Artificial Intelligence Enabled Internet of Things for Maize Plant Diseases Detection

EasyChair Preprint 10066

Abstract

Diseases affect the quality of corn crops and reduce the efficiency of agriculture production, resulting in a significant loss to the farmers. Currently, Rwandan farmers utilize naked-eye observation to identify maize diseases, which necessitates being well-trained and experienced, as some plant diseases are difficult to recognize. To overcome limitations presented by these techniques, the Internet of Things (IoT) and Artificial (AI) technologies are great imperative technologies for making farming more efficient. In this research, for early maize plant disease detection, an AI-enabled IoT mobile application is proposed to help farmers accurately detect maize plant diseases earlier during growth stages.  For detecting plant disease, a plant image is captured by the camera and transmitted to the local server running on raspberry pi with the help of an android application. Plant images are fed to the AI model which determines the types of disease and recommends further remedial steps to the farmer. Various performance indicators, such as classification accuracy and processing time, were used to evaluate our system. The model has an overall classification accuracy of 80%.

Keyphrases: Artificial Intelligence, Internet of Things, Raspberry Pi, mobile application

@booklet{EasyChair:10066,
  author    = {Rene Mitsindo and Emmanuel Masabo and Elias Bizuru and Joseph Habiyaremye and Viviane Ishimwe},
  title     = {Artificial Intelligence Enabled Internet of Things for Maize Plant Diseases Detection},
  howpublished = {EasyChair Preprint 10066},
  year      = {EasyChair, 2023}}