基于图像的形变远程监测方法及应用  

Image-based remote deformation monitoring method and application

作　　者：张宇 ZHANG Yu(China Railway Design Corporation,Tianjin 300000,China)

机构地区：[1]中国铁路设计集团有限公司,天津300000

出　　处：《北京测绘》2023年第11期1545-1551,共7页Beijing Surveying and Mapping

基　　金：中国铁路设计集团有限公司科技研究开发计划重点课题(2023A0253810);中国铁路设计集团有限公司科技研究开发计划(2023A0253810)。

摘　　要：随着信息技术和计算机科学的发展,对形变监测的智能化也提出了更高的要求。研究了一种基于图像的形变分析方法,对监测设备先进行畸变矫正,采集目标区域的序列影像,预处理后对采集设备进行采集得到影像进行预处理,然后使用尺度不变特征转换算法(SIFT)特征匹配找到诸多特征点,对序列影像采用全局几何匹配(APM)方法做进一步高精度配准达到子像素的精度,解算出其位移形变量。在此基础上设计开发了一套监测系统,通过与全站仪同步观测靶标对比实验,验证了该方法在10 m范围满足0.3 mm的监测精度,通过应用实践说明该方法可以有效解决工程中的问题。With the development of information technology and computer science,higher requirements for the intelligence of deformation monitoring have been put forward.An image-based deformation analysis method was studied.The distortion of the monitoring equipment was first corrected,and the sequence image of the target area was collected.After pre-processing,the image collected by the acquisition equipment was pre-processed,and then many feature points were found by using the scale-invariant feature transform(SIFT)algorithm for feature matching.The sequence image was further aligned with high precision by using the all-pixels participated matching(APM)method to achieve the accuracy of subpixels,and the displacement deformation variables were solved.On this basis,a monitoring system was designed and developed.Through the comparison experiment with the total station synchronous observation target,the method was verified to meet the monitoring accuracy of 0.3 mm in the range of 10 m,and the application practice shows that the method can effectively solve the problems in engineering.

关 键 词：自动化监测 图像 配准 子像素 位移 

分 类 号：P258[天文地球—测绘科学与技术]