Author :
Publisher :
ISBN 13 :
Total Pages : 7 pages
Book Rating : 4.:/5 (13 download)
Book Synopsis Comparing Derivatives from Lidar and Image-derived Point Clouds by :
Download or read book Comparing Derivatives from Lidar and Image-derived Point Clouds written by and published by . This book was released on 2016 with total page 7 pages. Available in PDF, EPUB and Kindle. Book excerpt: Accurate three-dimensional (3-D) geospatial data can provide valuable information on forest canopy structure, tree height, and topography. Often this 3-D data is gathered via lidar, but due to the cost of lidar it is rarely acquired for monitoring purposes, but rather as a one-time collect. While image matching techniques have been around for a number of years, advances in computer technology made execution of image matching algorithms practical. Stereo imagery, such as Forest Service resource photography, is less expensive than lidar, is acquired more frequently, and has a longer history in the agency, thus making it a potential tool for monitoring forest canopy structure and not just forest cover. To assess the value of image-derived point clouds we compared them with corresponding lidar point clouds to determine if they are a suitable replacement when lidar isnt available or is too expensive to collect a second time. To accomplish this, we derived and compared three productscanopy height, canopy cover and a digital elevation model (DEM)from both point cloud datasets for an area located on the Coconino National Forest and calculated error between the image-derived and lidar-derived layers. RMSE values ranged from 7.87 to 8.58 m for canopy height and 8.44 to 9.21 m for the DEM. The RMSE for the canopy cover metric was 22.82 percent. The acquisition and metadata characteristics (e.g., image overlap, ground control, absence of full aero-triangulation and lack of a neat model) of the aerial imagery used for this project were insufficient to allow the derivation of accurate 3-D point clouds. Future work can build upon this effort by performing similar comparisons with aerial imagery that has been acquired with image specifications better suited to generating point clouds, such as increased overlap and sidelap (i.e., 80 percent and 60 percent, respectively), more ground control points, and the inclusion of a neat model.