Character tracking method, system and server

Abstract

The embodiment of the invention provides a figure tracking method, system and server. The method comprises the steps: each camera group in a region is connected with a figure database; a regional camera monitoring network is established, the locations of a to-be-detected person photographed by a camera groups in the database can be quickly queried through to-be-detected features of the to-be-detected person, and the locations where the camera groups of the to-be-detected person appear are connected in a time sequence, so that the movement track of the to-be-detected person can be quickly obtained. On the other hand, in the embodiment of the invention, a front image and a back image of a to-be-detected person are respectively acquired by using a camera group; any one of a cross-parallel ratio function target detection method, a face recognition embedding method for tracking face detection and a gait similarity matching method is used for matching the characteristics of the person to bedetected with the person in the monitoring image information, and any one of the three methods can realize the function.

Description

technical field [0001] The present invention relates to the technical field of line tracking, in particular to a person tracking method, system and server. Background technique [0002] As a kind of information recognition technology, image recognition technology can effectively identify characteristic information such as faces, clothing, vehicles, and license plates. Image recognition technology has been applied in many fields of life, such as person trajectory retrieval. However, the traditional character trajectory retrieval is still stuck in artificial video playback, which is time-consuming and labor-intensive, and the timeliness is not high. If it is only used for criminal investigation or people and objects, it can also meet the application requirements, but if it is for the matching and trajectory tracking of people's trajectories from different cameras in multiple areas, it cannot meet the application requirements. [0003] In a specific area, how to quickly and a...

AI Technical Summary

Problems solved by technology

But now the traditional character trajectory retrieval still stays in artificial video playback, which is time-consuming and labor-intensive, and the timeliness is not high

If it is only used for criminal investigation or searching for people and objects, it can also meet the application requirements, but if it is for the matching and trajectory tracking of people's trajectories from different cameras in multiple areas, it cannot meet the application requirements.

Images