For Rwanda's small-scale farmers, the maize crop has grown significantly in importance as a crop for food security and income. Unfortunately, a number of illnesses continue to cause maize producers to have much lower yields. Diseases have a negative impact on the quality of maize harvests and impair agricultural production efficiency, which costs farmers a lot of money. Maize plant health conditions are essential for farmers to make a good profit; in addition, plant health conditions should be watched at all phases of plant growth to detect plant illnesses early and treat them.

Currently, Rwandan farmers rely on their naked eyes to diagnose maize illnesses. Because some plant illnesses are difficult to detect, observation needs extensive training and experience. Another method involves sending samples to a lab for testing; this takes money and time. To overcome the limitations of these techniques IoT and AI technologies are excellent must-haves for increasing farming productivity; they may assist farmers take precautions to prevent losses and provide decent food security in various regions of the nation.

In this thesis, an AI-enabled IoT mobile application for early maize plant disease detection was proposed to assist farmers in automatically detecting maize plant illnesses at an early stage of plant growth. To detect plant illness, an image of the plant is captured with the camera and uploaded to a local server using an Android application, the plant image is subjected to various image processing algorithms at the server to determine the disease, and the discovered disease is delivered back to the farmer's mobile application with remedies. Our system was assessed using a number of performance measures, including classification accuracy and processing speed. When it comes to differentiating the three most common disease groups that damage maize leaves, the model has an overall classification accuracy of 80%.