Landslide Segmentation with U-Net: Evaluating Different Sampling Methods and Patch Sizes

Landslide inventory maps are crucial to validate predictive landslide models; however, since most mapping methods rely on visual interpretation or expert knowledge, detailed inventory maps are still lacking. This study used a fully convolutional deep learning model named U-net to automatically segment landslides in the city of Nova Friburgo, located in the mountainous range of Rio de Janeiro, southeastern Brazil. The objective was to evaluate the impact of patch sizes, sampling methods, and datasets on the overall accuracy of the models. The training data used the optical information from RapidEye satellite, and a digital elevation model (DEM) derived from the L-band sensor of the ALOS satellite. The data was sampled using random and regular grid methods and patched in three sizes (32x32, 64x64, and 128x128 pixels). The models were evaluated on two areas with precision, recall, f1-score, and mean intersect over union (mIoU) metrics. The results show that the models trained with 32x32 tiles tend to have higher recall values due to higher true positive rates; however, they misclassify more background areas as landslides (false positives). Models trained with 128x128 tiles usually achieve higher precision values because they make less false positive errors. In both test areas, DEM and augmentation increased the accuracy of the models. Random sampling helped in model generalization. Models trained with 128x128 random tiles from the data that used the RapidEye image, DEM information, and augmentation achieved the highest f1-score, 0.55 in test area one, and 0.58 in test area two. The results achieved in this study are comparable to other fully convolutional models found in the literature, increasing the knowledge in the area.

[1]  Yongjun Zhang,et al.  End-to-End Change Detection for High Resolution Satellite Images Using Improved UNet++ , 2019, Remote. Sens..

[2]  Biswajeet Pradhan,et al.  Integration of LiDAR and QuickBird Data for Automatic Landslide Detection Using Object-Based Analysis and Random Forests , 2017 .

[3]  Qingyong Zhang,et al.  Automatic recognition of landslide based on CNN and texture change detection , 2016, 2016 31st Youth Academic Annual Conference of Chinese Association of Automation (YAC).

[4]  Andrea Manconi,et al.  Mapping Landslides on EO Data: Performance of Deep Learning Models vs. Traditional Machine Learning Models , 2020, Remote. Sens..

[5]  Trevor Darrell,et al.  Fully Convolutional Networks for Semantic Segmentation , 2017, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[6]  Andrea Manconi,et al.  Sentinel-1 SAR Amplitude Imagery for Rapid Landslide Detection , 2019, Remote. Sens..

[7]  Yifei Zhang,et al.  Automated Landslides Detection for Mountain Cities Using Multi-Temporal Remote Sensing Imagery , 2018, Sensors.

[8]  Chong Xu,et al.  Preparation of earthquake-triggered landslide inventory maps using remote sensing and GIS technologies: Principles and case studies , 2015 .

[9]  David Alexander,et al.  A brief survey of GIS in mass-movement studies, with reflections on theory and methods , 2008 .

[10]  Shamsul Fakhruddin,et al.  Application of Science and Technology for Sendai Framework for Disaster Risk Reduction 2015-2030 , 2016 .

[11]  Travis E. Oliphant,et al.  Guide to NumPy , 2015 .

[12]  A. Avelar,et al.  January 2011: The Extreme Landslide Disaster in Brazil , 2013 .

[13]  Dirk Tiede,et al.  DWELLING EXTRACTION IN REFUGEE CAMPS USING CNN – FIRST EXPERIENCES AND LESSONS LEARNT , 2018, The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences.

[14]  Xiao Xiang Zhu,et al.  Deep Learning in Remote Sensing: A Comprehensive Review and List of Resources , 2017, IEEE Geoscience and Remote Sensing Magazine.

[15]  Thomas Brox,et al.  U-Net: Convolutional Networks for Biomedical Image Segmentation , 2015, MICCAI.

[16]  B. Pradhan,et al.  Rainfall-induced landslide susceptibility assessment at the Chongren area (China) using frequency ratio, certainty factor, and index of entropy , 2016 .

[17]  André Stumpf,et al.  Object-oriented mapping of landslides using Random Forests , 2011 .

[18]  Willy A. Lacerda,et al.  Mechanisms of the Recent Catastrophic Landslides in the Mountainous Range of Rio de Janeiro, Brazil , 2013 .

[19]  Thomas Blaschke,et al.  Object based image analysis for remote sensing , 2010 .

[20]  W. Köppen Das geographische System der Klimate , 1936 .

[21]  Thomas Blaschke,et al.  Evaluation of Different Machine Learning Methods and Deep-Learning Convolutional Neural Networks for Landslide Detection , 2019, Remote. Sens..

[22]  Francisco Herrera,et al.  Deep-Learning Convolutional Neural Networks for scattered shrub detection with Google Earth Imagery , 2017, ArXiv.

[23]  G. Gois,et al.  VARIABILIDADE ESPAÇO-TEMPORAL E INTERANUAL DA CHUVA NO ESTADO DO RIO DE JANEIRO , 2018 .

[24]  Peng Gao,et al.  Landslide mapping with remote sensing: challenges and opportunities , 2020, International Journal of Remote Sensing.

[25]  Qiaosong Wang,et al.  Object Recognition in Aerial Images Using Convolutional Neural Networks , 2017, J. Imaging.

[26]  Markus Metz,et al.  GRASS GIS: A multi-purpose open source GIS , 2012, Environ. Model. Softw..

[27]  S. L. Kuriakose,et al.  Spatial data for landslide susceptibility, hazard, and vulnerability assessment: An overview , 2008 .

[28]  Biswajeet Pradhan,et al.  Landslide Detection Using Residual Networks and the Fusion of Spectral and Topographic Information , 2019, IEEE Access.

[29]  Alexandr A. Kalinin,et al.  Albumentations: fast and flexible image augmentations , 2018, Inf..

[30]  Fang Chen,et al.  Landslide detection based on contour-based deep learning framework in case of national scale of Nepal in 2015 , 2020, Comput. Geosci..

[31]  Thomas Blaschke,et al.  Geographic Object-Based Image Analysis – Towards a new paradigm , 2014, ISPRS journal of photogrammetry and remote sensing : official publication of the International Society for Photogrammetry and Remote Sensing.

[32]  F. Guzzetti,et al.  Landslide inventory maps: New tools for an old problem , 2012 .