Denoising of X-ray pulsar observed profile using biorthogonal lifting wavelet transform

作     者：Mengfan Xue Xiaoping Li Yanming Liu Haiyan Fang Haifeng Sun Lirong Shen 

作者机构：School of Aerospace Science and Technology Xidian University 

出 版 物：《Journal of Systems Engineering and Electronics》 (系统工程与电子技术（英文版）)

年 卷 期：2016年第27卷第3期

页      面：514-523页

核心收录：

学科分类：070207[理学-光学] 0808[工学-电气工程] 07[理学] 0809[工学-电子科学与技术（可授工学、理学学位）] 08[工学] 080203[工学-机械设计及理论] 0802[工学-机械工程] 0811[工学-控制科学与工程] 0702[理学-物理学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

主　　题：X ray pulsar denoising linear phase wavelet scale dependent threshold 

摘      要：In X-ray pulsar-based navigation, strong X-ray background noise leads to a low signal-to-noise ratio（SNR） of the observed profile, which consequently makes it very difficult to obtain an accurate pulse phase that directly determines the navigation precision. This signifies the necessity of denoising of the observed profile. Considering that the ultimate goal of denoising is to enhance the pulse phase estimation, a profile denoising algorithm is proposed by fusing the biorthogonal lifting wavelet transform of the linear phase characteristic with the thresholding technique. The statistical properties of X-ray background noise after epoch folding are studied. Then a wavelet-scale dependent threshold is introduced to overcome correlations between wavelet coefficients. Moreover, a modified hyperbola shrinking function is presented to remove the impulsive oscillations of the observed profile. The results of numerical simulations and real data experiments indicate that the proposed method can effectively improve SNR of the observed profile and pulse phase estimation accuracy, especially in short observation durations. And it also outperforms the Donoho thresholding strategy normally used in combination with the orthogonal discrete wavelet transform.