Video target automatic locking and tracking method under dynamic background

Abstract

The invention discloses a video target automatic locking and tracking method under a dynamic background. The method comprises the steps: 1, a multi-view video collection module is used to shoot a real-time panoramic video image, and the real-time panoramic video image is sent to a target recognition module; 2, a target identification module identifies a target similar to the specified moving target in the panoramic video image, finds a to-be-tracked target sequence, and acquires an initial position coordinate sequence of the to-be-tracked target; 3, the target pre-selection module selects a to-be-tracked target according to a preset priority sequence and then sends the to-be-tracked target to the target tracking control module; 4, the target tracking control module drives the spherical camera to move to enable the target to be located in the center of an imaging picture of the spherical camera; 5, after the tracked target moves in the image of the spherical camera, the position information of the tracked target is recalculated, so that the tracked target returns to the imaging center position of the spherical camera again. According to the invention, the accuracy and reliability of target tracking in a large-scale scene monitoring video system can be greatly improved.

Description

technical field [0001] The invention relates to a video target tracking method in the technical field of video monitoring, in particular to a method for automatically locking and tracking a video target under a dynamic background. Background technique [0002] At present, video surveillance systems are more and more widely used in various fields. Due to the large and complex amount of video information, how to stably and reliably obtain the state information of a specific target (including the shape, position, speed, and trajectory of the target) from the video is increasingly becoming a research hotspot. [0003] The status information of the video target is generally obtained by manually observing the video, which requires very high energy and attention of the observer, which not only consumes manpower, but also easily leads to missed and misjudged judgments. [0004] In addition, due to the interference of many factors such as weather, light, visibility, and environment,...

AI Technical Summary

Problems solved by technology

[0003] The status information of the video target is generally obtained by manually observing the video, which requires very high energy and attention of the observer, which not only consumes manpower, but also easily leads to missed judgments and misjudgments

[0004] In addition, due to the interference of many factors such as weather, light, visibility, and environment, the scenes in video surveillance are complex and changeable, and multiple identical or similar targets often appear in the same scene

Therefore, for machine recognition, the existing target tracking algorithms are still not very stable and accurate in obtaining specific target state information, which affects the behavior analysis and decision-making judgment of specific targets.

Images