| With the rapid development of urbanization,although the development of deep excavations is achieved,a series of geotechnical problems have been arisen.Considering the complexity of geotechnical engineering,the related theory and numerical calculation methods of Deep Excavations can't well reflect the deformation of the foundation pit supporting structure,the main structure and the surrounding building(structure),which often lead to accidents and great economic losses during the process of construction.Therefore,it seems to be important to monitor the process of construction of deep excavations.With effectively monitoring and acquiring the monitoring data of the foundation pit in time and performing analysis,the test of design expectations,as well as the grasp of the condition of the foundation pit,will be fully controlled.And It is of great significance to guide the construction of deep excavations and ensure the safety of deep excavations.At present,deep excavations are generally developed in the direction of "Being large,deep,tight,close,difficult and dangerous".The surrounding geological conditions and loading conditions are often complicated;the partial overloading phenomenon is serious in particular.However,the monitoring methods based on the traditional measurement methods are point to monitoring processes which have the problems that the Measuring points are far apart and the number of points is small relatively.So it's difficult to completely and timely know the situation in which there are no monitoring points to monitor the deformation of the deep excavation.Therefore,to realize the surface monitoring of deep excavation is of great significance to keep abreast of the secure status of deep excavation during the process of construction.On the basis of the digital close-range photogrammetry technology,this paper studies the external factors which influence the precision of calibration during the camera calibration process and it also studies subpixel localization method in the process of Landmark image point coordinated measurement;the use of three methods of three-dimensional coordinated resolution and image stitching method is to realize deformation monitoring of multiple sample points;the use of various compensation adaptive kalman filter algorithm is to deal with dynamic deformation monitoring data;the validity of the method is measured through accurate analysis test and field test.The paper mainly includes the following aspects:(1)Three external factors such as the number of pictures,light intensity,the size of the checkerboard which influence the precision of camera calibration are analyzed by means of laboratory experiments,and a reasonable field of view of camera calibration is used.Deformation monitoring scheme with multiple sample points is introduced,too.(2)The sub-pixel location of the vertex mark point is realized based on the digital image processing technology and the sub-pixel location technology,and the sub-pixel coordinates of the landmark point is extracted by a simple matching method.(3)Four methods including DLT,the rear-forward intersection method,relative-absolute orientation method,and image splicing method are used to solve the problem where control point and monitoring point occur in different regions.And the four methods are tested by measuring the accuracy of analysis test.(4)Adaptive kalman filtering algorithm with variance compensation filter and projections for deformation monitoring data,taking an engineering instance analysis classical kalman filter algorithm and variance compensation adaptive kalman filtering algorithm in practical application effect in dynamic data processing to prove the variance compensation adaptive kalman filtering algorithm is applied to the rationality of the deformation monitoring data processing.(5)The feasibility and rationality of the application that use digital close-range photogrammetry to achieve the in the monitoring of the deformation of deep excavations are achieves with multiple sample points. |
PDF Full Text Request