Accession Number : AD1009332


Title :   Terrain Classification Using Multi-Wavelength Lidar Data


Descriptive Note : Technical Report


Corporate Author : Naval Postgraduate School Monterey United States


Personal Author(s) : Thomas,Judson J


Full Text : https://apps.dtic.mil/dtic/tr/fulltext/u2/1009332.pdf


Report Date : 01 Sep 2015


Pagination or Media Count : 85


Abstract : With the arrival of Optechs Titan multispectral LiDAR sensor, it is now possible to simultaneously collect three different wavelengths of LiDAR data. Much of the work performed on multispectral LiDAR data involves gridding the point cloud to create Digital Elevation Models and multispectral image cubes. Gridding and raster analysis can have negative implications with respect to LiDAR data integrity and resolution. Presented here is a method of attributing the Titan LiDAR point cloud with the spectral information of all three lasers and the potential improvement of performing all analysis within the point cloud. Data from the Optech Titan are analyzed for purposes of terrain classification, adding the spectral component to the LiDAR data point cloud analysis. The approach used here combines the three spectral sensors into one point cloud, integrating the intensity information from the 3 sensors. Nearest-neighbor sorting techniques are used to create the merged point cloud. Standard LiDAR and spectral classification techniques are then applied. The ENVI spectral tool n-Dimensional Visualizer is used to extract spectral classes from the data, which can then be applied using supervised classification functions. The Maximum Likelihood classifier provided consistent results demonstrating effective terrain classification for as many as eleven classes.


Descriptors :   Terrain , classification , DIGITAL ELEVATION MODELS , MULTISPECTRAL , LIDAR , POINT CLOUDS , MAXIMUM LIKELIHOOD ESTIMATION , SUPERVISED MACHINE LEARNING


Subject Categories : Cartography and Aerial Photography
      Optical Detection and Detectors


Distribution Statement : APPROVED FOR PUBLIC RELEASE