An algorithm that combines a widely used sonar data processing method and a newly developed machine-learning-based algorithm to extract of shallow-water bathymetry from LiDAR point clouds was evaluated for accuracy and potential operationalisation. Data comprised 103 500 m-by-500 m data tiles located near the Florida Keys (United States) representing an operationally realistic range of environmental and data conditions. Tiles are processed tiles individually to classify each LiDAR pulse return (“sounding” in hydrographic terminology) as bathymetry or not.  Compared to a reference classification an average agreement of about 90% was produced; accuracy varied depending on ocean bottom and data conditions.  The average false negative rate – the most important metric in hydrographic mapping – was about 5%. Processing time for tiles containing the average number of soundings (seven million) on a desktop computer was approximately 100 minutes. A major advantage is that the algorithm does not require in situ ground-“truth” data for training or calibration.