| 摘要 |
PROBLEM TO BE SOLVED: To appropriately extract the video of an intersection which is beneficial to traffic safety guidance by accurately determining the position of a traveling object at a photographic point of time of each frame of a video photographed by a photographing device loaded on the traveling object regardless of failure in the positioning of a GPS.SOLUTION: A positioning acquisition part 12 acquires the position of a vehicle at a photographic point of time of each frame of a video on the basis of sensor data 32 including a GPS position recorded by a GPS device, and a behavior determination part 14 determines the behavior of the vehicle at each frame photographic point of time on the basis of sensor data 32, and an intersection selection part 16 selects an intersection from the position of the vehicle at each frame photographic point of time, and an intersection crossing determination part 18 calculates a vehicle existence area at each frame photographic point of time, and narrows down intersections to be determined from the behavior of the vehicle at each frame photographic point of time and attribute information of each of the intersections, and calculates the area of each of the narrowed-down intersections, and determines each frame photographed in a period in which the vehicle existence area intersects the area of any intersection. |
