Abstract:<p>A segmentation singular value decomposition (SVD)-lifting wavelet transform (LWT) denoising algorithm based on ensemble empirical mode decomposition (EEMD) was proposed to better suppress noise in an atmospheric lidar return signal. The EEMD method is used to distinguish inherent modal functions (IMFs) of the noise and signal, and remove the IMF with noise as its main component. Moreover, the SVD-LWT method is adopted to remove the noise existing in the signal-based IMF and thus finely extract the signal. The simulated Bumps signal with different sequences of Gaussian white noise was denoised, and the denoising effect of the EEMD-SVD-LWT algorithm was compared with the effects of the wavelet soft threshold, EEMD (correlation coefficient), and EEMD (difference value) methods. Simulation shows that the denoising effect of the EEMD-SVD-LWT algorithm was best. The EEMD-SVD-LWT algorithm was also used to denoise practical lidar signals and was better than that achieved with the other methods.</p>
What problem does this paper attempt to address?
The paper attempts to address the problem of effectively suppressing noise and extracting useful information from atmospheric lidar echo signals. Specifically, the paper proposes a joint denoising algorithm based on Ensemble Empirical Mode Decomposition (EEMD), Singular Value Decomposition (SVD), and Lifting Wavelet Transform (LWT), aiming to better remove noise from atmospheric lidar echo signals and finely extract the signal.
### Background and Problem
1. **Application of Lidar Technology**:
- Lidar technology is widely used in meteorological observation and atmospheric environment monitoring, with advantages such as wide detection range, high spatiotemporal resolution, and strong anti-interference capability.
- In practical applications, lidar echo signals decay with the square of the distance, making far-field signals very weak and easily overwhelmed by noise.
2. **Sources of Noise**:
- Dark current generated by photodetectors, thermal noise from signal processing circuit components, and stray light from the sun and sky background can all cause far-field signals to be overwhelmed by noise.
- The complex and variable detection environment means that lidar echo signals carry not only information components of particles or gases but also noise signals, with useful signals potentially being completely submerged.
3. **Limitations of Existing Denoising Methods**:
- Existing denoising methods include digital filtering techniques such as Fourier Transform, Moving Average, Wavelet Transform, and Empirical Mode Decomposition (EMD).
- Fourier Transform is suitable for linear stationary signals but not for nonlinear and non-stationary signals, introducing distortion during denoising.
- The Moving Average method is simple but loses effective data points, reducing the accuracy of lidar detection.
- Wavelet Transform is superior to Fourier Transform and Moving Average methods in handling nonlinear and non-stationary signals but requires the selection of wavelet basis functions and decomposition levels.
- Empirical Mode Decomposition (EMD) is suitable for nonlinear and non-stationary signals but has issues such as mode mixing, stopping criteria, and endpoint effects.
### Solution
The paper proposes a joint denoising algorithm combining EEMD, SVD, and LWT (EEMD-SVD-LWT), with the following specific steps:
1. **EEMD Method**:
- Use the EEMD method to decompose the signal into multiple Intrinsic Mode Functions (IMFs) and a residual term.
- Determine the boundary between signal and noise by calculating the correlation coefficient differences between IMFs and the original signal, identifying the IMFs containing the signal.
2. **SVD Method**:
- Perform segmented Singular Value Decomposition (SVD) on the IMFs containing the signal to remove noise.
- Use a continuous overlapping segmented SVD method to reduce abrupt changes between adjacent segments and improve denoising effectiveness.
3. **LWT Method**:
- Use Lifting Wavelet Transform (LWT) to further remove noise from the IMFs, preserving the detailed features of the signal.
### Experimental Results
- Through simulation experiments and the denoising effect on actual lidar signals, the EEMD-SVD-LWT algorithm outperforms traditional denoising methods in terms of denoising effectiveness, effectively preserving the characteristics of lidar echo signals, with a higher signal-to-noise ratio and lower error.
In summary, the paper proposes an effective joint denoising algorithm that addresses the problem of noise suppression and signal extraction in lidar echo signals, with significant application value.