Wind Lidar signal denoising method based onSingular Value Decomposition and Variational ModeDecomposition

A denoising method based on singular value decomposition (SVD) and variational mode decomposition (VMD) is proposed for wind lidar. Utilizing the covariance matrix based lidar signal simulation model, the performance of VMD, SVD, and VMD-SVD is evaluated. The results show that the VMD-SVD method is of better performance, and the output signal-to-noise ratio (SNR) is about 12 dB at the input SNR of −9dB. The actual lidar signals processing is performed with this combined denoising method, and the detection range and wind speed at pulse accumulation numbers of 50,100, and 300 are compared. We set the wind speed resulting from noisy signal with pulse accumulation number of 300 as the reference wind speed, and the mean value and standard deviation of wind differences are analyzed. The results show that the denoising method can not only increase the detection range while ensuring the accuracy of wind speed estimation but also achieve the same detection distance with fewer pulse accumulations, thereby improving the temporal resolution. For the pulse accumulation number of 50, the detection range is extended to 24 km from 18.45 km, and the standard deviation of speed difference is 0.88 m/s; for the same detection range, the temporal resolution is increased by about 6 times.