Electronic Image Stabilization Method Based on Visual Attention Mechanism

Abstract

An electronic image stabilization method based on a visual attention mechanism includes the following steps: conducting foreground moving area detection on a reference frame, marking foreground moving sub blocks, extracting overall situation remarkable characteristic points in the reference frame, matching characteristic point pairs, removing error-matched characteristic point pairs, acquiring moving parameters, conducting moving filtering, conducting rapid moving compensation, and rebuilding undefined boundary information to acquire a panoramic image. By means of the method, compensation parameters are acquired through extraction, matching, checking and moving parameter calculation of the overall situation remarkable characteristic point pairs and self-adaptation filtering smooth movement, vision stability and definition between video frames are improved, instability of video sequences is removed or reduced, and observation effects of a video monitoring or tracking system is improved.

Description

technical field [0001] The invention belongs to the technical field of digital image processing, in particular to an electronic image stabilization method based on a visual attention mechanism. Background technique [0002] Because the human visual system has the characteristic of persistence of vision, when the camera equipment shakes at a high frequency, the visual system is easy to capture the shaking of the picture in the image, so that the vision will be blurred and it is difficult to observe the target. The high-frequency jitter of the camera will cause the instability of the video sequence. The motion in the video sequence can be divided into target motion and camera motion, wherein the former is local motion and the latter is global motion. In order to facilitate observation, it is necessary to smooth and stabilize the motion in the video sequence, thus electronic image stabilization technology appears. Electronic image stabilization technology is a technology that ...

AI Technical Summary

Problems solved by technology

Among them, the attenuation coefficient in the motion attenuation method is set empirically in the experiment and cannot be applied to all video sequences; the mean filter uses a simple averaging operation, which will introduce redundant low-frequency noise; the Kalman filter requires process noise and observation The noise is known a priori and follows a Gaussian distribution with zero mean, which is difficult to satisfy in practical systems

[0008] For the motion estimation method and motion compensation method used in the aforementioned electronic image stabilization system, the speed of the algorithm depends on the extraction, matching and iterative operation of feature information, and there are local moving objects in complex motion scenes. Therefore, it is unavoidable that the feature points are selected on the foreground target, resulting in a decrease in the accuracy of global motion estimation; at the same time, it is difficult for the algorithm to deal with the complex random jitter and scanning motion of the camera at the same time, and it is prone to filter divergence and output false scenes, which are different from real The scanning scene varies greatly, which affects the observation effect; in addition, when compensating, the transformation parameters are calculated point by point for the current frame, which consumes computing time and affects the real-time processing capability of the system. The loss of boundary information will also affect visual observation.

Images