PL4U: Automated plane-based uncertainty evaluation and reduction method for indoor Mobile Laser Scanning systems

Mobile Laser Scanning (MLS) systems have obtained remarkable achievements in data acquisition and various scenario applications over the past decades. However, the investigation about their uncertainty evaluation has followed a different trend, significantly lagging behind the development pace of current MLS systems. A lack of automated, reliable, cost-effective, and commonly acceptable uncertainty evaluation solutions is evident. This paper presents an automated plane-based evaluation and reduction method for trajectory estimation errors of indoor MLS systems. The complete process can be split into plane correspondence establishment, parameter estimation and uncertainty reduction. The plane correspondence is mainly established by plane extraction, plane selection, plane matching, and matching verification. Then, a data-driven plane-based continuous random sampling estimation strategy is utilized to estimate 6 DoF trajectory estimation errors. Ultimately, a frame-wise error evaluation and reduction is achieved. For validation, two independent experiments are conducted, each covering distinct indoor scenarios and point clouds exhibiting different quality characteristics. The MLS point clouds evaluated in these experiments vary significantly in terms of noise levels and the presence of outliers, thus reflecting two common scenarios encountered in practical applications. Two independent experiments demonstrate that the results estimated by our method are highly consistent with the reference value and our method outperforms existing leading methods regarding overall effectiveness and robustness.