In this paper, we propose a lane-level dynamic traffic control (LDTC) system targeting at driving efficiency optimization. The LDTC system integrates vehicular networks with intersection cameras to collect fine-grained information of vehicles, such as turning intentions and lane positions. LDTC can maximize the intersection throughput and provide fairness among traffic flows. With vehicular networks, the traffic controller of an intersection can collect all turning information before vehicles make their turns. With intersection cameras, the lane positions of vehicles can be detected even if GPS is not accurate enough to provide lane localization. In addition, the traffic condition can be continually monitored as some vehicles are not equipped with onboard units for vehicular communications. In LDTC, while allocating the green light to the traffic flows with higher passing rates for throughput maximization, it also allocates the green light to the ones with lower passing rates for fairness provision. This paper demonstrates our current prototype.