Lens jitter prevention video movement target detection method for intelligent mobile terminal

Abstract

The invention relates to a lens jitter prevention video movement target detection method for an intelligent mobile terminal. The method comprises the steps: dividing a jitter video into a plurality of original frame images with independent numbers; setting a preset difference value as the interval; extracting the continuous original frame images as original images so as to guarantee that an original video does not have big pixel loss; carrying out the gray processing and two-dimensional Gaussian filtering of the original images through optimized image space conversion parameters, and then classifying the obtained images through building a background pixel model and a spherical model, and obtaining a background point; carrying out the movement compensation and correction of the current frame image through employing the offset of the image background pixels after the initialization and real-time updating of a background model, correcting the background model through employing the offset, and extracting a moving object. The method carries out the movement compensation and correction of the current frame image through employing the offset of the image background pixels, so as to compensate the jitter of a lens and to eliminate the adverse impact caused by an unstable jitter lens, thereby effectively detecting the jitter video moving object.

Description

Technical field [0001] The invention relates to the field of intelligent video processing, and in particular to a method for detecting a moving target in an intelligent mobile terminal anti-camera shake video. Background technique [0002] In recent years, various cameras have been installed in different public places, such as highways, street corners, large shopping malls, hospitals, schools, and airports, to achieve the purpose of video surveillance. Traditional video surveillance relies on artificially analyzing and judging these video data subjectively, understanding the action behavior, and then drawing conclusions and giving feedback. Traditional video surveillance also has many shortcomings, especially when the camera lens shakes, once the surveillance personnel cannot make a judgment on the moving target in the shaking video, once important information is omitted, it may cause irreversible consequences. At the same time, with the continuous emergence of smart mobile term...

AI Technical Summary

Problems solved by technology

[0004] However, in the case of a fixed camera, the existing methods default to the acquisition of video data on the premise that the camera is fixed. Once a video is shot by a smart mobile terminal, such as a camera or a mobile phone, it is prone to lens shake. The video moving target detection method for non-jittering shots will be useless, causing the shaking shots to interfere with the extraction of moving targets, resulting in large-scale false detections

Images