摘要
目前视频在教育、健康、工业等领域有着广泛的应用。这一领域的重要研究表明,视频和图像的处理仍然是一个挑战,进一步的深入研究是相当复杂的。在过去的几十年里,在这一领域的研究中,人们提出了不同的方法和应用。最常用的一种方法是通过关键帧提取技术来对视频数据进行压缩。提出了一种简单、快速、高效、复杂度低的方法,该方法利用直方图差异来实现视频关键帧。算法在读取视频、确定该视频的帧数以及通过计算两个连续帧绝对差的均值和标准差来提高这些帧之间的对比度之后,引入一种均衡化技术来计算每一帧的直方图,这样就可以计算出视频帧之间的直方图差值。之后再进行适当的阈值比较,如果连续两帧之间的直方图差值大于之前计算的阈值,那么当前帧就被选作关键帧。结果表明,该算法不仅复杂度低,而且简单有效。
Currently videos have a wide range of applications, such as education, health, industry, etc. Important studies and research in this field have shown that the processing of video and image remains a challenge and continues under a deep study and research with a high degree of complexity. In the last decades, different methodologies and applications for the study and research of this field have been proposed. One of the most used methods is the extraction of the key frames to compress the video data. In the article we propose a simple method, fast, efficient and with a low degree of complexity, Histogram Difference for the selection of the key frames of a video. First, the algorithm starts with the reading of a video. Second step, the total number of frames in this video is determined. Third, calculate the value of the mean and standard deviation of the absolute difference between two consecutive frames, to improve the contrast levels of these frames, a methodology called equalization is introduced and the histogram values of each frame are calculated and successively the histogram difference frame to frame is calculated. Then an appropriate threshold value is determined and a comparison is performed, if the histogram difference between two consecutive frames is greater than the previously calculated threshold and the current frame is selected as key frame. The results show that the algorithm not only presents a low level of complexity, it is also simple and efficient.
作者
Jorge Michel Díaz Rodriguez
姚品
万旺根
Jorge Michel Diaz Rodriguez;Yao Pin;Wan Wanggen(School of Communication and Information Engineering,Shanghai University,Shanghai 200444,China;Institute of Smart City,Shanghai University,Shanghai 200444,China)
出处
《电子测量技术》
2019年第19期170-175,共6页
Electronic Measurement Technology
基金
上海市科委港澳台科技合作项目(18510760300)资助
关键词
关键帧
直方图
阈值
均衡化
绝对差
key frame
histogram
threshold
equalization
absolute difference
