Three dimensional (3D) spatial metrics for objects

Abstract

The increasing availability of lidar data and structure from motion processing techniques is moving pattern metric research toward the development of three-dimensional (3D) analysis. There is a need to develop spatial pattern metrics that leverage 3D datasets, such as those derived from lidar or unmanned aircraft systems technology, that are meaningful and interpretable across landscape contexts. We introduce a suite of 3D spatial pattern metrics that can be computed on gradient surfaces such as digital surface models, but are rooted in traditional, patch-based landscape metrics that are familiar and interpretable across landscape contexts. We compute a suite of 3D metrics and demonstrate their use by analyzing a landscape pattern in the built environment of New Orleans in 2002 and 2008—pre- and post- Hurricane Katrina. Lidar data are used to segment individual buildings and calculate 3D patterns at the equivalent of the patch-, class- and landscape-levels for traditional landscape metrics. 3D spatial metrics can characterize landscape patterns at multiple spatial scales. These metrics capture aspect of pattern that traditional patch-mosaic and surface metrics cannot. Future research can build from these measures to develop other measures of 3D spatial patterns that are applicable for different ecological contexts. Continuing advances in full waveform lidar may contribute to the development of more complex metrics.