Estimation of urban building heights from ICESat-2 laser altimeter data and quality control

Abstract

As a new satellite laser altimetry technology, ICESat-2 altimeter data from photon counting detection could explore the potential for urban building height measurement. However, previous studies have rarely involved quality control in this field, especially under complex detection conditions, which may lead to unreliable applications. Additionally, most studies remain at the exploratory phase of algorithm development, requiring further research to address processing techniques. This article proposes a cable suspension criterion for estimating urban building heights with more efficient processing based on the data characteristics of ICESat-2 photons, and constructs a quality control method based on the photon spatial distribution and quality evaluation to ensure the reliability of applications. We selected Hong Kong as the study area as there are hundreds of ICESat-2 track data in this area for evaluating the performance of our methods. We used the airborne LiDAR data (vertical accuracy: 0.1 m at 95% confidence level) as the validation data, and the results indicate that the proposed methods can estimate urban building heights with high accuracy and reliability. The root mean square error (RMSE) of building height estimation results before quality control is about 23.61 m, and that after quality control is about 1.06 m. We further evaluated the proposed method in the areas with different topographic reliefs (slopes) and building heights. It is found that in the areas with lower slopes or high building heights, the RMSE of building height estimation after quality control can be smaller than 1 m. This study offers a valuable addition to the exploration of the potential of measuring building heights using ICESat-2 altimeter data under complex urban settings and promises a global urban mapping direction of satellite LiDAR.