Electronic image stabilization method based on characteristic coupling

Abstract

The invention provides an electronic image stabilization method based on characteristic coupling. The method comprises the following steps: (1) motion estimation: estimating motion parameters of a present frame and a previous frame to obtain a global motion vector and a local motion vector; (2) motion filtering: employing a Kalman filter to carry out filtering on the global motion vector, and obtaining a willful scan vector, wherein difference between the global motion vector and the willful scan vector is a vibration parameter needs to compensate; (3) motion compensation: utilizing a compensation parameter obtained by filtering carrying out frame by frame compensation; (4) video restoration: by utilizing space-time redundant information of video, carrying out restoration on the local motion vector with a motion restoration method and guiding pixel point filling. In order to solve motion carrier vibration influence on video shooting effect, the invention provides the electronic image stabilization method based on characteristic coupling. The method has the advantages of high accuracy, good anti-noise performance, strong tracking performance, good boundary consistency and the like.

Description

technical field [0001] The invention belongs to the technical field of image processing, relates to video image processing, and is an electronic image stabilization method based on feature matching. Background technique [0002] In recent years, with the development of digital multimedia technology and the continuous improvement of people's living standards, more and more video-based applications in safety, entertainment, and environmental protection have appeared in daily life. As far as environmental protection is concerned, the cyanobacteria outbreak in Taihu Lake in the summer of 2007 led to water cutoffs in Wuxi, which aroused great attention from the State Council. At present, Taihu Lake is listed as the first of the eight key areas in the National Science and Technology Major Special Project on Water Pollution Control and Treatment. During the implementation of the national major special project of "Taihu Lake Spirulina Video Surveillance", it was found that the actua...

AI Technical Summary

Problems solved by technology

[0007] CN201010179280.9 uses block matching in motion estimation. Although the search algorithm is the diamond search method, it is still difficult to achieve a good balance between accuracy and complexity, especially in the real-time video image stabilization occasions targeted by the present invention , it is difficult to apply in real time

In addition, the abnormal block distinguishing strategy is to compare the difference between the current movement mean and the last mean value, which is not suitable for large-scale movements, and the adaptability is not strong

[0008] CN201010179282.8 adopts the particle filter PF (Particle Filter) method to filter the motion parameters, but this algorithm needs a large number of samples to well approach the posterior probability density of the system, and cannot effectively distinguish between "intentional motion" and "random jitter". "

In addition, the hole filling after image stabilization adopts the method of mosaic of multiple images, the stitching traces are obvious, and the texture continuity at the stitching place cannot be maintained

[0009] CN200710036817.4 adopts SIFT algorithm to obtain global motion parameters. Although the matching accuracy is high, the algorithm has high time and space complexity and poor real-time performance

In addition, the method of combining Gaussian filtering and curve fitting is used in motion filtering. Although "over-stable" and "under-stable" can be avoided through parameter adjustment, the problems of delay and smearing have not been solved.

[0010] The above-mentioned pixel-based electronic image stabilization algorithms, such as grayscale projection method and bit-plane matching method, use the relationship between the grayscale values ​​of each pixel to estimate the motion parameters. The biggest feature is that the processing speed is fast and it is easier to realize real-time processing. However, This type of algorithm has certain limitations. First, it requires the image to have rich texture and obvious contrast. Secondly, it assumes that the rotation angle between two adjacent frames is small, and only considers translational motion, not rotational motion.

The post-stabilization video restoration algorithm based on motion restoration first repairs the local motion field, and then guides the pixel restoration through the complete motion field. However, some points to be repaired near the boundary may be "skipped" and cannot be used for optical flow vector estimation, thus The formation of "blind spots" that affect the restoration of pixels

Images