3 The use of remote sensing data to monitor Indonesian peatlands 3.1 Introduction to remote sensing
3.2 LiDAR data
Light Detection and Ranging (LiDAR) is an active remote sensing technique which is based on the transmission of laser pulses toward the ground surface and the recording of the return signal. By analyzing the time delay for each pulse back to the sensor, the heights of all reflecting objects can be measured in the range of a few centimeters. LiDAR systems are usually classified using three characteristics: (a) the type of recording the return signal, (b) footprint size, and (c) sampling rate and scanning pattern (Dubayah & Drake, 2000). Two recording types can be differentiated, the discrete-return and the full-waveform system (Figure I-8). For discrete-return systems, pulse detection is conducted in real-time on the returned signal, so that the system detector splits a continuous waveform into several time stamped pulses giving the position of the individual targets (Mallet & Bretar, 2009).
These laser scanning systems are called multi-echo or multi-pulse and typically collect first and last pulses but some are able to differentiate up to six individual returns from one pulse. The footprints of these systems are small reaching sizes of 0.2 to 0.9m. Full-waveform systems on the other hand record the amount of energy for a series of equal time intervals and give more control to the user as their processing methods increase pulse detection reliability, accuracy, and resolution. A certain amplitude against time waveform is obtained for each time interval. To understand these waveform pre-processing is necessary which is usually the
15 decomposition of these waveforms into a sum of echoes generating a three dimensional (3D) point cloud. Most commercial LiDAR systems nowadays are small- footprint systems (0.2 to 3.0m), depending on flying height and beam divergence, and a high repetition frequency. In this thesis data from an airborne and a spaceborne LiDAR system was analyzed.
Figure I-8: Conceptual differences between full-waveform and discrete-return LiDAR systems (Lefsky et al., 2002b; modified). In the left the intersection of a laser illumination area, or footprint, through a simplified tree crown is shown. In the center the hypothetical return signal collected by a full-waveform recording device is depicted. In the right three different discrete-return LiDAR sensors are indicated.
First-return LiDAR devices only record the position of the first object hit by the laser beam. Last-return LiDAR devices on the other had record the position of the last object hit by the laser beam and are especially useful for topographic mapping. Multiple-return LiDAR sensors record the positions of a smaller number of objects in the path of the illumination.
Airborne LiDAR data was acquired during a flight campaign conducted between 5 and 10 August 2007. A Riegl LMS-Q560 Airborne Laser Scanner was mounted to a
16
Bell 206 helicopter. Small-footprint full-waveform LiDAR data was collected from a flight altitude of 500m above ground over a scan angle of ±30° (swath width ±500m).
The laser sensor had a pulse rate of up to 100,000 pulses per second with a footprint of 0.25m and a wavelength of 1.5μm (near infrared). Due to the accurate time stamping (109 samples per second), the three dimensional coordinates of the laser beam reflections (x, y, and z), the intensity, and the pulse width can be extracted by a waveform decomposition, which fits a series of Gaussian pulses to the waveform.
This resulted in an average of 1.4 echoes per square meter. The Riegl LMS-Q560 Airborne Laser Scanner system allows height measurements of ±0.02m. Single beam measurements have an absolute horizontal accuracy of ±0.50m and vertical accuracy of ±0.15m Root Mean Square Error (RMSE).
The Ice, Cloud, and land Elevation Satellite (ICESat) has been orbiting the earth since 12 January 2003 at an altitude of 600km with a 94° inclination and during most of its operating life it has been programmed for a 91-day orbital repeat cycle and was decommissioned from operation on 14 August 2010. The Geoscience Laser Altimeter System (GLAS) onboard ICESat was a full waveform sensor using a 1,064nm laser operating at 40Hz. This resulted in a nominal footprint of about 65m diameter on the earth’s surface with each pulse separated by 172m postings (Schutz et al., 2005).
There were three lasers onboard ICESat of which the first one failed about 38 days into the mission (29 March 2003). The original temporally continuous measurements were replaced by three 33 day operating periods per year, so that the life of the second and third laser could be extended (Sun et al., 2008). The laser footprint on the earth’s surface actually was in the form of an ellipse and its size varied over time as a function of power output from the laser (Harding & Carajabal, 2005). As the GLAS sensor recorded the returned energy over time these waveforms represented the vertical distribution of the terrain and vegetation within each footprint. GLAS data have been demonstrated to accurately estimate forest height (Lefsky et al., 2007;
Rosette et al., 2008; Lefsky, 2010) and AGB (Harding & Carajabal, 2005; Boudreau et al., 2008). In this study we used the ICESat/GLAS data from release version 31.
According to The National Snow and Ice Data Center ICESat/GLAS this release version had an average horizontal geolocation error for all laser campaigns of 0.78
±5.09m (The National Snow and Ice Data Center, 2011b).
17