Obtaining ground points using CSF Filter algorithm in various airborne LIDAR point cloud data

Abstract

Airborne laser scanning (ALS) is a remote sensing method widely recognized for its efficiency
in acquiring data quickly and delivering accurate results. To ensure the reliability of ALS data,
effective decontamination is crucial. This study aims to enhance the data quality of three
distinct LIDAR datasets representing urban, rural, and forest environments by applying the
CSF Filter algorithm in the CloudCompare software, an open-source tool widely used in point
cloud processing. The impact of various data characteristics and input parameters on the
filtering results was assessed through a series of comprehensive tests. The results of our
analysis revealed a notable relationship between the selected parameters and the quality of
the filtered data. Specifically, when the cover value within the CSF Filter parameters was
increased, a corresponding increase in data loss was observed, leading to significantly flawed
outcomes. These findings emphasize the importance of carefully selecting and fine-tuning the
input parameters to avoid undesirable consequences. The findings underscore the importance
of combining automated filtering algorithms with manual cleaning to achieve high-quality and
reliable point cloud data for various geospatial analyses and applications.