Image noise is a variation of uneven pixel values that occurs randomly. A good estimation of image noise parameters is crucial in image noise modeling, image denoising, and image quality assessment. To the best of our knowledge, there is no single estimator that can predict all noise parameters for multiple noise types. The first contribution of our research was to design a noise data feature extractor that can effectively extract noise information from the image pair. The second contribution of our work leveraged other noise parameter estimation algorithms that can only predict one type of noise. Our proposed method, DE-G, can estimate additive noise, multiplicative noise, and impulsive noise from single-source images accurately. We also show the capability of the proposed method in estimating multiple corruptions.
Repetitive transients are usually generated in the monitoring data when a fault occurs on the machinery. As a result, many methods such as kurtogram and optimized Morlet wavelet and kurtosis method are proposed to extract the repetitive transients for fault diagnosis. However, one shortcoming of these methods is that they are constructed based on the index of kurtosis and are sensitive to the impulsive noise, leading to failure in accurately diagnosing the fault of the machinery operating under harsh environment. To address this issue, an optimized SES entropy wavelet method is proposed. In the proposed method, the optimized parameters including bandwidth and central frequency of Morlet wavelets are selected. Then, based on the wavelet coefficients decomposed using the optimized Morlet wavelet, the SES entropy is calculated to select the scales of wavelet coefficients. Finally, the repetitive transients are reconstructed based on the denoising wavelet coefficients of the selected scales. One simulation case and vibration data collected from the experimental setup are used to verify the effectiveness of the proposed method. The simulated and experimental analyses showed that the signal-to-noise ratio (SNR) of the proposed method has the largest value. Specifically, the SNR in the experimental analysis of the proposed method is 0.6, while that of the other three methods is 0.043, 0.0065, and 0.0045, respectively. Therefore, the result shows that the proposed method is superior to the traditional methods for repetitive transient extraction from the vibration data suffered from impulsive noise.