<p>Under adverse weather conditions, LiDAR beams are affected, resulting in noise in the point cloud data. This weakens the LiDAR system’s ability to accurately recognize the surrounding environment. Eliminating the effects of adverse weather on LiDAR sensing systems has become a hot research topic. To provide the foundation for eliminating the effect of adverse weather on point cloud data, this paper investigates point cloud datasets, enhancement methods and denoising methods under adverse weather. Point cloud datasets collected under adverse weather conditions, such as rain, snow, and fog, are summarized and categorized based on the specific type of weather. Point cloud data enhancement methods designed for adverse weather conditions are reviewed and systematically classified into three categories: empirical model-driven, physical model-driven, and deep learning-based methods. The denoising methods for point cloud data under adverse weather conditions are systematically sorted out and categorized into voxel-grid-based, outlier-removal-based, and deep-learning-based methods. The strengths and weaknesses of various types of enhancement methods and denoising methods are compared. Built on these reviews, further research directions on point cloud datasets, point cloud enhancement methods and denoising methods under adverse weather conditions are discussed.</p>