基于PSO-SVM方法的电源线传导泄漏信号识别与还原

Recognition and Reconstruction of Conduction Leakage Signal via Power Line Based on PSO-SVM Method

针对显示器电源线传导泄漏信号中红信号识别的难题,该文提出基于粒子群(PSO)算法优化支持向量机(SVM)的识别方法。首先对传导泄漏信号进行滤波预处理并分段,然后利用粒子群-支持向量机(PSO-SVM)对传导泄漏信号进行训练、分类并与SVM分类性能进行对比,最后应用PSO-SVM实现了显示图像的还原。结果表明此算法可以准确实现电源线传导泄漏信号中红信号的识别,且识别率明显高于SVM分类器。

In order to identify the red signal in the conduction leakage signal of the display power line effectively, a Particle Swarm Optimization-Support Vector Mechine （PSO-SVM） algorithm based on Particle Swarm Optimization （PSO） algorithm for parameter optimization is proposed. Firstly, the conducted leakage signal is filtered, then the PSO-SVM is used to train and classify the conducted leakage signals and compared with the SVM classification. Finally, the display image is reconstructed using PSO-SVM. The result shows that the the red signal can be effectively identified, and the identification rate is significantly higher than the SVM classifier.