露天矿滑坡位移监测的ASIFT算法

发布时间：2018-05-10 17:35

本文选题：露天矿 + ASIFT算法　；参考：《金属矿山》2017年05期

【摘要】：针对现有的大型露天矿边坡大变形监测方法存在的不足,提出了一种基于ASIFT图像特征匹配算法的滑坡位移监测方法。首先基于露天矿滑坡真实地形空间分布形态,设计了滑坡物理模型,并利用高分辨率数码相机获取模型滑动过程中的序列影像;然后利用ASIFT特征匹配算法对滑坡各阶段光学影像进行特征点提取与匹配,并利用RANSAC随机抽样一致算法剔除误匹配点;最后利用滑坡体运动特征矢量集模型计算出特征点空间矢量位移,从而实现了滑坡位移场标定与滑坡范围确定。研究表明:所提算法对于滑坡蠕变阶段的位移监测效果一般,但可有效监测滑坡中后期发生明显变形时的位移,该算法的特征点匹配数量及匹配正确率均明显优于SIFT算法,且位移矢量标定精度也得到明显提高,反映出该算法适合于大型滑坡后期发生大变形时的位移量监测。
[Abstract]:In view of the shortcomings of the existing monitoring methods for large slope deformation in large open-pit mines, a landslide displacement monitoring method based on ASIFT image feature matching algorithm is proposed. Firstly, the physical model of landslide is designed based on the spatial distribution of the real landform in open-pit mine, and the sequence image of the sliding process of the model is obtained by using high-resolution digital camera. Then the ASIFT feature matching algorithm is used to extract and match the feature points in the optical images of different stages of the landslide, and the RANSAC random sampling algorithm is used to eliminate the mismatch points. Finally, the space vector displacement of feature points is calculated by using the feature vector set model of landslide motion, which realizes the calibration of landslide displacement field and the determination of landslide range. The results show that the proposed algorithm is effective in monitoring the displacement of landslide in creep stage, but it can effectively monitor the displacement of landslide in the middle and late stages of deformation. The number of feature points matching and the matching accuracy of this algorithm are obviously better than that of SIFT algorithm. The accuracy of displacement vector calibration is also improved obviously, which shows that the algorithm is suitable for monitoring displacement in the late stage of large landslide.
【作者单位】：东北大学资源与土木工程学院;辽宁省交通高等专科学校测绘系;
【基金】：国家自然科学基金项目(编号:41074127) 国家重点基础研究发展计划(973计划)项目(编号:2011CB707102)
【分类号】：TD804;TP391.41

【相似文献】