摘要
研究一种适用于运动图像序列的移位帧累积技术.通过在目标运动方向上对运动图像进行移位,满足像素间的空间相关性,进而实现运动图像序列的帧累积.阐述了移位帧累积的具体实现方法,推导了运动图像间移位量的数学模型,并讨论了移位帧累积实时处理的可能性.实验表明:通过数学模型所确定的移位量与实验结果具有良好的一致性;处理后的微光运动图像的噪声得到了抑制,像质较处理前有了明显提高.所讨论的移位帧累积方法突破了帧累积平均仅适用于静态图像增强的限制,是一种提高运动图像像质的有效方法.
Studies the technique of frame shift and integral that is suitable for the processing of moving image sequence. In frame shift and integral, spatial relativity is obtained by the shift of the previous and subsequent frames to a central frame. Noise suppression is achieved by the integration and average of frames. Thus, the quality of moving images is improved and the moving image sequence is enhanced. In this paper, the method of frame shift and integral is described. Moreover, a model to calculate the shift value is established and the probability of real time processing for the method is also discussed. Experimental results show that the calculated values of the shift are well in accord with the actual values. The noise of low light level image is suppressed and the quality of moving image is improved. The technique of frame shift and integral breaks through the limit that frame integral can only be applied to the enhancement of static image sequence. It is an effective method to improve the quality of moving image sequence.
出处
《北京理工大学学报》
EI
CAS
CSCD
北大核心
2003年第4期517-520,共4页
Transactions of Beijing Institute of Technology
基金
国家"八六三"计划项目(2001AA423290)
国家教委跨世纪优秀人才计划基金(2003)
