Food shortages have intensified in the agricultural market due to the demand for food exceeding the supply. In particular, agriculture in Korea is suffering from a decrease in productivity due to a severely aging population and a lack of farm workforce. The agricultural population, which amounted to 14.42 million in 1970, has decreased by less than 10% to 1.03 million in 2021. The farm household population has also shrunk by 98,000 and is expected to continue declining.
In an attempt to address the limitations of modern agriculture facing the crisis of food and staffing shortages, various high-tech technologies have been introduced into agriculture. This shift is moving towards a new paradigm of robot and automation-based smart agriculture.
Agricultural robots play a crucial role in smart agriculture and can perform various tasks such as monitoring crops, increasing harvest yield, and removing weeds. In the coming years, agricultural robots are expected to become indispensable for farmers, as they can reduce working hours, labor force, and labor costs while ensuring farmers' interests. Agricultural robots are being developed alongside advancements in agriculture. However, efficient operation of agricultural robots requires software that can accurately recognize and track the surrounding environment and target objects.
Agricultural robots, along with object tracking software, can analyze imagery information captured by cameras or sensors to identify the location and movement of desired objects and control the operation of the agricultural robot accordingly. For example, harvesting robots require software that can identify and track objects such as fruits and vegetables to collect them at the appropriate time.
The development of object tracking software for agricultural robots is an important research topic that can contribute to the development of smart agriculture and productivity improvement. However, there have been few studies on object tracking algorithms and systems specifically designed for agricultural environments, and limited cases of actual application. Therefore, this study investigates object tracking algorithms and systems specialized for agricultural environments to enhance efficiency and accuracy in smart agricultural robots.