Pedestrian-vehicle cross-lens tracking method

Abstract

The invention relates to the technical field of image recognition, in particular to a pedestrian-vehicle cross-lens tracking method, which comprises the following steps of: extracting structural features of a target face video, a pedestrian video or a vehicle video; matching the structural features in an association database and determining the positions of people and vehicles, wherein human face structured information, pedestrian structured information and vehicle structured information which are mutually correlated are stored in the correlation database. According to the ivnention, pedestrian-vehicle positioning and cross-lens tracking are performed through an association database fusing structured information of human faces, pedestrians and vehicles, so that the problem in the prior art that efficient and accurate determination of human-vehicle positions and trajectory tracking cannot be met in massive monitoring videos is solved; and the technical problem that association detection of human faces, pedestrians and vehicles and cross-lens tracking of multiple cameras are difficult to realize is solved. Since the markers on the passengers and vechiles are unique, the conditions of driving escape, vehicle abandoning escape, seat replacement and the like can be continuously tracked from the mark unless the video range is separated from the video range.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a method for tracking people and vehicles across shots. Background technique [0002] At present, with the explosive growth of image surveillance videos, traditional video surveillance alone relies on human eyes to identify faces, pedestrians, and vehicles, which is far from meeting the demand. How to efficiently and accurately determine the location and track tracking of people and vehicles through massive surveillance videos? Face recognition, pedestrian recognition, vehicle recognition, and positioning of people and vehicles, as well as cross-camera tracking, have become urgent problems to be solved through videos captured by lenses under different viewing angles and different lighting conditions. [0003] Research has been carried out in the industry. For example, a Chinese patent discloses a method and device for cross-camera tracking of people and vehicles. The ap...

AI Technical Summary

Problems solved by technology

[0004] The disadvantage of the above technical solution is that people and vehicles are separated. In many safety incidents, there will be situations such as driving away, abandoning the car and fleeing, changing the car, etc. The above technical solution cannot make a reasonable response to this situation.

Images