A New Method for Passenger Flow Counting System Based on Surveillance Video

With the increasing of public safety demand, monitoring system has been used widely. However, there are not many mature passenger flow counting systems in the market, because the system´s key problem has not been satisfactory resolved. In this paper, the key algorithms in this system are discussed. The main algorithms have been improved, such as: the extraction of human, the segmentation of crowd and judgment of human moving direction. Finally, the improved algorithms integrate a system which achieves remarkable robust results through testing under different scenes.