Target image stabilizing method of binocular vision system with variable visual angle and resolution ratio

Abstract

The invention relates to a stabilizing method of high resolution video based on a binocular PTZ vision system, which belongs to the field of visual surveillance. The method is characterized in that the two videos with different resolution ratios are simultaneously adopted, wherein, one video adopts a fixed panorama image of a low-resolution visual angle, and the other adopts a local image of high-resolution changed visual angles, comprising interested moving targets. The method adopts two PTZ cameras, therefore the application is convenient. The high resolution video can be obtained by manually or automatically controlling the moving of the cameras through manpower or a computer. The target is to obtain a complete and stable video including interested moving targets, and high-resolution information is adopted as much as possible. The method firstly solves the problem of the image registration between videos with different resolutions, secondly provides four sequential filling modes, and fully utilizes prior high-solution information and historical high-solution information to fill target videos. The result processed in by the method can be used for crime forensics, monitoring record storage, behavioral analysis of moving targets and the like, and experimental results indicate that the method provided by the invention is practical.

Description

technical field

[0001] The invention relates to the technical field of computer vision video processing, and in particular to a technology for stabilizing high-resolution video using a binocular PTZ (an active lens that can change viewing angle and resolution parameters) lens. Background technique

[0002] With the development of computer vision technology and the improvement of security awareness around the world, intelligent visual surveillance technology has received more and more attention. This technology has gradually been widely used in military and civilian applications. In the early application, the intelligence level was very low, so the security protection function mainly relied on people to judge the events in the surveillance video, and the reliability and automation were very low. Its main purpose is to collect evidence after the crime. The development trend of the current intelligent visual surveillance system is to pay more attention to crime early warning ...

Examples

Embodiment Construction

[0134] The core idea of ​​the present invention is to first use the image registration information at different resolutions to determine the visible field of view of the output video, and then use the high-resolution information of the current frame and the historical high-resolution information to fill the field of view. The principle is to use as high resolution information as possible. Finally, for the areas that cannot be filled, the low-resolution images enlarged by interpolation are used to ensure the integrity of the video.

[0135] Mainly comprise following three technologies in the present invention:

[0136] First, the registration of images of different resolutions. The present invention combines two methods based on feature points and direct pixel gray scale, and designs a "three-step registration method", which can effectively solve this problem.

[0137] Second, the strategy of filling high-resolution information. The present invention first uses the high-reso...

technical field

[0001] The invention relates to the technical field of computer vision video processing, and in particular to a technology for stabilizing high-resolution video using a binocular PTZ (an active lens that can change viewing angle and resolution parameters) lens. Background technique

[0002] With the development of computer vision technology and the improvement of security awareness around the world, intelligent visual surveillance technology has received more and more attention. This technology has gradually been widely used in military and civilian applications. In the early application, the intelligence level was very low, so the security protection function mainly relied on people to judge the events in the surveillance video, and the reliability and automation were very low. Its main purpose is to collect evidence after the crime. The development trend of the current intelligent visual surveillance system is to pay more attention to crime early warning ...

Examples

Embodiment Construction

[0134] The core idea of ​​the present invention is to first use the image registration information at different resolutions to determine the visible field of view of the output video, and then use the high-resolution information of the current frame and the historical high-resolution information to fill the field of view. The principle is to use as high resolution information as possible. Finally, for the areas that cannot be filled, the low-resolution images enlarged by interpolation are used to ensure the integrity of the video.

[0135] Mainly comprise following three technologies in the present invention:

[0136] First, the registration of images of different resolutions. The present invention combines two methods based on feature points and direct pixel gray scale, and designs a "three-step registration method", which can effectively solve this problem.

[0137] Second, the strategy of filling high-resolution information. The present invention first uses the high-reso...