Abstract: In practical applications, lidar signals are often subject to interference from solar background light and the dark current of photodetectors, which compromises the accuracy and reliability of data. To effectively eliminate the noise within the signal, this paper introduces a denoising method based on successive variational mode decomposition (SVMD) and singular value decomposition (SVD). Firstly, the red-tailed hawk (RTH) algorithm is employed to optimize the parameters of SVMD, enabling a more precise decomposition of lidar signals and extraction of intrinsic mode functions (IMFs). Then, the entropy of each IMF is assessed by calculating permutation entropy (PE), which is categorized into effective components and noise component, and the effective component will be undergone SVD-based denoising. Simulation and empirical results demonstrate that, compared with the other methods, the proposed method yields the smoothest signal waveform after denoising, with the highest signal-to-noise ratio and the lowest root mean square error, and can effectively suppresses long-distance noise without distortion, showing superior denoising performance under the same conditions.

Key words: lidar, noise processing, successive variational mode decomposition, singular value decomposition