Artificial intelligence based spraying drone design and implementation

The world population is increasing rapidly and accordingly there is an increase in food demand. For this reason, it is necessary to increase crop productivity in agricultural lands in order to meet the food demand. Substances such as insects and pests that will adversely affect plants in agricultural lands are called "pesticides". Chemical spraying is needed to clean the plants from pesticides. Studies show that chemical spraying in agricultural products increases product yield by 60%. The biggest disadvantage of manual spraying is that it can cause respiratory ailments, heart diseases, etc. to the person spraying these fertilizers. that can cause health problems. For this reason, it should be carried out in a balanced way and without the use of human power as much as possible. To avoid this risk and to spray pesticides evenly, spraying of agricultural lands can be carried out by a drone. Drones can perform efficient spraying in a short time compared to manual methods. With this study, cherry trees are sprayed autonomously with the support of artificial intelligence. NVIDIA Jetson NANO development kit was used to run the artificial intelligence application. The YOLOv5 model was preferred because of its high accuracy for the detection of cherry trees. The drone Hexacopter was modeled using the SOLIDWORKS program, with a body structure and six engines, and its analyzes were carried out with the ANSYS program. Drone A spraying drone with a body length of 1150mm, a drug capacity of 5l and a flight time of 12 minutes has been developed. Thanks to the developed hexacopter drone, spraying of agricultural lands is carried out autonomously. Due to the fact that they fly high, a homogeneous spraying is carried out and the yield of the products is increased. At the same time, it will not only reduce the required manpower and spraying time, but also reduce the health hazard. According to the results of this study, the F1-score value of the deep learning model was determined as 0,980. When the autonomous (continuous) spraying and artificial intelligence-based spraying method are compared for the spraying method, 53% less medication is used and 50% less energy is used in the artificial intelligence-based spraying method. Thanks to the advantages it offers, artificial intelligence support can be added to spraying drone systems for spraying cherry trees.