An autoencoder-based model for forest disturbance detection using Landsat time series data

ABSTRACT Monitoring and classifying disturbed forests can provide information support for not only sustainable forest management but also global carbon sequestration assessments. In this study, we propose an autoencoder-based model for forest disturbance detection, which considers disturbances as anomalous events in forest temporal trajectories. The autoencoder network is established and trained to reconstruct intact forest trajectories. Then, the disturbance detection threshold is derived by Tukey’s method based on the reconstruction error of the intact forest trajectory. The assessment result shows that the model using the NBR time series performs better than the NDVI-based model, with an overall accuracy of 90.3%. The omission and commission errors of disturbed forest are 7% and 12%, respectively. Additionally, the trained NBR-based model is implemented in two test areas, with overall accuracies of 87.2% and 86.1%, indicating the robustness and scalability of the model. Moreover, comparing three common methods, the proposed model performs better, especially for intact forest detection accuracy. This study provides a novel and robust approach with acceptable accuracy for forest disturbance detection, enabling forest disturbance to be identified in regions with limited disturbance reference data.

[1]  Hankui K. Zhang,et al.  Characterization of Landsat-7 to Landsat-8 reflective wavelength and normalized difference vegetation index continuity. , 2016, Remote sensing of environment.

[2]  Zhiqiang Yang,et al.  Detecting trends in forest disturbance and recovery using yearly Landsat time series: 1. LandTrendr — Temporal segmentation algorithms , 2010 .

[3]  Yong Haur Tay,et al.  Abnormal Event Detection in Videos using Spatiotemporal Autoencoder , 2017, ISNN.

[4]  S. Goward,et al.  An automated approach for reconstructing recent forest disturbance history using dense Landsat time series stacks , 2010 .

[5]  Varun Chandola,et al.  Anomaly detection for symbolic sequences and time series data , 2009 .

[6]  J. Tukey,et al.  Performance of Some Resistant Rules for Outlier Labeling , 1986 .

[7]  R. Lucas,et al.  New global forest/non-forest maps from ALOS PALSAR data (2007–2010) , 2014 .

[8]  Sungzoon Cho,et al.  Variational Autoencoder based Anomaly Detection using Reconstruction Probability , 2015 .

[9]  Evan B. Brooks,et al.  How Similar Are Forest Disturbance Maps Derived from Different Landsat Time Series Algorithms , 2017 .

[10]  Rob J Hyndman,et al.  Detecting trend and seasonal changes in satellite image time series , 2010 .

[11]  Warren B. Cohen,et al.  Detecting trends in forest disturbance and recovery using yearly Landsat time series: 2. TimeSync — Tools for calibration and validation , 2010 .

[12]  Zhe Zhu,et al.  Mapping forest change using stacked generalization: An ensemble approach , 2018 .

[13]  Joanne C. White,et al.  Forest Monitoring Using Landsat Time Series Data: A Review , 2014 .

[14]  Arturo Sanchez-Azofeifa,et al.  Assessing the accuracy of detected breaks in Landsat time series as predictors of small scale deforestation in tropical dry forests of Mexico and Costa Rica , 2019, Remote Sensing of Environment.

[15]  W. Cohen,et al.  North American forest disturbance mapped from a decadal Landsat record , 2008 .

[16]  Mariana Belgiu,et al.  Random forest in remote sensing: A review of applications and future directions , 2016 .

[17]  Michael A. Wulder,et al.  Opening the archive: How free data has enabled the science and monitoring promise of Landsat , 2012 .

[18]  W. Buermann,et al.  Mapping tropical disturbed forests using multi-decadal 30 m optical satellite imagery , 2019, Remote Sensing of Environment.

[19]  Zhiqiang Yang,et al.  A LandTrendr multispectral ensemble for forest disturbance detection , 2018 .

[20]  Zhe Zhu,et al.  Cloud detection algorithm comparison and validation for operational Landsat data products , 2017 .

[21]  J. Im,et al.  Machine learning approaches for forest classification and change analysis using multi-temporal Landsat TM images over Huntington Wildlife Forest , 2013 .

[22]  Zhiqiang Yang,et al.  Continuous monitoring of land disturbance based on Landsat time series , 2020, Remote Sensing of Environment.

[23]  C. Woodcock,et al.  Continuous change detection and classification of land cover using all available Landsat data , 2014 .

[24]  Miroslav Svoboda,et al.  Forest disturbances under climate change. , 2017, Nature climate change.

[25]  C. Justice,et al.  High-Resolution Global Maps of 21st-Century Forest Cover Change , 2013, Science.

[26]  Michael Dixon,et al.  Google Earth Engine: Planetary-scale geospatial analysis for everyone , 2017 .

[27]  W. Cohen,et al.  Spatial and temporal patterns of forest disturbance and regrowth within the area of the Northwest Forest Plan , 2012 .

[28]  Zhiqiang Yang,et al.  Implementation of the LandTrendr Algorithm on Google Earth Engine , 2018, Remote. Sens..

[29]  Joseph Mascaro,et al.  Combating deforestation: From satellite to intervention , 2018, Science.

[30]  Dongdong Hou,et al.  Anomaly detection via adaptive greedy model , 2019, Neurocomputing.

[31]  Changqing Ke.  The discrimination of soil erosion features in the red soil region using remote sensing , 2005, SPIE Remote Sensing.

[32]  Randy C. Paffenroth,et al.  Anomaly Detection with Robust Deep Autoencoders , 2017, KDD.

[33]  Geoffrey E. Hinton,et al.  Reducing the Dimensionality of Data with Neural Networks , 2006, Science.

[34]  Giles M. Foody,et al.  Good practices for estimating area and assessing accuracy of land change , 2014 .

[35]  Zhe Zhu,et al.  Current status of Landsat program, science, and applications , 2019, Remote Sensing of Environment.

[36]  Charu C. Aggarwal,et al.  Outlier Detection with Autoencoder Ensembles , 2017, SDM.

[37]  Carol Miller,et al.  A New Metric for Quantifying Burn Severity: The Relativized Burn Ratio , 2014, Remote. Sens..

[38]  Michael Schultz,et al.  Performance of vegetation indices from Landsat time series in deforestation monitoring , 2016, Int. J. Appl. Earth Obs. Geoinformation.

[39]  Simon D. Jones,et al.  A fusion approach to forest disturbance mapping using time series ensemble techniques , 2019, Remote Sensing of Environment.

[40]  Takehisa Yairi,et al.  Anomaly Detection Using Autoencoders with Nonlinear Dimensionality Reduction , 2014, MLSDA'14.