Echo signal extraction method of laser radar based on improved singular value decomposition and wavelet threshold denoising

Aiming at the problem of the weak echo signal for laser radar, the probability detection model of the laser radar is set up, and the influence of different signal to noise ratio and the threshold noise ratio is analyzed. The denoised process can effectively improve the detection probability. Therefore, two-level denoising framework with singular value decomposition and adaptive wavelet denoising is proposed. The improved method of selecting singular value based on curvature spectrum is proposed. On this basis, the selection criterion of decomposition level of adaptive wavelet denoising is put forward, and the optimal threshold is obtained by using continuous derivable threshold function and gradient descent. The joint denoising performance of singular value decomposition and adaptive wavelet denoising under Gaussian white noise is simulated and analyzed. The echo signal denosing experiments show that the amplitude of the peak noise is reduced by 10 times on the basis of the effective retention of the echo signal.