Landslide detection using deep learning and object-based image analysis

Abstract

Recent landslide detection studies have focused on pixel-based deep learning (DL) approaches. In contrast, intuitive annotation of landslides from satellite imagery is based on distinct features rather than individual pixels. This study examines the feasibility of the integration framework of a DL model with rule-based object-based image analysis (OBIA) to detect landslides. First, we designed a ResU-Net model and then trained and tested it in the Sentinel-2 imagery. Then we developed a simple rule-based OBIA with only four rulesets, applying it first to the original image dataset and then to the same dataset plus the resulting ResU-Net heatmap. The value of each pixel in the heatmap refers to the probability that the pixel belongs to either landslide or non-landslide classes. Thus, we evaluate three scenarios: ResU-Net, OBIA, and ResU-Net-OBIA. The landslide detection maps from three different classification scenarios were compared against a manual landslide inventory map using thematic accuracy assessment metrics: precision, recall, and f1-score. Our experiments in the testing area showed that the proposed integration framework yields f1-score values 8 and 22 percentage points higher than those of the ResU-Net and OBIA approaches, respectively.

Landslide Susceptibility Mapping of Landslides with Artificial Neural Networks: Multi-Approach Analysis of Backpropagation Algorithm Applying the Neuralnet Package in Cuenca, Ecuador

“…A frequently applied method is artificial neural networks (ANNs), with several studies performed in different countries [7,[30][31][32][33][34][35]. In addition to the above, it is important to highlight other sophisticated ML methods for LSMs, including deep learning and object-based image analysis [36] and other specific deep learning applications such as automatic landslide detection [37,38] or convolutional neural networks for landslide detection [39]. The absence of a universally accepted method [30] allows investigating them in order to, with a suitable validation method, determine the ones that present a good performance.…”

The Outcome of the 2022 Landslide4Sense Competition: Advanced Landslide Detection From Multisource Satellite Imagery

“…A rulebased image classification based on several thresholds that vary depending on the study area distinguishes landslides from other features. In [16], [17], [27], [28], similar methodologies and rule-based segmentation and classification have been described. In the second step of our landslide annotation and image labeling procedure, a variety of further data sources were then used to visually correct landslide polygons, specifically Google Earth images, and previous landslide inventory data sets provided in a number of studies like [29], were used to enhance the first step's results.…”

Landslide4Sense: Reference Benchmark Data and Deep Learning Models for Landslide Detection

“…Also, including expert knowledge and user-defined parameters for determining the optimum parameters, thresholds for landslide segments in object-based methods help to provide higher accuracy than pixel-based methods [26]. However, current object-based methods often address mainly the low-level features of individual objects, which limits their generalization potential, since there is also a need to include high-level features such as object topologies [27]. Additionally, although creating handcrafted high-level features using expert knowledge is transparent and flexible compared to deep learning (DL) models, it is laborious, timeconsuming, and subjective [28].…”