摘要
为应对滑坡灾害突发性高、识别预测能力弱等问题,本文提出了一种基于无人机摄影测量的滑坡灾害识别与监测方法。首先,利用无人机快速获取滑坡体周围的高分辨率影像,构建影像数据集;然后,采取深度学习理论,选取CaffeNet网络模型,快速准确识别滑坡体范围;最后,结合据摄影测量理论,利用不同时期的无人机航拍影像制作DEM、DOM数字成果,根据特征断面和土方量变化预判出滑坡灾害的再发生程度。实验结果表明,采用无人机摄影测量进行滑坡灾害识别与监测可行性较高,灾害识别的精确度高,能满足地质灾害预警保障需求。
To address the issues of high suddenness and weak identification and prediction capabilities of landslide disasters,this paper proposes a landslide disaster identification and monitoring method based on unmanned aerial vehicle photogrammetry.Firstly,we use UAVs to quickly obtain high-resolution images of the landslide area and construct an image dataset;Then,using deep learning theory and selecting the CaffeNet network model,the landslide range can be quickly and accurately identified;Finally,based on the theory of photogrammetry,DEM and DOM digital results were produced using unmanned aerial vehicle(UAV)aerial images from different periods,and the degree of recurrence of landslide disasters was predicted based on characteristic sections and changes in soil volume.The experimental results show that using UAVs for landslide disaster identification and monitoring is highly feasible,with high accuracy in disaster identification,and meets the needs of geological disaster warning and protection.
作者
程涛
CHENG Tao(China Hebei Construction&Geotechnical Investigation Group Co.,Ltd.,Shijiazhuang 050227,China)
出处
《测绘与空间地理信息》
2025年第10期144-146,共3页
Geomatics & Spatial Information Technology
关键词
无人机
摄影测量
滑坡灾害
深度学习
变化监测
UAV
photogrammetry
landslide disasters
deep learning
change monitoring
