IoT-based Precision Agriculture System using Drone Vision for Sustainable Farming

Abstract

In this paper we proposed an IoT-integrated precision agriculture model that uses drones, wireless sensors, and artificial intelligence to make agriculture more efficient and sustainable. This model was designed with a five-layer structure that handles different stages, from collecting data to applying it in the real-world applications. Our proposed model was collected data from both aerial drones and ground sensors which was tested on the 5-hectares wheat agricultural land around 18 months. In this model we used a Convolutional Neural Network and it was trained on 35,000 of crop images out of 80% used for training and 20% for testing. The proposed model identified early signs of disease in the plants with 92% accuracy. This early detection of the disease helps the farmers to reduce the pesticide usage by 42%, to help effective resource management, reduced water usage by 30% and a 40% reduction in fertilizer usage. These savings were matched by a 20% increase in crop production and a 15% cut in the cost of running the farm. Looking at how the system performs financially over a long time, comparable for 10 years, it turned out to be a smart choice. It gave a return of 167% on the money invested and paid back the cost in around 7.3 years. Overall, the study shows that using drones and smart IoT tools can help make farming more sustainable, profitable, and ready for changing weather conditions.