Gaussian Process Regression for Forest Attribute Estimation From Airborne Laser Scanning Data

While the analysis of airborne laser scanning (ALS) data often provides reliable estimates for certain forest stand attributes-such as total volume or basal area-there is still room for improvement, especially in estimating species-specific attributes. Moreover, while the information on the estimate uncertainty would be useful in various economic and environmental analyses on forests, a computationally feasible framework for uncertainty quantifying in ALS is still missing. In this paper, the species-specific stand attribute estimation and uncertainty quantification (UQ) is approached using Gaussian process regression (GPR), which is a nonlinear and nonparametric machine learning method. Multiple species-specific stand attributes are estimated simultaneously: tree height, stem diameter, stem number, basal area, and stem volume. The cross-validation results show that GPR yields on average an improvement of 4.6% in estimate root mean square error over a state-of-the-art $k$ -nearest neighbors (kNNs) implementation, negligible bias and well performing UQ (credible intervals), while being computationally fast. The performance advantage over kNN and the feasibility of credible intervals persists even when smaller training sets are used.

[1]  Göran Ståhl,et al.  Model-assisted estimation of biomass in a LiDAR sample survey in Hedmark County, NorwayThis article is one of a selection of papers from Extending Forest Inventory and Monitoring over Space and Time. , 2011 .

[2]  Yong Pang,et al.  Characterizing forest canopy structure with lidar composite metrics and machine learning , 2011 .

[3]  Sébastien Da Veiga,et al.  Gaussian process modeling with inequality constraints , 2012 .

[4]  E. Næsset.  Practical large-scale forest stand inventory using a small-footprint airborne scanning laser , 2004 .

[5]  E. Næsset,et al.  Forestry Applications of Airborne Laser Scanning , 2014, Managing Forest Ecosystems.

[6]  S. Popescu,et al.  Lidar remote sensing of forest biomass : A scale-invariant estimation approach using airborne lasers , 2009 .

[7]  Carl E. Rasmussen,et al.  Gaussian processes for machine learning , 2005, Adaptive computation and machine learning.

[8]  Virpi Junttila,et al.  Sparse Bayesian Estimation of Forest Stand Characteristics from Airborne Laser Scanning , 2008 .

[9]  Andrew Gordon Wilson,et al.  Deep Kernel Learning , 2015, AISTATS.

[10]  Petteri Packalen,et al.  Comparing nearest neighbor configurations in the prediction of species-specific diameter distributions , 2018, Annals of Forest Science.

[11]  E. Næsset,et al.  Using pre-classification to improve the accuracy of species-specific forest attribute estimates from airborne laser scanner data and aerial images , 2014 .

[12]  Rajnish Chauhan,et al.  Light Detection and Ranging , 2017, Encyclopedia of GIS.

[13]  Petteri Packalen,et al.  Uncertainty Quantification in ALS-Based Species-Specific Growing Stock Volume Estimation , 2017, IEEE Transactions on Geoscience and Remote Sensing.

[14]  M. Maltamo,et al.  A Two Stage Method to Estimate Species-specific Growing Stock , 2009 .

[15]  S. Magnussen,et al.  Alternative mean-squared error estimators for synthetic estimators of domain means , 2016 .

[16]  M. Maltamo,et al.  Variable selection strategies for nearest neighbor imputation methods used in remote sensing based forest inventory , 2012 .

[17]  M. Maltamo,et al.  The k-MSN method for the prediction of species-specific stand attributes using airborne laser scanning and aerial photographs , 2007 .

[18]  S. Reutebuch,et al.  Light detection and ranging (LIDAR): an emerging tool for multiple resource inventory. , 2005 .

[19]  Mark J. Schervish,et al.  Nonstationary Covariance Functions for Gaussian Process Regression , 2003, NIPS.

[20]  R. Nelson,et al.  Model-based inference for biomass estimation in a LiDAR sample survey in Hedmark County, Norway , 2011 .

[21]  Andrew Gordon Wilson,et al.  Gaussian Process Kernels for Pattern Discovery and Extrapolation , 2013, ICML.

[22]  M. Maltamo,et al.  Effects of pulse density on predicting characteristics of individual trees of Scandinavian commercial species using alpha shape metrics based on airborne laser scanning data , 2008 .

[23]  E. Næsset.  Predicting forest stand characteristics with airborne scanning laser using a practical two-stage procedure and field data , 2002 .

[24]  William C. Horrace,et al.  Some results on the multivariate truncated normal distribution , 2005 .

[25]  José Cristóbal Riquelme Santos,et al.  A Preliminary Study of the Suitability of Deep Learning to Improve LiDAR-Derived Biomass Estimation , 2016, HAIS.

[26]  Virpi Junttila,et al.  Linear Models for Airborne-Laser-Scanning-Based Operational Forest Inventory With Small Field Sample Size and Highly Correlated LiDAR Data , 2015, IEEE Transactions on Geoscience and Remote Sensing.

[27]  Thomas B. Schön,et al.  Linearly constrained Gaussian processes , 2017, NIPS.

[28]  Daniel Mandallaz,et al.  Sampling Techniques for Forest Inventories , 2007 .

[29]  E. Næsset,et al.  Remote sensing and forest inventories in Nordic countries – roadmap for the future , 2018 .

[30]  Geoffrey E. Hinton,et al.  Using Deep Belief Nets to Learn Covariance Kernels for Gaussian Processes , 2007, NIPS.

[31]  John B. Bradford,et al.  Hierarchical Bayesian spatial models for predicting multiple forest variables using waveform LiDAR, hyperspectral imagery, and large inventory datasets , 2013, Int. J. Appl. Earth Obs. Geoinformation.

[32]  Edwin V. Bonilla,et al.  Multi-task Gaussian Process Prediction , 2007, NIPS.

[33]  Mikko Kolehmainen,et al.  Neural Networks for the Prediction of Species-Specific Plot Volumes Using Airborne Laser Scanning and Aerial Photographs , 2010, IEEE Transactions on Geoscience and Remote Sensing.