The Electronically Steerable Flash Lidar: A Full Waveform Scanning System for Topographic and Ecosystem Structure Applications

Abstract

The electronically steerable flash lidar (ESFL) is a waveform lidar sensor that incorporates two advances relevant to the design of future spaceborne lidar sensors. The first is a nonmechanical scanner that splits a single incoming beam into a variable number of output beams that can be aligned independently across track; the transmitted beam pattern can be changed up to 60 Hz. The second is a flash focal plane array (FFPA) capable of recording waveforms simultaneously from a 128 t 128 pixel grid with individual footprints spread over multiple pixels. In this paper, the incoming beam was used to illuminate eight 8.4-m footprints which were imaged simultaneously on 12 t 12 pixel subsets of the FFPA. The FFPA digitizes waveforms at a vertical resolution of 75 cm over 41 vertical bins to create waveforms of 30.75-m depth. Multiple waveforms obtained using range-gating were combined for these analyses. ESFL data were collected at Manitou Experimental Forest (MEF), located in the Pikes Peak National Forest, Colorado, USA and the Stephen F. Austin Experimental Forest (AEF), located in the Angelina Forest, Nacogdoches, TX. We evaluated the use of individual pixel-level and aggregated footprint-level waveforms and alternate approaches to define the extent of each footprint in the focal plane array. Using discrete return lidar data as a reference, we evaluated the ability of ESFL lidar to estimate canopy height and compared the two sensors' rates of penetration to the terrain surface. We found the footprint-level waveforms were better suited for use with existing waveform processing techniques, although techniques for processing at the pixel-level appear feasible. Relationships between height estimates from each lidar data set were most closely related when footprint-level ESFL waveforms were calculated after removing pixels that had less than 50% of the maximum energy within that footprint. Regressions between ESFL and reference lidar data estimates of height explained 84% (AEF) and 85% (MEF) of variance; this study could not say definitively which method yielded the more accurate estimate of height.