Slope surface deformation monitoring plays an important role in landslide risk assessment and early warning. Currently, the mainstream GNSS, as a point-measurement technique, is expensive to deploy, resulting in information on only a few points of displacement being obtained on a target slope in practical applications. In contrast, optical images can contain more information on slope displacement at a much lower cost. Therefore, a low-cost, high-spatial-resolution and easy-to-implement landslide surface deformation monitoring system based on close-range photogrammetry is developed in this paper. The proposed system leverages multiple image processing methods and monocular visual localization, combined with machine learning, to ensure accurate monitoring under time series. The results of several laboratory landslide experiments show that the proposed system achieved millimeter-level monitoring accuracy in laboratory landslide experiments. Moreover, the proposed system could capture slow displacement precursors of 5 mm to 10 mm before significant landslide failure occurred, which provides favorable surface deformation evidence for landslide monitoring and early warning. In addition, the system was deployed on a natural slope in Lanzhou, yielding preliminary effective monitoring results. The laboratory experimental results demonstrated the system’s effectiveness and high accuracy in monitoring landslide surface deformation, particularly its significant application value in early warning. The field deployment results indicated that the system could also effectively provide data support in natural environments, offering practical evidence for landslide monitoring and warning.

Raw terrain data acquired by sensing techniques such as SfM or LiDAR typically contain non-terrain components that require filtering, such as vegetation occlusion and other non-terrain features. While filtering helps remove non-terrain data, it can introduce discontinuities and local voids in the dataset. These data gaps can affect both the completeness of the terrain representation and subsequent analysis tasks. Therefore, it is crucial to develop effective terrain data completion methods for reliable terrain analysis.Traditional terrain data completion methods, such as interpolation-based algorithms and Poisson surface reconstruction, typically model and optimize data continuity from a mathematical perspective. Although these methods address local voids to some extent, they generally fail to exploit terrain features and semantic information, limiting their effectiveness in completing complex terrain scenarios.To address these issues, we propose a deep learning-based framework for terrain data completion. Our methodology explores different neural network designs with supervised and unsupervised learning, incorporating geomorphological constraints to improve terrain feature representation and semantic understanding. The framework leverages the representational capabilities of deep learning to improve the robustness of terrain data completion, contributing to a more consistent and reliable basis for subsequent terrain analysis and applications.