The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part Bl. Beijing 2008
178
point of view. In real-life practice, a user’s top priorities,
typically data quality and project cost-efficiency, may or may
not be directly fulfilled by the “better” numbers presented by
the instrument manufacturer.
Moreover, owing to a lack of generally accepted guidelines for
lidar performance characterization, lidar system manufacturers
may choose different methodologies to characterize and to
present system performance capabilities. That is why it is very
important for the user of a commercial lidar system to
understand the underlying technical premises behind values on
a specification sheet and to make informed decisions to fulfil
project requirements (Figure 1). This paper will help lidar
system users to understand the underlying relationships among
various numbers on an airborne lidar specification sheet and to
bridge the gap between the “bare” numbers and the expected
real-life performance capabilities of a lidar system.
2. DATA COLLECTION EFFICIENCY VERSUS
PERFORMANCE SPECIFICATIONS
How fast can the system collect data? How quickly can the
project be completed? In other words, how cost-efficient is the
lidar system? Contrary to a mistaken assumption, the most cost-
efficient approach is not simply to set every operating
parameter of a lidar system to its maximum capacity. In fact,
the operating parameter that contributes most pertinently to
maintaining high density of points and achieving maximum
area coverage rate is laser pulse repetition frequency (PRF).
Because of its direct connection with data collection rate for
achieving survey time cost-effectiveness, PRF has become a
prime differentiating factor in the marketing of both lidar
sensors and data collection services (Flood, 2001). However,
considering PRF as a sole figure of merit without its connection
to other lidar parameters can be misleading. We will describe
how different mechanisms used for laser beam deflection and
scan pattern may affect point density and area coverage rate and,
hence, the operating parameters and cost-efficiency of a survey.
2.1 Link: PRF and Scan Patterns
Several scanning techniques, each with advantages and
disadvantages, are employed in airborne lidar systems. The
most common are (a) constant-velocity rotating polygon mirror
and (b) oscillating mirror (Figure 2). The advantage of a
rotating polygon mirror is its scan pattern, which appears as
linear unidirectional and parallel scan lines on the target.
However, its primary disadvantage is that for a certain period of
time during each rotation cycle, range measurement is either not
taken or, if taken, should then be discarded. As a result, with
this type of scanning mechanism, laser PRF does not equate
with data collection rate; hence, in most cases, manufacturers
specify the PRF and data measurement rate separately.
The oscillating mirror scan mechanism seems to be more
popular for airborne lidar systems. The mirror is always
pointing to the ground, and the system’s laser PRF is equivalent
to its data collection rate. However, laser PRF does not
immediately translate to area coverage rate for a given point
spacing because two distinct oscillating scan patterns—
sawtooth and sinusoidal—manifest two different laser point
distribution outcomes (Figure 3). In a sawtooth scan pattern,
scanner velocity is kept constant for most of the swath. This
gives an almost uniform point distribution across the swath with
slightly increasing point density towards the scan edges. In a
sinusoidal scan, point density is the lowest at the centre of the
swath and grows toward the edges of the scan line. That is why
a lidar with a sinusoidal scan pattern has to operate at a much
higher laser PRF to maintain the same nadir point density as a
lidar with a sawtooth scan pattern. It was shown (Ussyshkin et
al., 2008b) that at a 1-km flying altitude, 30-Hz scan frequency,
and ±25° scan angle, a lidar with a sinusoidal scan pattern has to
operate at a 158-kHz PRF to achieve the same point density at
nadir as a sawtooth scanner operating at 100 kHz.
Figure 2. Scan patterns: (a) constant velocity rotating polygon
mirror and (b) oscillating mirror
Sinusoidal Pattern, Non-Uniform Sawtooth Pattern, Mostly Uniform
Figure 3. Sawtooth and sinusoidal scan patterns from an
oscillating mirror scanner
Thus, from the user’s point of view, laser PRF cannot be the
only figure of merit for data collection efficiency since scan
pattern significantly changes cross-track point distribution and
affects one of the most important project requirements—ground
point density. To meet project-required point density on the
ground and to maximize area coverage rate, the lidar system
user must carefully consider the choice of scan pattern along
with laser PRF.
2.2 Link: Scan Frequency and Scan Angle
Maximum scan frequency, as specified on a lidar system
specification sheet, is another very important instrument
parameter that affects data collection efficiency achievable in
the field. Again, comparing “bare” numbers of the maximum
scan frequency may be quite misleading.
Scan rate for rotating polygon mirror versus oscillating
mirror: With a rotating polygon mirror, the scan rate is the
number of scan lines per second (Figure 2a). For example, a
100 Hz scan rate means that the scanner can provide 100
parallel scan lines every second. With an oscillating mirror, a
scan frequency of 100 Hz means that the mirror completes 100