Video Location Positioning Study Based on Two Steps Greedy Algorithm

摘要

Determining the location and date of photos and videos is needed in some fields, such as production, public security, public opinions and information. But most photos and videos coming from networks or unclear anonymous sources have no clear information about the photographers and no signs of the location and date, which causes great difficulties in telling the true from the false and maintaining the public liability. It is also an obstacle of confirming and obtaining the evidence in cases. Thus, it is of great significance to clearly confirm the location and date of photos and videos rapidly. Base on the data of the changing lengths and angles on shadow objects in photos and videos, this paper use two-stage greedy algorithm to reduce the errors between the calculation place and the real place gradually for consequently to get the general latitude and longitude of the place. It also discusses the differences of knowing the specific date of the target and not. This method can be used to calculate the latitude and longitude with the shadow information provided by the photos or videos whose time spans are more than 30 minutes. And the error will be in 0.5 degree. The running time is saved above 70% related to the Grid Search.