In order to determine the people’s location and number in the disaster area, such as fire, earthquake and other emergency accidents region, and raise the efficiency of search and rescue,
we proposed a cloudy data-based saving regional real-time demographic statistical system, which is mainly constituted by four modules, namely, the target recognition, the local area
network server (LANS), the integrated management platform and the remote client. This system applied the strategy of vertical shooting and the technology of head recognition as well as the
target detection, and it also analyzed the headcount information accurately in the specific zones by LANS and management platform to share data with other similar module units in the cloud then generated a ground headcount statistics with huge data. The system combined the map’s locating display function and utilized PC or other mobile terminal easily to count the headcount in the specific area. It can not only provide scientific data support to the government rescues department to rescue the victims quickly and effectually from the disaster while decreasing the casualty maximally, but can function in other applications such as school, shopping center and enterprise, etc.. It would be widely used in the future.