EEMD 近似熵和支持向量机的管道泄漏特征向量提取研究

摘要

针对管道泄漏检测中常存在误报、错报及漏报等问题，提出一种管道泄漏检测高准确率的检测方案。该方案将EEMD近似熵和支持向量机相结合应用到管道泄漏检测研究中，为消除端点效应和模态混叠现象，将改进经验模态分解和近似熵相结合对管道泄漏信号进行泄漏特征提取，组成特征向量组作为支持向量机分类的输入，实现较好的分类效果。%Considering the false declaration or report in the pipeline leak detection, a scheme of improving leakage detection accuracy was proposed.This scheme has the EEMD approximate entropy combined with SVM and then has it applied to the pipeline leak detection.With a view to eliminating the endpoint effect and modal overlap, the improved empirical mode decomposition and combination of approximate entropy were a-dopted to extract the pipeline leakage’ s signal feature and to input it as SVM to achieve better classification effect.