Method for detecting and tracking moving object under complex background

Abstract

A method for detecting and tracking a moving object under a complex background belongs to the field of object tracking, and a method of stably tracking the object under the complex background is involved particularly. The method comprises the following steps: detecting the acquired current frame image to obtain detection object information satisfying preset conditions; at the current frame image,selecting a candidate region by taking the position of a tracking object as the center; and obtaining the object position corresponding to a candidate object in the candidate region by using a classifier model. The method of the invention sets the tracking loss judgment condition by the response diagram oscillation between a candidate sample and a classifier, so as to accurately judge whether theobject meets the conditions such as occlusion, loss or blur or not. At the same time, the historical value of a response value in a response diagram is used to judge the model update, which reduces the model drift and the number of model update. In a complex scene, the tracking object can be accurately selected from the candidate sample, and the fast and stable tracking of the object can be achieved.

Description

technical field [0001] The invention belongs to the field of target tracking, in particular to a method for stably tracking a target under complex backgrounds. Background technique [0002] Whether in the military or civilian fields, target tracking technology has a wide range of applications. Fully automatic or semi-automatic target tracking tasks in battlefield reconnaissance, low-altitude defense, traffic monitoring, and homeland security can greatly reduce staff and working hours. However, although many effective video target tracking algorithms have been proposed, they still face many difficulties in practical applications, such as lighting changes in the environment, nonlinear deformation of the target, camera shake, and noise interference in the background. , which poses a great challenge to target tracking. [0003] At the same time, most of the existing target tracking methods can only track the target in a short period of time, and there are few researches on the...

AI Technical Summary

Problems solved by technology

However, although many effective video target tracking algorithms have been proposed, they still face many difficulties in practical applications, such as lighting changes in the environment, nonlinear deformation of the target, camera shake, and noise interference in the background. , which poses a great challenge to target tracking

[0003] At the same time, most of the existing target tracking methods can only track the target in a short period of time, and there are few studies on the stable tracking method.

Images