Features of point clouds synthesized from multi-view ALOS/PRISM data and comparisons with LiDAR data in forested areas

LiDAR waveform data from airborne LiDAR scanners (ALS) e.g. the Land Vegetation and Ice Sensor (LVIS) have been successfully used for estimation of forest height and biomass at local scales and have become the preferred remote sensing dataset However, regional and global applications are limited by the cost of the airborne LiDAR data acquisition and there are no available spacebome LiDAR systems. Some researchers have demonstrated the potential for mapping forest height using aerial or spaceborne stereo imagery with very high spatial resolutions. For stereo images with global coverage but coarse resolution new analysis methods need to be used. Unlike most research based on digital surface models, this study concentrated on analyzing the features of point cloud data generated from stereo imagery. The synthesizing of point cloud data from multi-view stereo imagery increased the point density of the data. The point cloud data over forested areas were analyzed and compared to small footprint LiDAR data and large-footprint LiDAR waveform data. The results showed that the synthesized point cloud data from ALOS/PRISM triplets produce vertical distributions similar to LiDAR data and detected the vertical structure of sparse and non-closed forests at 30 m resolution. For dense forest canopies, the canopy could be captured but the ground surface could not be seen, so surface elevations from other sources would be needed to calculate the height of the canopy. A canopy height map with 30 m pixels was produced by subtracting national elevation dataset (NED) from the averaged elevation of synthesized point clouds, which exhibited spatial features of roads, forest edges and patches. The linear regression showed that the canopy height map had a good correlation with RHSO of LVIS data with a slope of 1.04 and R-2 of 0.74 indicating that the canopy height derived from PRISM triplets can be used to estimate forest biomass at 30 m resolution. (C) 2014 Elsevier Inc All rights reserved.