Aircraft Continuous Tracking Method Based on Combination of Video Analysis and Positioning Information

Abstract

The invention relates to an aircraft continuous tracking method based on the combination of video analysis and positioning information, which calculates and obtains the image coordinates of the target aircraft in the video image according to the positioning information of the target aircraft, performs single-target tracking on the target aircraft according to the surveillance video, and combines the two For comparison, if the distance between the two continues to exceed the effective tracking threshold, it is considered that the single target tracking has lost the target. After confirming that the single target tracking has lost the target, take the image coordinate point of the target aircraft in the video image as the center point, and based on the image coordinate point The effective tracking range determines the search frame for re-identification of the target aircraft. The HOG feature is preferably used as the feature descriptor for target aircraft detection, and the position model and scale model of the tracking filter are obtained by calculating the DSST algorithm. The invention can judge the loss of the target aircraft in time, and the amount of data processing for retrieving the lost target is small, and can be mainly used for continuous tracking of targets such as aircraft and similar occasions.

Description

technical field [0001] The invention relates to an aircraft continuous tracking method based on the combination of video analysis and positioning information, which belongs to video monitoring and computer information technology, and can be mainly used for continuous tracking of aircraft and other targets and similar occasions. Background technique [0002] There are two main types of existing aircraft continuous tracking technologies: one is based on the existing aircraft positioning information, and maps the latitude and longitude of the aircraft used in the positioning information to the image coordinates of the aircraft monitoring or tracking system, thereby realizing aircraft identification and tracking. The longitude and latitude information (positioning information) can be obtained by various methods, such as surface surveillance radar (SSR), automatic dependent surveillance-broadcast (ADS-B), multi-base station positioning system (Multilateration), etc.; another type ...

AI Technical Summary

Problems solved by technology

These technical difficulties will cause the selected tracking target to be lost, that is, the tracking frame is separated from the actual target in the video, which will affect the actual promotion and utilization

[0004] At present, the first problem that needs to be dealt with in the face of the loss problem is: how to judge the loss of target tracking and the time of loss. However, among all the existing algorithms, none of them can really judge when the target is lost.

The existing technology usually regards the disappearance of the tracking frame in the video as the loss of the target. In this case, target re-identification can be used to retrieve the lost target. For example, the characteristics of the tracking target can be recorded before the target disappears. Do global target detection, and then perform feature matching on the detected target. If the matching is successful, the target can be considered to be retrieved. Since the existing target re-identification algorithms are all based on the area selection strategy of sliding windows on the entire video image, they are not targeted. High time complexity and redundant windows; even target detection based on deep learning requires target detection on the entire video image, which is slow

In particular, most of the target loss in reality is that the tracking frame is still in the picture, but not at the actual position of the target, resulting in false tracking. This kind of target loss is more harmful than the disappearance of the tracking frame, and it is difficult to find and correct. , the existing technology is still unable to judge and correct the target loss in the presence of such a tracking frame

Images