Noise and outlier points are inherent in point cloud data collected using vehicle-mounted lidar systems. This study introduces a statistical parameter adaptive filtering (SPAF) technique for denoising lidar point cloud data to mitigate the impact of automotive noise. First, based on the area of interest of various driving conditions, the application condition of filter for point cloud coarse noise reduction, and then use point of statistical features and interval statistics as clustering, calculate the neighborhood of each point cloud point number as well as multiple statistical filter cutoff threshold, and lastly suggested to enhance statistical filtering method to extract the noise reduction. The method completely takes into account the features of the point cloud in various driving settings after theoretical study and experimental verification, requiring no user input parameters. This study compares the radius filtering and bilateral filtering algorithms in the experiment with the suggested method. The results demonstrate that the proposed algorithm achieves an accuracy of up to 94%, outperforming conventional filtering methods.