Interpine recently presented an update in the airborne LiDAR sensor technology available for use in New Zealand. This was given to the Forest Owners Association GCFF Phenotyping technical steering group. This focused on what sensor capability has historically been deployed in forestry applications and where our airborne service providers are moving in terms of latest LiDAR sensor adoption and application. Much of the LiDAR research work in NZ has focused on use of the OPTECH ATLM3100EA sensor. This sensor was released in 2006 and while having been very good to introduce LiDAR to the forest industry through initial adoption in the LUCAS (Ministry for the Environment, Landuse and Carbon Analysis System), and many subsequent research and commercial projects is now almost a decade old.
Looking to the immediate future there seems to be 4 major changes in the Airborne LiDAR effecting both the costs associated with flight planning, specifications and vegetation data quality being detected.
Digitised Full Waveform and Post Processing for Discrete Range Measurements (Returns, Points)
Some of the sensors now available are capable of storing in flight the digitised full waveform of each pulse. This results in a post processing workflow for the extraction of discrete returns LiDAR points from each pulse. The LiDAR systems deployed in New Zealand to date, such as the ATLM3100EA (2007-2014), the Pegasus HD500 (2014) and more recently the Orion HD300 (2015) are limited to discrete range measurements stored directly during flight (*optional full waveform storage have not been deployed in any of these systems). With the availability of the Riegl LMS-Q1560 and Trimble AX60 sensors we are moving to full digital waveform storage. This allows for extraction of potentially unlimited range measurements in each pulse. Therefore the definition of the sub canopy in our typical forests types may resulting in 7-15+ returns per pulse. This is a stark difference to what we have been used to with only 4 returns possible from the previous sensors. The realm of full waveform direct analysis for vegetation is still in the research camp at time of writing this article, but discrete return analysis workflows are well established for LiDAR.
Figure 1 – Showing the move from discrete returns storage only, to now post processing of full waveform to potentially unlimited returns per pulse.
Reflecting Mirror Types
The ALTM3100EA, Optech Pegasus HD5600 (Dual laser) and Orion HD300 units available implement an oscillating mirror technology for reflecting the laser energy across the swath. This creates a typical saw tooth or zig zag scan. One of the key limitations of this technology is the mechanical nature of the mirror design. This being the as the mirror moves from side to side it travels at changing velocity. As the mirror slows down to turn around and then subsequently speeds up again and point distribution is effected. This results in the uneven point distribution across the swath. Another approach now available in the Trimble AX60 and Riegl LMS-Q1560 units is the change to a rotating polygon mirror. This mechanical limitation is removed from the mirror and it spins at a constant velocity as it rotates through 360 degrees while reflecting the laser energy across the swath.
Figure 2 – Reflecting mirror design changes and the potential effect on point density and spacing.
This change in point distribution can be seen in the 3D z-raster whereby z is presented as 1m pulse density (analysis of last return points). The instability of the pulse density from an oscillating scanner is obvious, and results in the need to restrict scan angles and trim overage to remove much of this instability. Much of the research around scan angle limits and stability of vegetation metrics is built from scanners with this oscillating mirror technology. In this image field of view used at data collection was 28 degrees or 14 of Nadir.
Figure 3 – Reflecting mirror design changes and the potential effect on point density and spacing.
Reviewing a dataset from a rotating polygon mirror such as the Riegl LMS-Q1560 is shown below in the same 3D z-raster format. Although the colour range is different it is quickly obvious the pulse density variation is no longer evident. The other consideration here is this image is from a scan angle of 58 Field of View (FOV) or 29 degrees of Nadir.
Figure 4 – Reflecting mirror design changes and the potential effect on point density and spacing.
The ALTM3100EA unit used in most forestry projects through to 2014 in New Zealand had a minimum 3m echo separation distance. This echo separation or distance between vertical range measurements (m) therefore impacts on the definition of the vegetation canopy and therefore the stability and reliability of sub canopy metrics used for later analysis. With the sensors now available such as the Optech Orion HD300, Optech Pegasus HD500, Riegl LMS-Q1560 or Trimble AX60 this to 0.4-0.7m. Combining this with the move to digitised full waveform and user defined post processing of this to discrete returns means we are moving to potentially unlimited returns from a single pulse. This greatly increases the definition of the vegetation structure and more importantly in analysis to date brings stability in new metrics entering as predictors for our important response variables.
Figure 5 – Showing the technology changes from 3m echo separation to now the potential of 0.4m separation with full waveform post processing of discrete returns.
Multiple Pulses in the Air
LiDAR mapping sensors use time of flight of laser pulses and detected returns as the basis for measurement. LiDAR sensors such as the ALTM3100EA were only able to track one outgoing pulse in the air at a time (T0_1), and so before T0_2 could be transmitted the returns from T0_1 had to be detected. The most recent generation sensors operate with more than one outgoing pulse in the air at a time. This is often referred to multi-time-around or multi pulse. Sensors like the Pegasus HD500 or Orion HD300 bring this ability to tracking 2 pulses per laser in the air at any one time. A sensor which is dual laser such as the Pegasus HD500 can therefore track 2x 2 pulses in the air or 4 outgoing pulses in the air at a time. The Riegl LMS-Q1560 or Trimble AX60 units have the ability to track up to 12 pulses in the air per laser. In the case with the dual laser Riegl LMS-Q1560 this means up to 2x 12 or 24 pulses in the air at a time. The potential impact of this on data, flight planning and cost effectiveness is obvious. Interpine are currently still reviewing this in more detail and this is worth a blog article in itself so as we review more datasets from these sensors in differing terrain we will post more information.
This presentation explored the currently available sensors from our local service providers. A future review is looking at where LiDAR sensors are heading for future vegetation assessment.
If you would like to know more about LiDAR in general and its application in vegetation assessment, don’t hesitate to contact our team. The presentation can also be made available.