Efficient and accurate road extraction from remote sensing imagery is important for applications related to navigation and Geographic Information System updating. Existing data-driven methods based on semantic segmentation recognize roads from images pixel by pixel, which generally uses only local spatial information and causes issues of discontinuous extraction and jagged boundary recognition. To address these problems, we propose a cascaded attention-enhanced architecture to extract boundary-refined roads from remote sensing images. Our proposed architecture uses spatial attention residual blocks on multi-scale features to capture long-distance relations and introduce channel attention layers to optimize the multi-scale features fusion. Furthermore, a lightweight encoder-decoder network is connected to adaptively optimize the boundaries of the extracted roads. Our experiments showed that the proposed method outperformed existing methods and achieved state-of-the-art results on the Massachusetts dataset. In addition, our method achieved competitive results on more recent benchmark datasets, e.g., the DeepGlobe and the Huawei Cloud road extraction challenge.

Landslides recurrently cause severe damage and, in some cases, the full disruption of many highways in mountainous areas, which can last from a few days to even months. Thus, there is a high demand for monitoring tools and precipitation data to support highway alignment selections before construction. In this study, we proposed a new system highway alignment selection method based on coherent scatter InSAR (CSI) and ~1 km high-spatial-resolution precipitation (HSRP) analysis. Prior to the CSI, we calculated and analyzed the feasibility of Sentinel-1A ascending and descending data. To illustrate the performance of the CSI, CSI and SBAS–InSAR were both utilized to monitor 80 slow-moving landslides, which were identified by optical remote-sensing interpretation and field investigation, along the Barkam–Kangting Highway Corridor (BKHC) in southwestern China, relying on 56 Sentinel-1A descending images from September 2019 to September 2021. The results reveal that CSI has clearer deformation signals and more measurement points (MPs) than SBAS-InSAR. And the maximum cumulative displacements and rates of the landslides reach −75 mm and −64 mm/year within the monitoring period (CSI results), respectively. Furthermore, the rates of the landslides near the Jinchuan River are higher than those of the landslides far from the river. Subsequently, to optimize the highway alignment selection, we analyzed the spatiotemporal evolution characteristics of feature points on a typical landslide by combining the −1 km HSRP, which was calculated from the 30′ Climatic Research Unit (CRU) time-series datasets, with the climatology datasets of WorldClim using delta spatial downscaling. The analysis shows that the sliding rates of landslides augment from the back edge to the tongue because of fluvial erosion and that accelerated sliding is highly related to the intense precipitation between April and September each year (ASP). Consequently, three solution types were established in our method by setting thresholds for the deformation rates and ASPs of every landslide. Afterward, the risk-optimal alignment selection of the BKHC was finalized according to the solution types and consideration of the construction’s possible impacts. Ultimately, the major problems and challenges for our method were discussed, and conclusions were given.

Saline soil harbors crystallized salt in the solid phase and salt solution in the liquid phase. By the natural environmental and human activity factors, migration and accumulation of the salt and crystallization process alternated with dissolving, which makes saline soil with significant seasonal differences appear in the deformation and regional scattering characteristic. These phenomena raise the limitation of conventional multi-temporal interferometric synthetic aperture radar (MT-InSAR) technique and consequently high-precision deformation monitoring of strong saline soil subgrade and pavement. To overcome the limitations, this study aimed to propose an advanced MT-InSAR method, which considers the seasonal interferometric coherences caused by precipitation and the temporal physical deformation evolution of the subgrade and pavement over strong saline soil. To present the better performance of the advanced method, a segment of the Qarhan–Golmud Expressway (QGE), which is the first expressway built in the strong saline soil area in China, was selected for this study. Two sets of SAR images acquired from January 2018 to January 2022 from Sentinel-1A ascending and descending orbits over the QGE were utilized, 774 and 928 seasonal filtered interferograms are optimized for deformation monitoring based on the deformation Poisson curve (PC) model. Compared with the method in the previous studies, the accuracy, efficiency, and reliability of the monitoring results increased dramatically. Subsequently, further discussions are conducted in detail on the regularity of expressway deformation over strong saline soil, especially from physical and chemical perspectives. Findings show that the ratio between Cl⁻ and SO₄^2- determines the deformation distribution and morphology. Also, the precipitation and temperature affect the seasonal characteristics. The contributions of this investigation might provide technical references for related expressway management and policy-making departments to ensure the long-term safe operation and stability of the QGE.