The Application of Entropy Analysis Based on Wavelet Threshold Denoising in Radiation Source Feature Extraction
Feature extraction is an important step in the modulation recognition of communication signals. With advantages in denoising methods based on wavelet threshold denoising, the paper proposes a method of entropy analysis to extract the features of signals. After the treatment of wavelet threshold denoising it discoveries the Shannon entropy and Index entropy as the identification features and identifies the modulations of the signals. Characteristics extracted by traditional methods often need high signal-tonoise ratio, and are not stable sometimes. However, entropy characteristics after wavelet threshold denoising overcome it. Simulation results show that certain disparities exist between different signals entropy values. Signals could be well distinguished by setting thresholds after wavelet threshold denoising. Good recognition effects are found in different circumstances under two transformations. The conclusion is drawn that using the entropy analysis method to extract the signal characteristics is less affected by noise, and the entropy curves are relatively stable. Its more conducive to distinguish the signal modulation and has a very good application value.
feature extraction Shannon entropy Index entropy wavelet threshold denoising
Li Yi-bing Li Jing-chao Lin Yun
Harbin Engineering University Harbin, China
国际会议
桂林
英文
5-8
2010-11-17（万方平台首次上网日期，不代表论文的发表时间）