The Installation Deviation Detection Method And Implementation Of Prefabricated Building Components Based On Image Modeling

Since 2016,China has begun to develop prefabricated buildings vigorously,the transportation engineering also responds to the call of the country actively,more and more new built traffic stations begin to adopt prefabricated building method one after another.Prefabricated building components may exist some deviation problems during installation.This existence of these deviations may affect the subsequent construction,and even endanger the use safety of buildings.Therefore,the installation deviation detection,which has significance for construction and management of prefabricated buildings,shall be conducted.The current installation deviation detection for building components often adopts the method of 2D plane measurement to conduct sampling inspection for some components,which is complex and time-consuming with low efficiency and heavy workload and difficult to evaluate the installation deviation of all components with low safety of field operation.The installation deviation detection method of prefabricated building components is put forward in this dissertation based on the image modeling for the above defects.The main contents of this paper are as follows:1、The 3D reconstruction method integrating UAV and camera(air + ground)images.This dissertation put forward 3D modeling method of integrating UAV(unmanned aerial vehicle)and camera images for the problem that it is difficult to collect the image data of components completely,which will affect the accuracy of component 3D model by using UAV singly in the construction site where the scene is complex and the components are covered with each other: Combining with the advantages of the wide range and high efficiency of unmanned collection and free angle and flexible operation of camera collection,the fast and comprehensive image data collection for components is completed,so as to complete the 3D reconstruction of components based on the collected images.The modeling experiment result of Jinjiling Campus Gymnasium of Guilin University of Electronic Technology proved that this method improved the limitation of UAV tilt photographic surveying in the scenes with the existence of obstruction,effectively solved the some problems of local distortion and holes in 3D model caused by the existence of obstruction,and enhanced the integrity and accuracy of 3D model.2、Accuracy analysis and verification of 3D model.In order to verify whether the 3D model accuracy built by 3D modeling method integrating UAV and camera(air + ground)images can satisfy the accuracy requirements of installation deviation detection for prefabricated building components,the method was adopted to complete the 3D reconstruction.The accuracy of the 3D model built by this method were evaluated through the comparison and analysis on the 25 pre-arranged inspection points’ measuring coordinates in the 3D model and total station.The experiment result showed that the the maximum error of point position accuracy in the 3D model built by this method was 2.305 mm with the average middle error of 2.147 mm,which can satisfy the accuracy requirements of installation deviation detection for prefabricated building components.3.Installation deviation detection of prefabricated building components based on 3D model.In order to verify the feasibility in practical application based on the installation deviation detection method of prefabricated building components for image modeling,the 4 wooden columns were taken as test objects to simulate the installation of precast columns in construction site,and completed the reconstruction work of 3D model;The deviation detection was conducted for wooden columns through the measurement for the 3D model.The experiment result verified that the method can intuitively,quickly,accurately and batchly complete the deviation detection for the projects of axis position,elevation and verticality of components,and compared with the traditional detection method,it has the advantages of high data acquisition efficiency,small labor,batch detection of components,non-contact mapping and so on.