Research On EMD Denoising Method Of Photon Count Lidar Signal

Proceeding from the signal feature of time-correlated single-photon counting laser radar, we discuss the radar signal noise reduction based on Empirical Mode Decomposition, which could analysis nonlinear and non-stationary time series automatically. First of all, the principle of TCSPC laser radar and EMD shifting process are recommended, and on the base of existing EMD reducing methods, a new distinguishing approach of correlative modes is proposed to pick out noise dominant IMFs. Secondly, by improving direct-threshold and indirect-threshold de-noising processes, we come up with an adaptive threshold method, in addition, we also contrast the results of some stationary and non-stationary signal simulation experiments using different reducing methods. Finally, we design variable radar ranging experiments, and put forward a new smooth threshold method to lower noise level of the measured signals, besides, we also design a piece of software to do de-noising works.The results show that:the distinguish approach of correlative modes may discriminate noise dominant IMFs effectively while the SNR ranges. In addition, compared with direct or indirect threshold de-noising process, the adaptive threshold method shows better denoising results, then, it also does better in adaptation, noise reduction effects of signals with different SNR and details preserving by contrast with Fourier transform, Wavelet Transform and EMD time scale methods. Beyond that, the smooth threshold method could reduce the noise of actual measurements with effect.