Improving the efficiency of forest inventory with Terrestrial LiDAR: what about Hand-held Mobile LiDAR?

For a decade, studies of the application of static Terrestrial Laser Scanner (TLS) in plotwise forest inventories are giving more and more effective results. In spite of the improvements occurring in processing scan data to extract forest attributes, the occlusion effect is still limiting the processing efficiency. A multi-scan approach is recommended to reduce this effect. However, such approach needs pre-scan preparations (setting up the plot, targets positioning), it requires data registration and it comes at a higher data collection cost. In this study we explore the potential of a Hand-held mobile LiDAR System (HMLS) as new LiDAR tool to scan forest plots. HMLS data are compared to static TLS data (single and multi-scan) in terms of data acquisition, registration time and quality of automatic DBH extraction. The low weight, small size of the instrument and no targets requirements reduce the time of pre-scan preparations to the time needed for single scan which is 6 times less than scanning a plot with 5 scans. The registration time depends of the time spent to scan the plot and it is of the same magnitude than single scan. The resulting point cloud of the HMLS is noisier than TLS point clouds. Nevertheless, error on DBH estimations is similar to scanning a plot with a TLS positioned at 5 locations. RMSE is higher than multi-scan and close to single scan for trees detected by the both LiDAR technologies. This first study exhibits the high potential of HMLS by its simple use, which needs only one operator while presenting similar results in automatic DBH extraction than static TLS. Technology and registering method improvements of this type of mobile LiDAR will reduce the noise of the point cloud, which might reduce the DBH RMSE.