Abstract: With the advantage of multi-scale and multi-resolving power of wavelet transform, the leakage features of crude petroleum pipeline were extracted by wavelet analysis. Using the regression method, the real-time measured data of pipeline leakage can be analyzed and recognized. The location can be obtained by the extracted hybrid features. The experiments show the validity of the proposed method. 

Key words: Wavelet analysis, Crude petroleum pipeline, Leakage detection 

摘要： 利用小波变换在突变信号处理中的优势，提取出原油管道泄漏信息经过小波变换后的泄漏点特征，运用回归分析方法，对泄漏信号的实测数据进行分析和定量识别。设计了反映泄漏信息的混合波形特征，并基于该特征实现了原油管道泄漏检测系统设计，由此定位管道泄漏位置，通过实验验证了方法的可行性。

关键词: 小波分析 , 原油管道 , 漏检测