Video image stabilization motion estimation method based on reference frame optimization and foreground and background separation

Abstract

The invention discloses a video image stabilization motion estimation method based on reference frame optimization and foreground and background separation, which is used for solving the problem thatthe effect after image stabilization processing is influenced by inaccurate motion vector estimation. The method comprises steps of firstly, carrying out reference frame selection optimization by taking adjacent frame priority as a principle, and then carrying out clustering and preliminary screening on feature points in video frames in a grid clustering and density clustering algorithm fusion mode; solving an optimal homography matrix by using a random sampling consistency algorithm, and solving the distance between the feature point pairs by combining perspective projection transformation; and removing the feature point pairs which do not meet a preset condition according to a distance criterion to obtain purified feature point pairs; and finally, using the obtained accurate background feature point pairs for calculating a global motion vector. According to the video image stabilization motion estimation method disclosed by the invention, the influence of the foreground on motion estimation is effectively removed, the estimated global motion vector precision is improved, the image stabilization restoration effect is improved, and the method can be widely applied to image stabilization restoration of videos shot by different shooting devices.

Description

technical field [0001] The invention belongs to the technical field of video image processing, and relates to a motion estimation method for video image stabilization, which can be widely used in mobile shooting equipment. Background technique [0002] With the increasingly wide range of shooting equipment and the diversification of shooting platforms, various photography and camera equipment can be seen everywhere in life. The video collected by these devices will inevitably have some jitter, which is not conducive to manual observation and subsequent video image processing. The main purpose of video image stabilization is to correct and eliminate the jitter and blur caused by external factors or its own mobility in the video, so as to improve the video quality and viewing experience. Image stabilization technology can be used in scientific research surveying and mapping, law enforcement recorders, mobile phone cameras and other fields. Its advantages of high flexibility, ...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to solve the problems existing in the existing technical solutions, such as random selection of reference frames, inability to effectively remove the influence of foreground points, resulting in inaccurate global motion vectors calculated in the motion estimation link, etc., and proposes a method based on reference frame optimization and A motion estimation method for separating foreground and foreground feature points, the method comprising:

Images