CAJ | 학술논문

在对高速运动目标的多媒体视频采集中，因为风力、高速抖动等原因导致视频和图像采集模糊，稳定性差，需要进行对视频和图像等多媒体素材进行电子稳像处理，提高对图像目标的识别和检测能力。传统方法中采用块匹配方法构建电子稳像算法，因为运动帧补偿效果不好，稳像效果较差。提出一种基于运动帧补偿的多媒体块匹配电子稳像算法，进行GSM随机场建模和进行像素点关联传递迭代优化，将图像的前一帧作为参考帧对当前帧进行参数补偿，求得的全部帧图像对应的运动参数，实现对电子稳像算法改进。仿真实验表明，算法能保留图像中的特征的同时去除图像中含有的抖动，稳像处理性能较传统方法优越。改进方法在电子侦察、远程目标识别等领域具有优越的应用前景。
재대고속운동목표적다매체시빈채집중，인위풍력、고속두동등원인도치시빈화도상채집모호，은정성차，수요진행대시빈화도상등다매체소재진행전자은상처리，제고대도상목표적식별화검측능력。전통방법중채용괴필배방법구건전자은상산법，인위운동정보상효과불호，은상효과교차。제출일충기우운동정보상적다매체괴필배전자은상산법，진행GSM수궤장건모화진행상소점관련전체질대우화，장도상적전일정작위삼고정대당전정진행삼수보상，구득적전부정도상대응적운동삼수，실현대전자은상산법개진。방진실험표명，산법능보류도상중적특정적동시거제도상중함유적두동，은상처이성능교전통방법우월。개진방법재전자정찰、원정목표식별등영역구유우월적응용전경。
In the multimedia video capture of high speed moving target, because the wind, high speed jitter causes such as video and image blur, it has poor stability, we need for electronic multimedia video and image stabilization processing, to improve the identification and the ability to detect the image target. In the traditional method using block matching method to construct the electronic image stabilization algorithm, because the motion compensation effect is not good, stabilization effect is poor. This paper proposed a matching electronic image stabilization algorithm based on motion compensation of multimedia, GSM random field modeling and pixel related transfer iterative optimization are obtained, the previous frame image as the reference frame of parameter compensation for the current frame is designed, corresponding to image motion, the parameters are obtained, the electronic image stabilization algorithm is optimized. Simulation results show that the algo-rithm can retain the characteristics of the image, the image contains the simultaneous removal of jitter, image stabilization processing performance is better than the conventional methods. The improved method has a good application prospect in the field of electronic reconnaissance, remote object recognition.