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

TL;DR

This study evaluates how patch size, sampling method, and dataset composition affect the accuracy of U-Net models for landslide segmentation using satellite and DEM data.

Contribution

It systematically compares different patch sizes and sampling methods, demonstrating their impact on landslide detection accuracy with U-Net models.

Findings

128x128 patch size with random sampling yields highest F1-score.

Smaller patches increase recall but also false positives.

DEM and data augmentation improve model accuracy.

Abstract

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.