International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B7, 2012
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia
mapping (Simard et al, 2008) and above-ground biomass
extraction (Lefsky et al., 2005). A high-altitude airborne sensor
tested was SLICER, providing ~10 m footprints at ~5000 m
altitude. Harding et al. (2001) proved that canopy height profiles
can be extracted from waveform data with correlation (R?)
values up to 0.75 can compared with ground truth. Lefsky et al.
(2005) used SLICER data to estimate above ground biomass
(AGBM) over different biomes, reaching correlation (R2) values
up to 0.85.
A strategy of processing waveform is to attempt for an
improvement in peak detecting in the post-processing phase.
Several methods have been tested and results show that the
Generalized Gaussian fitting method gives best results (Chauve
et al, 2007). Improved peak detection increases the number of
significant returns obtained from a survey. Tests have shown
that the number of returns — thus point density — can be
increased by a factor of two compared to conventional discrete-
return data (Reitberger et al, 2009). Point clouds with high
densities (> 10 point/m2) are necessary for extracting metrics
which use the distribution of 3D positions and intensities
(Reitberger et al., 2009) for the identification of single trees and
tree structure.
Metrics which use all the information obtainable from an ALS
survey have also been tested on large footprint (Drake et al,
2002) and small-footprint surveys. The objective of this paper is
to extract information on low vegetation (renovation) by
discrimination of terrain from low vegetation returns.
2. MATERIALS AND METHODS
2.1 Study area and ALS survey
The survey was done the 20th of June 2011 with a helicopter
carrying Optech’s ALTM 3100 sensor and a Rollei AIC modular
P45 digital metric camera. The surveyed area is a region whose
area is approximately 2.70 km?, located in the northwest part of
Italy. The region of interest (ROI) for this study is a smaller
portion (center at 7?29'54" longitude 45°46’18” latitude at
WGSS84 datum). Part of the ROI encompasses and area which
endured a severe fire event in 2005 which caused the destruction
of the forest stand. Figure 1 shows the ROI with circular sample
plots on five the five sectors which were tested for variability in
vegetation characteristics.
ctors with circular sample plots
Figure 1. ROI with the five se
Characteristic Value
Vehicle Helicopter
Sensor Optech ALTM 3100
Date of survey June 20 2011
Mean relative flight height ~525 m above ground
Scan angle* z2].5?
Scan frequency* 71.5 KHz
Output Datum ETRS2000 (2008) — WGS84
524
*Calculated directly from the output data
Table 1. Characteristics of the survey flight.
2.2 ALS data
In the text the term “waveform” will be used to indicate
amplitude as a function of time, specifically a vector of energy
values sampled at 1 ns intervals. The length of the vector
depends on the number of sampled energy values, which, in the
case of Optech’s waveform file format, is a variable number for
the return echo. In the case of the outgoing pulse (TO), the
energy is sampled constantly every 1 ns for 40 ns thus the TO
waveform is recorded in the NDF file using 40 samples. As can
be seen in figure 2 the maximum value does not always
correspond with the same sample time..
t9 (outgaing signal) waveform
M ES E
Figure 2. Plot of eight waveforms of the outgoing pulse.
2.3 Waveform recording format
All laser scanner data were stored by the system in Optech's
waveform data file formats, which consist in four types of files.
To process the waveform the following files were necessary:
- NDF file, where the recorded waveform data are stored
consisting of variable length records, each of which hold the
following information relative to one laser pulse: the GPS
timestamp, the outgoing pulse waveform and up to 7 significant
segments of the corresponding return echo waveform. The NDF
data are divided into frames, each containing 16838 records.
- IDX file, which is an index, relating starting and ending GPS
timestamps of a specific frame in the NDF file.
- CSD file holds information on trajectory, thus the position
and orientation of the vehicle at each laser pulse, as well as scan
angle and range of up to four returns.
- DGT file holds the index between frame start time and CSD
record number.
2.4 Sampling design
The ROI was divided into 5 sectors with different characteristics
in terms of forest cover and renovation technique; each area
was sampled with circular plots of 12.6 m radius, -500 m?
(figure 1). The characteristics for each area and the number of
plots are reported on the table below.
