Denoising effect evaluation indexes for EME measurement spectrum
Zili Yuan, Dan Long, Yafen Wang · 2016
Since some Gaussian noises may attach to the Electromagnetic Environment (EME) measurement spectrum, it's difficult to analyze test data or extract feature. There are some denoising methods, but we do not have any evaluation indexes to judge the effect of denoising methods. In view of this situation, a set of denoising effect evaluation indexes based on spectrum features is proposed in this paper. The indexes are including the Ratio of Gaussian Noise Square Sum (), the envelope compression ratio and the trend fitting degree. Using these indexes, we can put forward the denoising effect evaluation standard of the EME measurement spectrum. Moreover, we can use these indexes to optimize the parameters of denoising method. The effectiveness of these indexes are also showed using experiment validation in this paper.