|
LiDAR Research: Fuels,
Structure, and Inventory
The members of the lab in conjunction with a
number of regional collaborators are involved in a strong collection of
integrated LiDAR research and extension activities. In addition to
cutting-edge research we have organized several workshops and conference
sessions aimed at disseminating our research to as many stakeholders as
possible. As part of this collaboration we are putting together a web-based
mini-tutorial on LiDAR and how it can be useful for forestry:
Follow This Link for our new Lidar Web Tutorial ...
Production and Assessment of LiDAR Bare Earth
Models.
Jeffery S. Evans and Andrew T Hudak at the
USDA Forest
Service Rocky Mountain Research Station, Moscow are leading the
development of a bare earth model (BEM) algorithm that is suitable in large
biomass areas. As highlighted in J.Evans and A.Hudak (IEEE Transactions on Geoscience and
Remote Sensing, in press) they have developed a new algorithm entitled the
'Progressive Curvature Filter'. A comparison of the
Progressive Curvature Filter with
the commonly applied 'Block Minimum' is shown below.
  
All LiDAR Returns
Block Minimum
Progressive Curvature Filter
This algorithm is
already widely used by the USDA Agricultural Research Service (ARS), the
USDA FS
Remote Sensing Applications Center (RSAC), in addition to our research
group. As part of ongoing research the lab is collaborating with
Dr Tim
Link (University of Idaho) and Dr Danny Marks (ARS) in the
USDA-ARS Owyhee Wildlands LiDAR Experiment (OWLX). As part
of this experiment the lab will be hiring a new MSc student [click here
for advert] to conduct accuracy assessment of the Bare Earth Models
collected here and in other regional forest datasets.
LiDAR Characterization of Vegetation
Structure and Fuels .
In addition to BEM algorithms, our
lab have been developing
novel methods to extract tree-level characteristics from canopy height
models (CHMs). Work by A.M.S. Smith, M.J. Falkowski et al (Falkowski et
al 2006; Canadian Journal of Remote Sensing) highlights how
'Spatial Wavelet Analysis'
can be applied to extract tree crown widths and maximum tree heights from an
open-canopy mixed conifer forest stand in northern Idaho.
Andrew Hudak et al
also has considerable experience in integrating LiDAR with
multispectral satellite
sensor imagery. Following on from Hudak et al (2002) in which LiDAR and
Landsat ETM+ data were integrated, Hudak et al (2006) used both ALI (EO-1
Satellite) with LiDAR to model parameters such as Basal Area and Tree Density
on Moscow Mountain (see Right).
Complimentary Regional LiDAR Collaborators.
The University of Idaho's
Department of Geography, Lidar research by J Rooker Jensen, K Humes,
and others have focused on evaluating the technology's capabilities to
measure forest inventory within the Nez Perce Reservation, northern Idaho.
At the
University of Montana's National Center for Fire Landscape Analysis, C
Sielestad, L Queen, et al are involved in the assessment of surface
fuel models. On a related topic, R.Lawrence at Montana State University is
investigating the fusion of polarimteric synthetic aperture radar and
hyperspectral imagery of landcover classification in Yellowstone National
Park. The research lab is also coordinating with the CLICK
group at the USGS - EROS Data Center, who have recently set-up a national
depositary for XYZ LiDAR Data.
|
