人工智能算法在滑坡监测与预测技术中的研究与应用

程刚; 吴勇飞; 曹德胜; 吴亚熹

doi:10.19509/j.cnki.dzkq.tb20250115

您当前的位置：

首页 >

文章列表页 >

人工智能算法在滑坡监测与预测技术中的研究与应用

更新时间：2026-11-05

- 人工智能算法在滑坡监测与预测技术中的研究与应用
- Bulletin of Geological Science and Technology Vol. 44, Issue 5, Pages: 302-316(2025)
- 作者机构：
  
  1. 华北科技学院计算机科学与工程学院
  2. 华北科技学院河北省物联网监控技术创新中心
  3. 南京大学地球科学与工程学院
  4. 华北科技学院矿山安全学院
- 作者简介：
- 基金信息：
- DOI：10.19509/j.cnki.dzkq.tb20250115
  CLC： TP18;P642.22
- Online First：24 June 2025，
  
  Published：24 June 2025
- 稿件说明：
移动端阅览
程刚, 吴勇飞, 曹德胜, et al. 人工智能算法在滑坡监测与预测技术中的研究与应用[J]. Bulletin of Geological Science and Technology, 2025, 44(5): 302-316.
DOI：

程刚, 吴勇飞, 曹德胜, et al. 人工智能算法在滑坡监测与预测技术中的研究与应用[J]. Bulletin of Geological Science and Technology, 2025, 44(5): 302-316. DOI： 10.19509/j.cnki.dzkq.tb20250115.

摘要

为减轻滑坡灾害风险，进一步保障区域可持续发展，开展有效的滑坡监测与预测研究具有重要的现实意义。通过研究滑坡监测与预测中的关键技术与方法，分析各类算法在滑坡监测与预测场景中的效率和精度，不断提升滑坡灾害防治水平。在特征提取技术方面，对比分析了尺度不变特征变换（SIFT）、加速鲁棒特征（SURF）和自适应尺度不变特征变换（ASIFT）3种基于图像特征匹配算法的性能，其中ASIFT在匹配数量、精确率和召回率方面具有显著优势，尤其适用于准确性要求较高的复杂环境场景；在光流分析技术方面，探讨了基于Lucas-Kanade稀疏光流法和Horn-Schunck稠密光流法的应用效果，其中Lucas-Kanade稀疏光流法计算效率高，适合实时应用场景，但存在遗漏重要运动信息风险，Horn-Schunck稠密光流法能够提供全面的光流场信息，适用于环境复杂场景，但存在计算复杂度较高的不足，因而难以用于实时处理；在滑坡易发性预测方面，详细介绍了支持向量机（SVM）、决策树（DT）和随机森林（RF）等经典机器学习方法在滑坡预测中的应用优缺点，并重点研究了基于粒子群优化支持向量机（PSO-SVM）的模型性能，该模型通过优化超参数，显著提高了模型的分类准确度、泛化能力和预测精度。此外，通过引入Faster R-CNN模型，利用其先进的卷积神经网络架构，实现了复杂场景下滑坡事件的自动识别与分类，进一步提升了滑坡监测预警的效率和准确率。研究表明，ASIFT局部特征提取的精确率为0.84

Lucas-Kanade稀疏光流法的跟踪误差为0.12

PSO-SVM模型的均方根误差为0.52

Faster R-CNN模型在滑坡图像自动识别与分类方面的置信度可达0.98，综合性能较其他算法显著提升。综上所述，通过引入人工智能算法，结合多学科技术手段，全方面提升了滑坡监测与预测技术的效率和精度，研究成果为滑坡地质灾害防治提供了更有力的技术保障。

Abstract

ObjectiveTo reduce the risk of landslide disasters and further ensure regional sustainable development

conducting research effectively on landslide monitoring and prediction is of great practical significance. This paper aims to continuously improve the prevention and control level of landslide disasters by studying the key technologies and methods in landslide monitoring and prediction

and will analyze the efficiency and accuracy of various algorithms in scenarios of landslide monitoring and prediction. MethodsIn terms of feature extraction technology

the performance of three image feature matching algorithms

namely scale-invariant feature transform (SIFT)

speeded up robust features (SURF) and assessment of scale-invariant features transform (ASIFT)

were compared and analyzed. Comparative results demonstrate that ASIFT has significant advantages in the number of matches

precision rate and recall rate

and is especially suitable for complex environmental scenarios with high-accuracy requirements. In terms of optical flow analysis technology

the application effects based on the Lucas-Kanade sparse optical flow method and the Horn-Schunck dense optical flow method were discussed. Among them

the Lucas-Kanade sparse optical flow method has high computational efficiency and is suitable for real-time application scenarios

but there is a risk of missing important motion information. The Horn-Schunck dense optical flow method can provide comprehensive optical flow field information and is suitable for complex environmental scenarios. However

it has the drawback of high computational complexity and thus is difficult to be used in real-time processing. In terms of landslide susceptibility prediction

the advantages and disadvantages of the application of classic machine learning methods such as support vector machine (SVM)

decision tree (DT)

and random forest (RF) in landslide prediction are introduced in detail. The model performance based on particle swarm optimization-support vector machine (PSO-SVM) is mainly studied. This model optimizes hyperparameters

the classification accuracy

generalization ability and prediction accuracy of the model have been significantly improved. Furthermore

by introducing the Faster R-CNN model and leveraging its advanced convolutional neural network architecture

this paper realizes the automatic identification and classification of landslide events in complex scenarios

further enhancing the efficiency and accuracy of landslide monitoring and early warning. ResultsThe research shows that the accuracy rate of local feature extraction by ASIFT is 0.84

the tracking error of the Lucas-Kanade sparse optical flow method is 0.12

the root mean square error of the PSO-SVM model is 0.52

and the confidence level of the Faster R-CNN model in the automatic recognition and classification of landslide images can reach 0.98. The comprehensive performance is significantly improved compared with other algorithms in this paper. ConclusionIn summary

by introducing artificial intelligence algorithms and integrating multi-disciplinary technical means

this paper has comprehensively enhanced the efficiency and accuracy of landslide monitoring and prediction technology. The research results provide more powerful technical support for the prevention and control of landslide geological disasters.

关键词

Keywords

references

张慧慧.抚顺西露天矿大变形滑坡的多源协同监测方法研究[D].东北大学,2019(04).

黄玲.基于Django的山体滑坡监测系统设计与实现[D].贵州大学,2022(02).

石化波,王刚,曾怀恩.一种ICEEMDAN-CNN-SVR滑坡位移组合预测模型[J].三峡大学学报(自然科学版),2025(01).

马宁,李韶凯,田峰,叶霄,朱鸿鹄.滑坡光纤神经感测系统：技术与应用[J].地质科技通报,2024(06).

吴迪,梁泰铭,吴静,吴建建,易杨,娄万鹏.基于MEMS传感器的滑坡监测算法设计与试验[J].地质科技通报,2024(06).

蒲磊,杨启贵,程翔,周伟.基于SBAS-InSAR技术的HD库区CYP滑坡形变监测分析[J].中国农村水利水电,2024(11).

唐烽顺,郝利娜,宋雨洋,武德宏.目标检测算法在滑坡识别中的应用[J].兰州大学学报(自然科学版),2024(02).

尚硕,曹建荣,汪明,郑学汉,高鹤.机器视觉中角点检测算法研究[J].计算机测量与控制,2024(01).

朱鸿鹄.工程地质界面：从多元表征到演化机理[J].地质科技通报,2023(01).

窦杰,向子林,许强,郑鹏麟,王协康,苏爱军,刘军旗,罗万祺.机器学习在滑坡智能防灾减灾中的应用与发展趋势[J].地球科学,2023(05).

张抒,唐辉明,龚文平,邹宗兴.基于物理力学机制的滑坡数值预报模式:综述、挑战与机遇[J].地质科技通报,2022(06).

唐辉明.重大滑坡预测预报研究进展与展望[J].地质科技通报,2022(06).

程刚,王振雪,李刚强,倪钰菲,徐靓.我国滑坡监测文献计量研究的可视化分析[J].中国安全科学学报,2022(07).

简小婷,赵康,左小清,朱琪,朱文.基于Faster R-CNN目标检测的滑坡隐患识别——以福贡县城区为例[J].化工矿物与加工,2022(12).

赵玉兰,王兴伟,黄敏,姜春风.基于区域统计直方图与自适应规则的图像匹配算法[J].河南理工大学学报(自然科学版),2017(06).

曹兰柱,王珍,王东,王前领.露天矿滑坡预警理论与方法研究[J].中国安全科学学报,2017(03).

张俊,殷坤龙,王佳佳,黄发明.基于时间序列与PSO-SVR耦合模型的白水河滑坡位移预测研究[J].岩石力学与工程学报,2015(02).

贾丰蔓,康志忠,于鹏.影像同名点匹配的SIFT算法与贝叶斯抽样一致性检验[J].测绘学报,2013(06).

Rongjie He,Wengang Zhang,Jie Dou,Nan Jiang,Huaixian Xiao,Jiawen Zhou.Application of artificial intelligence in three aspects of landslide risk assessment: A comprehensive review[J].Rock Mechanics Bulletin,2024.

Ziyao Xu,Ailan Che,Hanxu Zhou.Seismic landslide susceptibility assessment using principal component analysis and support vector machine.[J].Scientific reports,2024.

Sun Chuanmeng,Xu Ruijia,Wang Chong,Ma Tiehua,Chen Jiaxin.Coal rock image recognition method based on improved CLBP and receptive field theory[J].Deep Underground Science and Engineering,2022.

Hakim Wahyu Luqmanul,Rezaie Fatemeh,Nur Arip Syaripudin,Panahi Mahdi,Khosravi Khabat,Lee Chang-Wook,Lee Saro.Convolutional neural network (CNN) with metaheuristic optimization algorithms for landslide susceptibility mapping in Icheon, South Korea[J].Journal of Environmental Management,2022.

Yin Chao,Li Haoran,Che Fa,Li Ying,Hu Zhinan,Liu Dong.Susceptibility mapping and zoning of highway landslide disasters in China.[J].PloS one,2020.

Adriaan L. van Natijne,Roderik C. Lindenbergh,Thom A. Bogaard.Machine Learning: New Potential for Local and Regional Deep-Seated Landslide Nowcasting[J].Sensors,2020.

Yi Wang,Zhice Fang,Haoyuan Hong.Comparison of convolutional neural networks for landslide susceptibility mapping in Yanshan County, China[J].Science of the Total Environment,2019.

Jie Dou,Ali P. Yunus,Dieu Tien Bui,Abdelaziz Merghadi,Mehebub Sahana,Zhongfan Zhu,Chi-Wen Chen,Khabat Khosravi,Yong Yang,Binh Thai Pham.Assessment of advanced random forest and decision tree algorithms for modeling rainfall-induced landslide susceptibility in the Izu-Oshima Volcanic Island, Japan[J].Science of the Total Environment,2019.

Beibei Yang,Kunlong Yin,Suzanne Lacasse,Zhongqiang Liu.Time series analysis and long short-term memory neural network to predict landslide displacement[J].Landslides,2019.

original> .Prediction of landslide displacement with step-like behavior based on multialgorithm optimization and a support vector regression model[J].Landslides,2018.

Chao Zhou,Kunlong Yin,Ying Cao,Bayes Ahmed.Application of time series analysis and PSO–SVM model in predicting the Bazimen landslide in the Three Gorges Reservoir, China[J].Engineering Geology,2016.

Jean-Michel Morel,Guoshen Yu.ASIFT: A New Framework for Fully Affine Invariant Image Comparison[J].SIAM Journal on Imaging Sciences,2009.

F. van den Bergh,A.P. Engelbrecht.A study of particle swarm optimization particle trajectories[J].Information Sciences,2005.

Alex J. Smola,Bernhard Schölkopf.A tutorial on support vector regression[J].Statistics and Computing,2004.

Hongche Liu,Tsai-Hong Hong,Martin Herman,Ted Camus,Rama Chellappa.Accuracy vs Efficiency Trade-offs in Optical Flow Algorithms[J].Computer Vision and Image Understanding,1998.

Corinna Cortes,Vladimir Vapnik.Support-Vector Networks.[J].Machine Learning,1995.

Berthold K.P. Horn,Brian G. Schunck.Determining optical flow[J].Artificial Intelligence,1981.

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

岩石突破压力与破裂压力分析及应用

降雨作用下鄂西山区麻池村1号滑坡变形破坏过程试验

强变形构造区多层含水系统承压水运移规律识别

基于迁移学习和多策略改进的山区公路落石目标检测

青海野马泉铁多金属矿床两期成岩成矿事件——来自锆石、石榴子石U-Pb定年和全岩地球化学的证据

Related Author

No data

Related Institution

No data

AI问答

⁰