滑坡是一种最为常见且威胁极大的地质灾害，在我国，由于其复杂的地质构造和环境特征，滑坡灾害发生的频率和程度更为严重，这些灾害对当地社会、经济和人民生活造成了巨大的影响。目前已有的滑坡监测方法中比较常用的有GNSS技术、三维激光扫描技术、合成孔径雷达技术等，但其数据获取成本高。为实现低成本滑坡自动化监测，本文引入无控制点的固定位置摄影测量技术实现了滑坡自动化监测流程，包括滑坡图像采集、图像质量评估、天气图像识别、监测图像优化、生成三维模型、对滑坡多期密集点云配准以及点云比较，最终得到滑坡灾害区域。本文的主要研究工作如下:

（1）针对影响三维重建精度的天气因素进行分析并引入相应的图像优化算法优化图像质量，且为实现自动化滑坡监测引入基于迁移学习的天气识别算法。首先分析气候因素即雨天、雾天图像对三维重建的影响，通过比较清晰场景与有雾、有雨场景三维重建结果，证明了有雾有雨均会对三维重建有影响，然后在图像质量评估的基础上利用基于迁移学习的天气识别方法对所获取到的图像进行识别，最后针对不同的气候因素根据相应的图像优化方法对监测图像优化，并利用优化后的图像进行三维重建，结果表明经过暗通道先验去雾的去雾图像与基于深度学习网络模型的去雨图像都可以成功重建出三维场景信息。

（2）设计了一种改进迭代最近点的异尺度点云配准算法。根据无控制点的固定位置摄影测量技术得到连续点云的相似性，设计一种改进迭代最近点的异尺度点云配准算法，通过对一组标准点云数据利用迭代最近点算法、相干点漂移算法、尺度配准算法以及本文算法进行配准比较，表明本文所设计的改进迭代最近点的异尺度点云配准算法可以配准尺度不一的相似点云，并通过一组实例数据利用本文算法进行验证，配准后的两个点云均方根误差为0.176cm，进一步说明本文算法可用于异尺度相似点云配准。

（3）引入基于多尺度模型到模型点云比较算法识别滑坡灾害。本文首先利用一组室内实验对变化前后的两组点云使用多尺度模型到模型点云比较结果，证明了该方法的可行性与有效性，然后通过一组模拟滑坡变化前后点云数据根据此方法自动识别出滑坡灾害区域。

（4）实现滑坡灾害自动化监测的流程。使用Python语言设计了滑坡自动化监测程序，在没有用户干预的情况下，通过设计的程序可以实现实时获取图像并在几分钟后得到滑坡变形区域。首先通过一组室内试验证明该流程的有效性，在提前设置好图像蒙版以及参考点云的前提下，可在无用户干预的情况下进行图像质量评估、三维重建、异尺度点云配准以及多尺度模型到模型点云比较后得到变化区域，然后将此流程应用到王家坡滑坡工程实例中，通过采集到的两期数据实现了实验区域的监测。

Landslides are one of the most common and threatening geological hazards. In China, due to its complex geological structure and environmental features, landslide hazards occur more frequently and to a greater extent, and these hazards have a huge impact on local society, economy and people's lives. The more commonly used landslide monitoring methods are GNSS technology, 3D laser scanning technology and synthetic aperture radar technology, but their data acquisition costs are high. In order to achieve low-cost landslide automatic monitoring, this paper introduces fixed-position photogrammetry without control points to realize the landslide automatic monitoring process, including landslide image acquisition, image quality assessment, weather image recognition, monitoring image optimization, generating 3D models, aligning multi-period dense point clouds of landslides and point cloud comparison, and finally obtaining landslide hazard areas. The main research work of this paper is as follows.

(1) The weather factors that affect the accuracy of 3D reconstruction are analysed and corresponding image optimisation algorithms are introduced to optimise image quality, and a migration learning-based weather recognition algorithm is introduced for automated landslide monitoring. By comparing the 3D reconstruction results of clear scenes with those of foggy and rainy scenes, it is proved that both foggy and rainy scenes have an impact on the 3D reconstruction, and then the images are identified using the migration learning-based weather recognition method on the basis of image quality assessment. The results show that both the fogged image and the rain image can be successfully reconstructed based on the deep learning network model.

(2) A heteroscale point cloud alignment algorithm with improved iterative nearest point is designed. Based on the similarity of continuous point clouds obtained from fixed position photogrammetry without control points, an improved iterative nearest point heteroscale point cloud alignment algorithm is designed and compared with a set of standard point cloud data using the iterative nearest point algorithm, the coherent point drift algorithm, the scale alignment algorithm and the algorithm in this paper, showing that the improved iterative nearest point heteroscale point cloud alignment algorithm designed in this paper can align similar point clouds with different scales. It is also verified by a set of example data using the algorithm in this paper that the root mean square error of the aligned two point clouds is 0.176cm, further demonstrating that the algorithm in this paper can be used to align similar point clouds at different scales.

(3) Introduction of a multi-scale model-to-model comparison algorithm based on landslide hazard identification. This paper first uses a set of indoor experiments to compare the results of two sets of point clouds before and after the change using a multi-scale model-to-model point cloud comparison to demonstrate the feasibility and effectiveness of the method, and then automatically identifies landslide hazard areas based on this method through a set of simulated point cloud data before and after the landslide change.

(4) Implementing a process for automated landslide hazard monitoring. Using the Python language to design an automated landslide monitoring process, the designed process enables real-time acquisition of images and obtains landslide deformation areas after a few minutes without user intervention. The effectiveness of the process is first demonstrated through a set of indoor tests, where image quality assessment, 3D reconstruction, heteroscale point cloud alignment and multi-scale model-to-model comparison can be carried out without user intervention, provided that image masks and reference point clouds are set in advance, and then the process is applied to the Wangjiapo landslide project example, where the experimental area is realised through the data collected in two phases The monitoring of the experimental area was achieved through the data collected in two phases.

[1] 许强, 朱星, 李为乐, 等. “天-空-地”协同滑坡监测技术进展[J]. 测绘学报, 2022, 51(7): 1416–1436.

[2] 毛伊敏, 张茂省, 李林, 等. 不确定密度聚类分析算法的滑坡危险性评价[J]. 自然灾害学报, 2019, 28(1): 139–148.

[3] 张鹏飞. GNSS载波相位时间传递关键技术与方法研究[J]. 测绘学报, 2020, 49(5): 669.

[4] 朱真, 江思义, 刘小明, 等. 基于广播RTK边缘计算的北斗高精度地质灾害监测系统及应用分析[J]. 水文地质工程地质, 2021, 48(5): 176–183.

[5] 张云龙. 北斗双天线高精度变形监测关键技术及其应用研究[D]. 北京:北京交通大学, 2018.

[6] 李得军. 采动区高等级公路三维激光扫描监测及变形信息提取[D]. 徐州:中国矿业大学, 2019.

[7] Li L, Lan H, Peng J. Loess erosion patterns on a cut-slope revealed by LiDAR scanning[J]. Engineering Geology, 2020, 268: 105516.

[8] Yin C, Li H, Hu Z, et al. Application of the Terrestrial Laser Scanning in Slope Deformation Monitoring: Taking a Highway Slope as an Example[J]. Applied Sciences, Multidisciplinary Digital Publishing Institute, 2020, 10(8): 2808.

[9] 程刚, 王振雪, 李刚强, 等. 滑坡一体化全维度监测预警技术进展与展望[J]. 中国安全科学学报, 2023, 33(1): 105–114.

[10] 薛现凯. 基于无人机倾斜摄影的边坡三维重建与灾害识别[D]. 长沙:湖南大学, 2021.

[11] 刘昌华, 王成龙, 李峰, 等. 数字近景摄影测量在山地矿区变形监测中的应用[J]. 测绘科学, 2009, 34(4): 197–199.

[12] 查燕萍, 张华平. 数字摄影测量发展现状与趋势初探[J]. 江西测绘, 2009(4): 4–6.

[13] 程效军, 朱东丽, 解洪伟. 近景摄影测量技术在广东大峡谷工程中的应用[J]. 地矿测绘, 2010, 26(3): 10–12.

[14] 项鑫, 王艳利. 近景摄影测量在边坡变形监测中的应用[J]. 中国煤炭地质, 2010, 22(6): 66–69.

[15] 李彩林, 张剑清, 郭宝云. 利用近景摄影测量技术的滑坡监测新方法[J]. 计算机工程与应用, 2011, 47(3): 6–8.

[16] 李春梅, 张连蓬. 采石场边坡多基线数字近景摄影测量精度的影响因素分析[J]. 测绘通报, 2011(10): 23–26.

[17] 李天子, 郭辉. 多基线近景摄影测量的平面地表变形监测[J]. 辽宁工程技术大学学报(自然科学版), 2013, 32(8): 1098–1102.

[18] 周弈, 李浩, 王田芳, 等. 露采矿山近景影像光束法三维建模及动态监测系统[J]. 测绘工程, 2016, 25(2): 61–65.

[19] 刘志奇, 李天子, 刘昌华, 等. 基于单像近景摄影测量的滑坡裂缝探测方法[J]. 金属矿山, 2018(8): 108–113.

[20] 杨松勇. 近景摄影测量技术在露天矿边坡变形监测中的研究[D]. 赣州:江西理工大学, 2019.

[21] Fraser C S, Riedel B. Monitoring the thermal deformation of steel beams via vision metrology[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2000, 55(4): 268–276.

[22] Ohnishi Y, Nishiyama S, Yano T, et al. A study of the application of digital photogrammetry to slope monitoring systems[J]. International Journal of Rock Mechanics and Mining Sciences, 2006, 43(5): 756–766.

[23] Ordóñez C, Martínez J, Arias P, et al. Measuring building façades with a low-cost close-range photogrammetry system[J]. Automation in Construction, 2010, 19(6): 742–749.

[24] Matori A N, Cahyono B K, Mokhtar M R M, et al. Close-range photogrammetric data for landslide monitoring on slope area[A]. 2012 IEEE Colloquium on Humanities, Science and Engineering (CHUSER)[C]. 2012: 398–402.

[25] Akca D. Photogrammetric Monitoring of an Artificially Generated Shallow Landslide[J]. The Photogrammetric Record, 2013, 28(142): 178–195.

[26] Kaufmann V, Dorn M, Patri G-P S, et al. Digital Camera Nikon D300 in Support of High Mountain Studies in the Langtang Valley, Central Himalaya, Nepal[J]. Universal Journal of Geoscience, 2013, 1(1): 1–9.

[27] Kim D H, Gratchev I, Berends J, et al. Calibration of restitution coefficients using rockfall simulations based on 3D photogrammetry model: a case study[J]. Natural Hazards, 2015, 78(3): 1931–1946.

[28] Alameda-Hernández P, El Hamdouni R, Irigaray C, et al. Weak foliated rock slope stability analysis with ultra-close-range terrestrial digital photogrammetry[J]. Bulletin of Engineering Geology and the Environment, 2019, 78(2): 1157–1171.

[29] Kromer R, Walton G, Gray B, et al. Development and Optimization of an Automated Fixed-Location Time Lapse Photogrammetric Rock Slope Monitoring System[J]. Remote Sensing, Multidisciplinary Digital Publishing Institute, 2019, 11(16): 1890.

[30] Roncella R, Forlani G. A Fixed Terrestrial Photogrammetric System for Landslide Monitoring[A]. M. Scaioni. Modern Technologies for Landslide Monitoring and Prediction[M]. Berlin, Heidelberg: Springer, 2015: 43–67.

[31] Eltner A, Kaiser A, Abellan A, et al. Time lapse structure-from-motion photogrammetry for continuous geomorphic monitoring[J]. Earth Surface Processes and Landforms, 2017, 42(14): 2240–2253.

[32] Zhan Z, Zhou G, Yang X. A Method of Hierarchical Image Retrieval for Real-Time Photogrammetry Based on Multiple Features[J]. IEEE Access, 2020, 8: 21524–21533.

[33] Micheletti N, Chandler J H, Lane S N. Investigating the geomorphological potential of freely available and accessible structure-from-motion photogrammetry using a smartphone[J]. Earth Surface Processes and Landforms, 2015, 40(4): 473–486.

[34] Roncella R, Forlani G, Fornari M, et al. Landslide monitoring by fixed-base terrestrial stereo-photogrammetry[J]. ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Copernicus GmbH, 2014, II–5: 297–304.

[35] Mallalieu J, Carrivick J L, Quincey D J, et al. An integrated Structure-from-Motion and time-lapse technique for quantifying ice-margin dynamics[J]. Journal of Glaciology, Cambridge University Press, 2017, 63(242): 937–949.

[36] Parente L, Chandler J H, Dixon N. Optimising the quality of an SfM-MVS slope monitoring system using fixed cameras[J]. The Photogrammetric Record, 2019, 34(168): 408–427.

[37] Blanch X, Eltner A, Guinau M, et al. Multi-Epoch and Multi-Imagery (MEMI) Photogrammetric Workflow for Enhanced Change Detection Using Time-Lapse Cameras[J]. Remote Sensing, Multidisciplinary Digital Publishing Institute, 2021, 13(8): 1460.

[38] 王果, 蒋瑞波, 肖海红, 等. 基于无人机倾斜摄影的露天矿边坡三维重建[J]. 中国矿业, 2017, 26(4): 158–161.

[39] 黄海宁, 黄健, 周春宏, 等. 无人机影像在高陡边坡危岩体调查中的应用[J]. 水文地质工程地质, 2019, 46(6): 149–155.

[40] 胡胜. 黄土高原滑坡空间格局及其对地貌演化的影响[D]. 西安:西北大学, 2019.

[41] Clapuyt F, Vanacker V, Schlunegger F, et al. Unravelling earth flow dynamics with 3-D time series derived from UAV-SfM models[J]. Earth Surface Dynamics, 2017, 5(4): 791–806.

[42] Seier G, Sulzer W, Lindbichler P, et al. Contribution of UAS to the monitoring at the Lärchberg-Galgenwald landslide (Austria)[J]. International Journal of Remote Sensing, 2018, 39(15–16): 5522–5549.

[43] Peppa M V. Morphology-based landslide monitoring with an unmanned aerial vehicle[D]. Newcstle University, 2018.

[44] Eltner A, Schneider D. Analysis of Different Methods for 3D Reconstruction of Natural Surfaces from Parallel-Axes UAV Images[J]. The Photogrammetric Record, 2015, 30(151): 279–299.

[45] Imaizumi F, Masui T, Yokota Y, et al. Initiation and runout characteristics of debris flow surges in Ohya landslide scar, Japan[J]. Geomorphology, 2019, 339: 58–69.

[46] 冯林, 李斌兵. 利用无人机倾斜影像与GCP构建高精度侵蚀沟地形模型[J]. 农业工程学报, 2018, 34(3): 88–95.

[47] Stumpf A, Malet J-P, Delacourt C. Correlation of satellite image time-series for the detection and monitoring of slow-moving landslides[J]. Remote Sensing of Environment, 2017, 189: 40–55.

[48] Messerli A, Grinsted A. Image georectification and feature tracking toolbox: ImGRAFT[J]. Geoscientific Instrumentation, Methods and Data Systems, Copernicus GmbH, 2015, 4(1): 23–34.

[49] Turner D, Lucieer A, De Jong S M. Time Series Analysis of Landslide Dynamics Using an Unmanned Aerial Vehicle (UAV)[J]. Remote Sensing, Multidisciplinary Digital Publishing Institute, 2015, 7(2): 1736–1757.

[50] Lucieer A, Jong S M de, Turner D. Mapping landslide displacements using Structure from Motion (SfM) and image correlation of multi-temporal UAV photography[J]. Progress in Physical Geography: Earth and Environment, 2014, 38(1): 97–116.

[51] Lague D, Brodu N, Leroux J. Accurate 3D comparison of complex topography with terrestrial laser scanner: Application to the Rangitikei canyon (N-Z)[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2013, 82: 10–26.

[52] James M R, Quinton J N. Ultra-rapid topographic surveying for complex environments: the hand-held mobile laser scanner (HMLS)[J]. Earth Surface Processes and Landforms, 2014, 39(1): 138–142.

[53] Smith M W, Vericat D. From experimental plots to experimental landscapes: topography, erosion and deposition in sub-humid badlands from Structure-from-Motion photogrammetry[J]. Earth Surface Processes and Landforms, 2015, 40(12): 1656–1671.

[54] 张祖勋, 张剑清. 数字摄影测量学[M]. 武汉大学出版社, 1997.

[55] 张剑清, 潘励, 王树根. 摄影测量学[M]. 武汉大学出版社, 2009.

[56] 赵明宇. 基于无人机摄影测量的岩体结构信息获取[D]. 长春:吉林大学, 2020.

[57] 毕瑞, 甘淑, 袁希平, 等. 复杂地貌无人机遥感3D场景构建[J]. 山地学报, 2022, 40(1): 151–164.

[58] Lowe D G. Distinctive Image Features from Scale-Invariant Keypoints[J]. International Journal of Computer Vision, 2004, 60(2): 91–110.

[59] Furukawa Y, Ponce J. Accurate, Dense, and Robust Multiview Stereopsis[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2010, 32(8): 1362–1376.

[60] Mariotti A, Ruti P M, Rixen M. Progress in subseasonal to seasonal prediction through a joint weather and climate community effort[J]. Climate and Atmospheric Science, Nature Publishing Group, 2018, 1(1): 1–4.

[61] Helstrom C W. Image Restoration by the Method of Least Squares[J]. JOSA, Optica Publishing Group, 1967, 57(3): 297–303.

[62] Wu Z, Chen H, Lei Y. Recognizing Non-Collaborative Radio Station Communication Behaviors Using an Ameliorated LeNet[J]. Sensors, Multidisciplinary Digital Publishing Institute, 2020, 20(15): 4320.

[63] He M, Zhao X, Lu Y, et al. An improved AlexNet model for automated skeletal maturity assessment using hand X-ray images[J]. Future Generation Computer Systems, 2021, 121: 106–113.

[64] Qiu J, Lu X, Wang X, et al. Research on Rice Disease Identification Model Based on Migration Learning in VGG Network[J]. IOP Conference Series: Earth and Environmental Science, IOP Publishing, 2021, 680(1): 012087.

[65] Huang X, Chen W, Yang W. Improved Algorithm Based on The Deep Integration of Googlenet and Residual Neural Network[J]. Journal of Physics: Conference Series, IOP Publishing, 2021, 1757(1): 012069.

[66] 郭心悦, 胡沁涵, 刘纯平, 等. 基于迁移学习和批归一化的菜肴图像识别方法[J]. 计算机应用与软件, 2021, 38(3): 124–133.

[67] 朱劲松, 李欢, 王世芳. 基于卷积神经网络和迁移学习的钢桥病害识别[J]. 长安大学学报(自然科学版), 2021, 41(3): 52–63.

[68] He K, Sun J, Tang X. Single Image Haze Removal Using Dark Channel Prior[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(12): 2341–2353.

[69] 刘海波, 杨杰, 吴正平, 等. 基于暗通道先验和Retinex理论的快速单幅图像去雾方法[J]. 自动化学报, 2015, 41(7): 1264–1273.

[70] McCartney E J. Optics of the atmosphere: Scattering by molecules and particles[M]. New York, 1976.

[71] He K, Sun J, Tang X. Guided Image Filtering[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013, 35(6): 1397–1409.

[72] Fu X, Huang J, Zeng D, et al. Removing Rain from Single Images via a Deep Detail Network[A]. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)[C]. Honolulu, HI: IEEE, 2017: 1715–1723.

[73] Krizhevsky A, Sutskever I, Hinton G E. ImageNet classification with deep convolutional neural networks[J]. Communications of the ACM, 2017, 60(6): 84–90.

[74] He K, Zhang X, Ren S, et al. Deep Residual Learning for Image Recognition[J]. arXiv, 2015.

[75] Luo Y, Xu Y, Ji H. Removing Rain from a Single Image via Discriminative Sparse Coding[A]. 2015 IEEE International Conference on Computer Vision (ICCV)[C]. 2015: 3397–3405.

[76] 胡志刚, 秦启飞. 基于凸包的最小体积有向包围盒生成算法[J]. 湖南大学学报(自然科学版), 2019, 46(2): 105–111.

[77] 汪威, 张开颜, 刘亚川, 等. 三维点云的两步校准法及其应用研究[J]. 工程设计学报, 2020, 27(5): 560–567.

[78] Zhong Y. Intrinsic shape signatures: A shape descriptor for 3D object recognition[A]. 2009 IEEE 12th International Conference on Computer Vision Workshops, ICCV Workshops[C]. Kyoto, Japan: IEEE, 2009: 689–696.

[79] Besl P J, McKay N D. A method for registration of 3-D shapes[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1992, 14(2): 239–256.

[80] Cao Y, Lu Y, Pan X, et al. An improved global best guided artificial bee colony algorithm for continuous optimization problems[J]. Cluster Computing, 2019, 22(2): 3011–3019.

[81] Zhu G, Kwong S. Gbest-guided artificial bee colony algorithm for numerical function optimization[J]. Applied Mathematics and Computation, 2010, 217(7): 3166–3173.

[82] Karaboga D, Basturk B. On the performance of artificial bee colony (ABC) algorithm[J]. Applied Soft Computing, 2008, 8(1): 687–697.

[83] Ying S, Peng J, Du S, et al. A Scale Stretch Method Based on ICP for 3D Data Registration[J]. IEEE Transactions on Automation Science and Engineering, 2009, 6(3): 559–565.

[84] Schölkopf B, Platt J, Hofmann T. Non-rigid point set registration: Coherent Point Drift[A]. Advances in Neural Information Processing Systems 19: Proceedings of the 2006 Conference[M]. MIT Press, 2007: 1009–1016.