基于多样本容量近景摄影测量的基坑稳定性评价

Evaluation for the foundation pit stability based on the multi-sample capacity close-range photogrammetry

基坑变形监测对评价基坑稳定性具有重要意义。针对基坑监测过程中控制点与监测点难以在同一测区内构像的问题,首先,通过分析图片数量、光照强度、棋盘格尺寸3个外界影响因素确定靶标视场范围影响下的相机标定结果的可靠性;然后,通过非量测数码相机采集多测区基坑图像,运用灰度矩亚像素边缘定位算法,对得到照片标志点像点坐标进行亚像素级定位;最后,通过3种近景摄影测量解析方法对基坑多测区坐标的解算实现基坑的多样本容量变形监测。结果表明:当靶标所占像幅比例介于0.2~0.8时,标定结果更可靠;直接线性变换算法(DLT)实现多样本容量(多测点、多测区)基坑变形监测的精度更高;通过测量精度试验及工程实例验证了该方法的可行性与合理性。

Aiming at the problem that the control point and the monitoring point are difficult to form in the same measurement area during the monitoring process of the foundation pit,the basis of the influence of the 3 external influential factors of the picture’s number,the light intensity and checker board size on the calibration accuracy of the camera has to be analyzed through the room-in experiments,so as to determine the optimal field of the target view.For the sub-pixel edge positioning problem of the top corner landmarks has to be solved based on the digital image processing technology and that of the sub-pixel positioning.Moreover,the sub-pixel image point coordinates of the landmarks have to be determined and solved through the straight line fitting so as to find the intersections,so as for the DLT algorithm to be dealt with and solved by using the rear-front intersection method,the relative orientation and absolute orientation method.What is more,since the control point and the monitoring point are not lying in the same measuring areas during the monitoring process of the foundation pit,it would be necessary for us to realize the multi-sample volume deformation monitoring by the virtue of the feasibility and rationality of the digital close-range photogrammetry application to the foundation pit through the measuring accuracy analysis tests and the on-the-spot authentic tests.Thus,starting from such a stand point,the research results can prove that the calibration results turns to be more reliable on the condition if the number of the target image is greater than 20,whereas the accuracy of the calibration results can be effectively improved with its uniformly high light intensity.In addition,when the ratio of the target’s image frame is between 0.2-0.8,the calibration results can be made more reliable.Besides,in the process of measuring the coordinates of the landmark points,the contour marks of the corners instead of the square marks of the corners can be made much more convenient for the Hough transform so as to extract the pixel-level straight edges constituting the coordinates of the center point so as to make much more stable the positioning results of the gray moment sub-pixel edge localization algorithm.And,except that,the measuring accuracy of the above said 3 solution methods can on the whole meet the monitoring demands of the foundation pit deformation,while the DLT algorithm can help to produce the highest measuring accuracy among the above 3 methods.Hence,the above mentioned deformation monitoring method can be adopted to the foundation pit deformation monitoring practice.Moreover,it can also help to solve difficult problems involving the control points and the monitoring points involving the monitoring process of the foundation pit in formulating the same measurement areas.