Automatic object tracking method for dome cameras

Abstract

The invention discloses an automatic object tracking method for dome cameras. During a detection phase, moving objects are detected for a current monitoring scene; after the moving objects are detected, an object is determined as a tracking object, position information of the tracking object is obtained, and then the tracking object position information is sent to a dome camera; and the dome camera quickly turns to the tracking object to make the tracking object be in the central area of the monitoring screen. The above process is called as the detection phase. During the detection phase, thedome camera does not make a zooming action. During a tracking phase, the moving object detection is performed on the current monitoring scene; after the moving objects are detected, an object is determined as a tracking object, the position information of the tracking object is obtained, and then the tracking object position information is sent to the dome camera; and the dome camera rapidly turnsto the tracking object through control by a cloud terrace and makes the tracking object in the central area of the monitoring screen and automatically changes the zoom during the rotation according to the area of a moving area of the tracking object, so as to make the tracking object more clear in the central area of the monitoring screen.

Description

technical field [0001] The invention relates to the field of intelligent monitoring, in particular to a method for a dome camera to automatically track a target. Background technique [0002] Intelligent monitoring is to use computer vision technology to process, analyze and understand video signals. Without human operation, it can locate, zoom and track moving targets in the monitoring scene, as well as cruise monitoring, perimeter and mixing. Line, behavior analysis and other intelligent processing. At present, the automatic tracking methods on the market mainly include: [0003] Method 1: Realize target tracking based on gun-ball linkage. In the monitoring scene of the gun machine, the moving target is detected by algorithms such as image processing and pattern recognition, and the position information of the moving target is obtained, and then the position information is transformed through coordinates to control the rotation of the dome machine, and lock the target on...

AI Technical Summary

Problems solved by technology

The disadvantage of this method is that the monitoring scene of the gun is fixed, and the range that can be monitored is limited. The second is that there is coordinate conversion between the gun and the dome, the calibration process is complicated, and the installation accuracy is high. The third is that the equipment requires at least one gun. machine and a ball machine, resulting in increased cost

The disadvantage of this method is that the coordinate conversion between the panoramic camera and the dome camera is complicated, and the installation accuracy is high, and the second is that the cost increases

The disadvantages of this method are, first, for scenes with low contrast, the background interference is large, and there are few corners that actually fall on the target, especially for non-rigid targets such as the human body, the results of optical flow tracking corners Sometimes there is a big deviation

Second, it is easily affected by swinging leaves and changes in light

The third is that it is not advisable to change the magnification during the tracking process, otherwise it is more likely to fail to track, or there is a large deviation in the target position

This method mainly relies on the target's motion information rather than feature information, so the anti-interference is better. The disadvantage is that it is necessary to estimate the target's motion speed in real time, so it is best to do it only once when the target is detected for the first time. In the zoom action, it is not advisable to change the magnification during the subsequent tracking process, otherwise the calculation of the target position will easily cause a large deviation

Images