Efficient video surveillance mobile object tracking way

Abstract

The invention provides an efficient video surveillance mobile object tracking way. The efficient mobile object tracking way comprises the following steps: S1, acquiring an effective connected region and an external rectangular region thereof detected by video mobile detection in advance; S2, deleting the region obtained in the step S1 to remove large area regions with width ratio and height rationof 80% and more, and removing the small area region with width ratio and height ration of 10% and less; and S3, removing the region of the connection region area occupying relatively small proportionof the whole external rectangular region area, namely, removing the region with proportion less than 10%. Through the tracking way provided by the invention, a mobile tracking algorithm is realized at a front-end chip without computing and transmitting data through a background server; a mobile tracking function can run all the time once being started in the equipment, the computation can be performed in real time according to different video scenes, and a shaking machine can be scheduled in real time to track a designated target, thereby achieving the ideal result in the end.

Description

[technical field] [0001] The invention relates to the technical field of moving object tracking methods, in particular to a high-efficiency video monitoring moving object tracking method with low computational complexity and low cost. [Background technique] [0002] In the field of security monitoring, video and image capture is an extremely important part. Most of the current home security monitoring equipment shaking heads and commercial high-speed balls and low-speed balls are controlled by preset trajectories and preset points. For cruising, very few devices have intelligent moving object tracking functions, and even if they do, they cannot efficiently achieve mobile tracking. The main reasons are in the following aspects: 1) It is difficult to implement pure algorithms in the front-end, and the calculation complexity is high; 2) It is difficult for the front-end chip to support such high-complexity calculations; 3) There are also images collected by the front-end and an...

AI Technical Summary

Problems solved by technology

[0002] In the field of security monitoring, video and image capture is an extremely important part. Most of the current home security monitoring equipment shaking heads and commercial high-speed balls and low-speed balls are controlled by preset trajectories and preset points. For cruising, very few devices have intelligent moving object tracking functions, and even if they do, they cannot efficiently achieve mobile tracking

The main reasons are in the following aspects: 1) It is difficult to implement pure algorithms in the front-end, and the calculation complexity is high; 2) It is difficult for the front-end chip to support such high-complexity calculations; 3) There are also images collected by the front-end and analyzed by the back-end server Yes, but it is inefficient to transfer data back and forth in this way, and the video surveillance equipment needs to constantly interact with the server, the server is under heavy pressure and the cost is high

[0003] Based on the above problems, how can we design a mobile tracking method, bypass the above points, and find another way to achieve real-time tracking of moving targets, which can better solve the difficulties of existing monitoring equipment in front-end processing and back-end processing servers The problem of high pressure and high cost is often considered by technicians in this field, and a lot of research and development and experiments have been carried out, and good results have been achieved

Images