Regional landslide identification is important for the risk management of landslide hazards. The traditional methods of regional landslide identification were mainly conducted by a human being. In previous studies, automatic landslide recognition mainly focused on new landslides distinct from the environment induced by rainfall or earthquake, using the image classification method and semantic segmentation method of deep learning. However, there is a lack of research on the automatic recognition of old loess landslides, which are difficult to distinguish from the environment. Therefore, this study uses the object detection method of deep learning to identify old loess landslides with Google Earth images. At first, a database of loess historical landslide samples was established for deep learning based on Google Earth images. A total of 6111 landslides were interpreted in three landslide areas in Gansu Province, China. Second, three object detection algorithms including the one-stage algorithm RetinaNet and YOLO v3 and the two-stage algorithm Mask R-CNN, were chosen for automatic landslide identification. Mask R-CNN achieved the greatest accuracy, with an AP of 18.9% and F1-score of 55.31%. Among the three landslide areas, the order of identification accuracy from high to low was Site 1, Site 2, and Site 3, with the F1-scores of 62.05%, 61.04% and 50.88%, respectively, which were positively related to their recognition difficulty. The research results proved that the object detection method can be employed for the automatic identification of loess landslides based on Google Earth images.
Interferometric synthetic aperture radar (InSAR) technology is known as one of the most effective methods for active landslide identification and deformation monitoring in large areas, and thus it is conducive to preventing and mitigating the losses caused by landslides. However, great uncertainty inevitably exists due to influences of complex terrains, dense vegetations, and atmospheric interferences in the southwestern mountainous area of China, and this is associated with false or erroneous judgment during the process of landslide identification. In this study, a landslide identification method is put forward by integrating InSAR technology and airborne light detection and ranging (LiDAR) technology. Via this method, surface deformation characteristics detected by InSAR technology and micro-geomorphic features reflected by LiDAR technology were used to identify and map landslides of large areas. Herein, the method was applied to process 224 Sentinel-1 images covering Danba County and its surrounding areas (540 km2) from October 2014 to September 2020. Firstly, 44 active landslides with total areas of 59 km2 were detected by stacking InSAR technology. Then, major regions up to 135 km2 were validated by data gained from the airborne LiDAR technology. Particularly, several large landslides with lengths and/or widths of more than 2 km were found. Further, the precipitation data were integrated with the above results to analyze the temporal deformation characteristics of three typical landslides from major regions via SBAS InSAR technology. The key findings were as follows: (1) The combination of InSAR and LiDAR technologies could improve the accuracy of landslide detection and identification; (2) there was a significant correlation between temporal deformation characteristics of some landslides and monthly rainfall, with an obvious hysteretic effect existing between the initiation timing of rainfall and that of deformation; (3) the results of this study will be important guidance for the prevention and control of geological hazards in Danba County and areas with similar complex geomorphological conditions by helping effectively identify and map landslides.
Landslides and collapses are common geological hazards in mountainous areas, posing significant threats to the lives and property of residents. Therefore, early identification of disasters is of great significance for disaster prevention. In this study, we used Small Baseline Subset Interferometric Synthetic Aperture Radar (SBAS-InSAR) technology to process C-band Sentinel-1A images to monitor the surface deformation from Songpinggou to Feihong in Maoxian County, Sichuan Province. Visibility analysis was used to remove the influence of geometric distortion on the SAR images and retain deformation information in the visible area. Hot spot and kernel density analyses were performed on the deformation data, and 18 deformation clusters were obtained. Velocity and slope data were integrated, and 26 disaster areas were interpreted from the 18 deformation clusters, including 20 potential landslides and 6 potential collapses. A detailed field investigation indicated that potential landslides No. 6 and No. 8 had developed cracks and were severely damaged, with a high probability of occurrence. Potential collapse No. 22 had developed fissures, exposing a dangerous rock mass and posing significant threats to the lives and property of residents. This study shows that the proposed method that combines visibility analysis, InSAR deformation rates, and spatial analysis can quickly and accurately identify potential geological disasters and provide guidance for local disaster prevention and mitigation.
The land reclaimed from the seaside may have a long-term subsidence trend, which poses a potential geohazard in the future land use. Xiamen Xiang’an New Airport (XXNA) is built on reclaimed land since 2016. Based on the spaceborne Sentinel-1 data between January 2018 to April 2019 and the time series interferometric synthetic aperture radar (InSAR) technique, this paper analyzed the reclaimed land subsidence evolution at XXNA in this period. InSAR measurements show that XXNA is suffering from severe subsidence, mainly in three regions because of the earth and sand compacting. By analyzing the spatial subsidence characterizations of the main subsiding areas combined with historical land reclamation and future land use planning, we find the potential threat of subsidence to future land use. Correlation between subsidence and the period of reclamation was found, indicating that the consolidation and compression in dredger fill is the main cause of subsidence. By combining subsidence monitoring results with different land use types and adopting the Expectation (Ex) and Entropy (En) methods, we analyzed the key area with potential subsidence geo-hazard. This work shows that with SAR interferometry, it is possible to find the large area ground subsidence in the airport reclaimed area. The areas with potential subsidence geo-hazards are consistent with the deep reclaimed earth, which means high subsidence risk in the future.
Dalian Jinzhou Bay International Airport (DJBIA) is an offshore artificial island airport, where the reclaimed land is prone to uneven land subsidence due to filling consolidation and construction. Monitoring and predicting the subsidence are essential to assist the subsequent subsidence control and ensure the operational safety of DJBIA. However, the accurate monitoring and prediction of reclaimed subsidence for such a wide area under construction are hard and challenging. This paper utilized the Small Baseline Subset Synthetic Aperture Radar (SBAS-InSAR) technology based on Sentinel-1 images from 2017 to 2021 to obtain the subsidence over the land reclamation area of the DJBIA, in which the results from ascending and descending orbit data were compared to verify the reliability of the results. The SBAS-InSAR results reveal that uneven subsidence is continuously occurring, especially on the runway, terminal, and building area of the airport, with the maximum subsidence rate exceeding 100 mm/year. It was found that there is a strong correlation between the subsidence rate and backfilling time. This study provides important information on the reclaimed subsidence for DJBIA and demonstrates a novel method for reclaimed subsidence monitoring and prediction by integrating the advanced InSAR technology and Terzaghi Consolidation Theory modeling. Moreover, based on the Terzaghi consolidation theory and the corresponding geological parameters of the airport, predicted subsidence curves in this area are derived. The comparison between predicted curves and the actual subsidence revealed by InSAR in 2017–2021 is highly consistent, with a similar trend and falling in a range of ±25 mm/year, which verifies that the subsidence in this area conforms to Terzaghi Consolidation Theory. Therefore, it can be predicted that in the future, the subsidence rate of the new reclamation area in this region will reach about 80 mm/year ± 25 mm/year, and the subsidence rate will gradually slow down with the accumulation of reclamation time. The subsidence rate will slow down to about 30 mm/year ± 25 mm/year after 10 years.
Beijing Capital International Airport (BCIA) has suffered from uneven land subsidence since 1935, which affects the smoothness of airport runways and seriously threatens the safety of aircrafts. In this paper, a spaceborne interferometric synthetic aperture radar (InSAR) with high-resolution Cosmo-SkyMed SAR data was utilized at BCIA for the first time to diagnose the subsidence hazard. The results show that subsidence is progressing at BCIA at a maximum rate of 50 mm/year, which is mainly distributed in the northwest side of the airport. It was found that the Shunyi-Liangxiang fault directly traverses Runway2 and Runway3 and causes uneven subsidence, controlling the spatial subsidence pattern to some degree. Four driving factors of subsidence were investigated, namely: the over-exploitation of groundwater, active faults, compressible soil thickness, and aquifer types. For the future sustainable development of BCIA, the influence of Beijing new airport and Beijing Daxing International Airport (BDIA), was analyzed and predicted. It is necessary to take relevant measures to control the uneven subsidence during the initial operation of BDIA and conduct long-term monitoring to ensure the regular safe operation of BCIA. This case demonstrates a remote sensing method of diagnosing the subsidence hazard with high accuracy and non-contact, providing a reliable alternative for the geohazard diagnosis of key infrastructures in the future.