Comparing features of single and multi-photon lidar in boreal forests

Abstract

The emerging single-photon laser scanning has made technological breakthrough in the collection of airborne laser scanning data. In principle, single-photon systems require only one detected photon for successful ranging. Further, the point density on the ground can be 10–100 times higher for single-photon lidar data than that obtained with multi-photon systems at the same flight altitude. This has great potential to reduce operation costs. Single-photon lidar technology is assumed to be the best for data acquisition when high point densities are required over very large areas, or when improvements in measurement rates can significantly reduce data acquisition costs, such as in nationwide laser scanning programmes, where the whole country is repeatedly covered with data every 5–10 years. This study investigates single-photon lidar and conventional multi-photon laser scanning data for their potential in characterizing ground and forest attributes. Performance is evaluated in a boreal forest by a comparative analysis, where single-photon lidar measurements with SPL100 (Leica/Hexagon) from two flight heights (1900 m and 3800 m) are compared with data from the Optech Titan (400 m) multi-photon airborne laser scanning (ALS) under summer conditions (i.e. leaves on). We found that SPL100 from both altitudes provides forest attribute estimates with comparable accuracy to that of Optech Titan from 400 m using an area-based method. This demonstrates that point density and flight altitude do not have significant impact on forest attribute estimation using the area-based approach. As a result, SPL100 is a cost-efficient alternative to a conventional laser scanner for forest inventories at large scale. There are systematic differences in behavior of the data sets due to differences in ranging sensitivity, beam size, and point density. We observed a higher proportion of ground returns in the SPL100 (3800 m) than in SPL100 (1900 m) data. Both SPL100 data in general produced a higher proportion of ground returns than Titan single channel did in structurally more homogeneous and one layer stands while higher proportion of ground returns from Titan than from SPL100 data in multi-layer stands. Forest structure and flight altitude has a notable impact on the distribution of points and further characteristics of the vertical structures. The pulse of Titan sensor penetrated deeper into the canopy than SPL100.