Due to the rise of sport cycling, mixed traffic of motorized vehicles and bicycles is common on Spanish two-lane rural roads. Road administrations face the challenge of managing roads integrating all road users. The first step towards a correct integration is to understand the phenomenon by characterizing cycle traffic and its interactions with other road users. Knowing how many cyclists, when and how they use the road is essential to ensure good road management. To this end, it is necessary to develop a counting system adapted to the specifics of road cycle traffic, which allows the volume of cyclists to be known in real time, as well as the distribution of cycle traffic over time. This study proposes a system for detecting and counting bicycles on rural roads based on video image processing using artificial intelligence. This system is equipped with a double connection, which allows sending the data to a traffic control center and controlling the system remotely. The system has been implemented and validated in a real environment, achieving acceptable results even when cyclists circulate in groups. Knowing the volume of bicycles in real time allows dynamic and intelligent management measures to be applied, which improve safety and traffic operation by integrating all users of the rural road.