基于重叠式元胞移位的音频广告统计方法

Audio Commercials Statistics Method Based on Overlapping Cellular Shift

根据广播广告监测的需求,提出一种基于重叠式元胞移位的音频广告检测算法.由于广告类型的多样性,该算法以多种类型的广告作为研究对象,并结合元胞自动机的基本理论和音频本身所具有的特点,采用重叠式元胞移位与互相关函数相结合的检测算法,实现音频广告的检测.实验结果表明:基于重叠式元胞移位的音频广告检测算法可以快速准确地从音频广播节目中检测出广告所在时间段,统计得到广告出现的次数,具有很好的鲁棒性.

To meet the demand of commercials-monitoring needed by the radio stations,an algorithm of audio commercials detection is proposed based on the overlapping cellular shift（ OCS）. Different types of commercials are used as research objects in the algorithm because of the diversity of commercials. Based on cellular automaton theory and the characteristics of audio,audio commercials are detected by using the method of the OCS combined the cross-correlation function. The experiment results show that the OCS-based algorithm can quickly and accurately detect the time section when commercials arise,and count the frequency of the commercials broadcasting from the radio,and the algorithm proposed is robust.