Landslide detection from bitemporal satellite imagery using attention-based deep neural networks

Amankwah, Solomon Obiri Yeboah; Wang, Guojie; Gnyawali, Kaushal Raj; Hagan, Daniel Fiifi Tawia; Sarfo, Isaac; Dong, Zhen; Nooni, Isaac Kwesi; Ullah, Waheed; Duan, Zhenfeng

doi:10.1007/s10346-022-01915-6

Cited by 30 publications

(13 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although, in this model, a small area of landslide cannot be recognized after layer-wise convolution and semantic information is ignored. The improvement of a single module does not solve the problem of various types of landslides in the dataset well: the edge information of large landslides, the identification of small landslides, the interference of vegetation, bare soil, living areas and other environments, etc., all add difficulties to the accurate identification of landslides [51]. Therefore, in this paper, the three modules are fused, and Figure 9(Af-Df) shows the effect of the improved model recognition.…”

Section: Visualization Comparison Of Landslide Extraction Resultsmentioning

confidence: 99%

Enhance the Accuracy of Landslide Detection in UAV Images Using an Improved Mask R-CNN Model: A Case Study of Sanming, China

Zhang

Zheng

et al. 2023

Sensors

View full text Add to dashboard Cite

Extracting high-accuracy landslide areas using deep learning methods from high spatial resolution remote sensing images is a hot topic in current research. However, the existing deep learning algorithms are affected by background noise and landslide scale effects during the extraction process, leading to poor feature extraction effects. To address this issue, this paper proposes an improved mask regions-based convolutional neural network (Mask R-CNN) model to identify the landslide distribution in unmanned aerial vehicles (UAV) images. The improvement of the model mainly includes three aspects: (1) an attention mechanism of the convolutional block attention module (CBAM) is added to the backbone residual neural network (ResNet). (2) A bottom-up channel is added to the feature pyramidal network (FPN) module. (3) The region proposal network (RPN) is replaced by guided anchoring (GA-RPN). Sanming City, China was selected as the study area for the experiments. The experimental results show that the improved model has a recall of 91.4% and an accuracy of 92.6%, which is 12.9% and 10.9% higher than the original Mask R-CNN model, respectively, indicating that the improved model is more effective in landslide extraction.

show abstract

Section: Visualization Comparison Of Landslide Extraction Resultsmentioning

confidence: 99%

Enhance the Accuracy of Landslide Detection in UAV Images Using an Improved Mask R-CNN Model: A Case Study of Sanming, China

Zhang

Zheng

et al. 2023

Sensors

View full text Add to dashboard Cite

show abstract

“…Cheng et al (2021) introduced an attention module designed based on the visual system and incorporated it into the yolov4 model for training, which improved the attention to landslide features and reduced background noise. Amankwah et al (2022) introduced an attention module into the deep network structure to improve the ability to suppress background noise. The research results show that the attention module can significantly improve the landslide detection performance.…”

Section: Open Access Edited Bymentioning

confidence: 99%

A novel Dynahead-Yolo neural network for the detection of landslides with variable proportions using remote sensing images

Han

FANG²,

Li³

et al. 2023

Front. Earth Sci.

View full text Add to dashboard Cite

Efficient and automatic landslide detection solutions are beneficial for regional hazard mitigation. At present, scholars have carried out landslide detection based on deep learning. However, continuous improvement regarding the accuracy of landslide detection with better feature extraction of landslides remain an essential issue, especially small-proportion landslides in the remote sensing images are difficult to identify up to date. To address this issue, we propose a detection model, the so-called Dynahead-Yolo which is designed by combining unifying scale-aware, space-aware, and task-aware attention mechanisms into the YOLOv3 framework. The proposed method focuses on the detailed features of landslide images with variable proportions, improving the ability to decode landslides in complex background environments. We determine the most efficient cascade order of the three modules and compare previous detection networks based on randomly generated prediction sets from the three study areas. Compared with the traditional YOLOv3, the detection rate of Dynahead-Yolo in small-proportion landslides and complex background landslides is increased by 13.67% and 14.12%, respectively.

show abstract

“…Spaceborne remote sensing provides an alternative data source for large‐scale landslide assessment and has been rapidly developed in recent years in response to more widespread data availability and improved processing workflows (Casagli et al, 2023; Kirschbaum et al, 2019; Mantovani et al, 1996; Metternicht et al, 2005). Remotely sensed landslide mapping techniques range from visual interpretation of satellite images (e.g., Kincey et al, 2021) and automated change detection (Amankwah et al, 2022; Scheip & Wegmann, 2021), to mapping of topographic change (Bernard et al, 2021; Dai et al, 2020; Dille et al, 2021) and detection of surface displacements (Bickel et al, 2018; Dille et al, 2021; Lombardi et al, 2017; Manconi, 2021; Manconi et al, 2018; Rosi et al, 2018). Each remote‐sensing and ground‐based method is associated with different spatiotemporal resolutions, uncertainties and intrinsic limitations.…”

Section: Introductionmentioning

confidence: 99%

Detection of slow‐moving landslides through automated monitoring of surface deformation using Sentinel‐2 satellite imagery

Van Wyk de Vries,

Arrell,

Basyal

et al. 2024

Earth Surf Processes Landf

View full text Add to dashboard Cite

Landslides are one of the most damaging natural hazards and have killed tens of thousands of people around the world over the past decade. Slow‐moving landslides, with surface velocities on the order of 10−2–102 m a−1, can damage buildings and infrastructure and be precursors to catastrophic collapses. However, due to their slow rates of deformation and at times subtle geomorphic signatures, they are often overlooked in local and large‐scale hazard inventories. Here, we present a remote‐sensing workflow to automatically map slow‐moving landslides using feature tracking of freely and globally available optical satellite imagery. We evaluate this proof‐of‐concept workflow through three case studies from different environments: the extensively instrumented Slumgullion landslide in the United States, an unstable lateral moraine in Chilean Patagonia and a high‐relief landscape in central Nepal. This workflow is able to delineate known landslides and identify previously unknown areas of hillslope deformation, which we consider as candidate slow‐moving landslides. Improved mapping of the spatial distribution, character and surface displacement rates of slow‐moving landslides will improve our understanding of their role in the multi‐hazard chain and their sensitivity to climatic changes and can direct future detailed localised investigations into their dynamics.

show abstract

Landslide detection from bitemporal satellite imagery using attention-based deep neural networks

Cited by 30 publications

References 42 publications

Enhance the Accuracy of Landslide Detection in UAV Images Using an Improved Mask R-CNN Model: A Case Study of Sanming, China

Enhance the Accuracy of Landslide Detection in UAV Images Using an Improved Mask R-CNN Model: A Case Study of Sanming, China

A novel Dynahead-Yolo neural network for the detection of landslides with variable proportions using remote sensing images

Detection of slow‐moving landslides through automated monitoring of surface deformation using Sentinel‐2 satellite imagery

Contact Info

Product

Resources

About