Shake compensation method and device for image acquisition device

Abstract

Embodiments of the present application provide a shake compensation method and device for an image acquisition device, which are used for real-time anti-shake of a video, and relate to the field of image processing. In this method, feature point detection is performed on the first frame of pictures, and a specified number of feature points are selected as feature points to be tracked in the second frame of pictures. For each current frame picture except the first frame picture, the feature points determined to belong to the moving target object are eliminated from the feature points to be tracked in the current frame picture, and the first motion parameter estimation is performed using the remaining feature points. After path planning or filtering is performed according to the estimation result of the first motion parameter, shake compensation is performed on the picture of the current frame. And selecting at least one feature point from the current frame picture; the selected at least one feature point and the remaining feature points are used as feature points to be tracked in the next frame picture. In the above method, since the feature points determined to belong to the moving target object are eliminated, the amount of calculation is reduced, and the accuracy of motion estimation is increased.

Description

technical field [0001] The present application relates to the field of image processing, and in particular to a shake compensation method and device for an image acquisition device. Background technique [0002] Camera equipment will inevitably cause video shake when shooting, which will affect the visual experience, so it is necessary to eliminate video shake. The existing anti-shake methods usually include optical anti-shake or electronic anti-shake. Optical anti-shake is to achieve anti-shake through the movement of the lens or photosensitive components. Electronic image stabilization includes two types: sensor and pure algorithm. Optical image stabilization and sensor-based electronic image stabilization are more costly and less flexible than pure algorithms. [0003] The existing pure-algorithm electronic anti-shake technology is usually an anti-shake method based on global motion parameters, which has a large amount of calculation and slow processing speed and is di...

AI Technical Summary

Problems solved by technology

Optical image stabilization and sensor-based electronic image stabilization are more costly and less flexible than pure algorithms

[0003] The existing pure algorithm electronic anti-shake technology is usually an anti-shake method based on global motion parameters, which has a large amount of calculation and slow processing speed and is difficult to apply to real-time video anti-shake

Images