基于亚像素无标点法的掌子面变形监测技术

Deformation Monitoring for Tunnel Face based on Subpixel Non-target Method

为了改变传统的变形监测方法存在的耗时、耗力及危险的处境,基于当前的图像处理技术,提出无标点亚像素监测法。以灰度图像相关性算法为基础,设计室内试验,把角点设为特征点,对无光线变化且没有受到意外扰动的照片进行分析,将试验结果与理论精度相比较。结果表明,无标点亚像素监测法在距离1m情况下,像素精度为0.286pix,位移识别精度为0.065mm,精度较高且误差波动不大。依托江西萍乡长平隧道掌子面进行现场试验,监测结果与理论研究及室内试验结果相比误差有所提高,但仍保持了较高的监测精度,能够满足工程变形监测的需求。表明该方法有较高的研究与应用价值。

The traditional deformation monitoring is time-consuming, labor-intensive, sometimes is even dangerous. This paper proposes the no-marking sub-pixel monitoring method basect on tne current image processing technique. The indoor experiment is conducted on the basis of image correlation calculation method, the corner is set as the feature point, and the pictures with no accidental disturbance and no light changes are analyzed. By comparing the research result with the theoretical accuracy, it can come to the following conclusion- the pixel accuracy is 0. 286pix and the displacement accuracy is 0. 065mm under the distance of lm. The no-marking sub-pixel monitoring method provides high accuracy and small deviation. In-site tests are conducted for Changping Tunnel face in Pingxiang, Jiangxi Province and the deviation increases compared to the theoretical research and indoor experiment, but the method still keeps a good monitoring accuracy and meets the requirement of engineering deformation monitoring. This proves the research and application value of the method.