Estimation of Rice Field Area Using YOLO Method to Support Smart Agriculture System

The agricultural system is an important sector in the world economy that is responsible for providing food for the growing population around the world, especially in Indonesia, which is an agricultural country, which makes the agricultural sector an important sector in supporting food needs. According to BPS (Central Bureau of Statistics) data from the last three years, Indonesia has experienced fluctuations in terms of harvest area. This is caused by several factors, including the lack of maximum utilization of agricultural land, limited human resources, and challenges in the use of advanced technology. Therefore, an innovation is needed to help farmers maximize the area of agricultural land. So in this research, drones are used to conduct mapping from various heights to optimize the measurement of land area and the use of the YOLOv8 (You Only Look Once, Version 8) method to detect rice fields in order to calculate the area of agricultural land. The result is that the YOLOv8n deep learning model produces an accuracy rate of 0.765, which will detect rice fields planted with rice plants. And the measurement of the original land area results in image sample 1 show that the use of YOLOv8n only detects land around 613.61 m², with an error rate of 38%. Then, image sample 2 shows that the use of YOLOv8n only detects land around 387.75 m², with an error rate of 56%.