrammed routes, (2) automatic photographing of each area, (3) prevention of collision between the small, self-made model vehicles, and (4) data collection of each area's temperature and humidity conditions. In this study, deep learning is employed mainly towards training and prediction of conditions based on the crop's temperature and humidity data as well as weather information. Deep learning is established first, followed by the Internet of Things, in order to achieve intelligent agriculture. The study must first acquire the farmers' cultivation skills; since such knowledge is passed on through experience and without any recordings, it can only be retrieved by means of the small model vehicles. In addition, the study also conducted simulation through hands-on experiments. The proposed scheme first collects all environment-related factors and conducts training on related coefficients, then utilizes the automatic irrigation system to allow the same kind of crops to grow in the most ideal environment. The system focuses on outdoor farms; it establishes an intelligent agriculture using minimal costs, resolving issues such as number of sensors needed, farm network accessibility, and electrical wiring.