| Underground space(underground mine,subway,tunnel,civil air defense engineering,underground car yard,etc.)has increasingly become an important place for people’s production activities,which plays an important role in the national economic construction,even plays a very key role.True and accurate 3D scene information of underground space plays a very important role in the safe,stable and efficient operation and management of underground space.However,there are many adverse factors in underground space operation environment(such as narrow space,insufficient light,no GNSS signal,humid air,unequal convex and concave ground),which cause great difficulties in data acquisition.Conventional measurement methods are difficult to meet the requirements for rapid and accurate acquisition of 3D spatial information of the entire underground scene.The mobile LiDAR survey technology is a rapidly developing frontier technology in the field of surveying and mapping.Laser scanner,IMU,GNSS and other ancillary equipment are highly integrated on the mobile platform,and it can quickly obtain detailed 3D spatial information of ground objects along the route during the moving process,which has an incomparable advantage over other data acquisition modes for obtaining banded underground spatial information.Although mobile LiDAR survey provides a good solution for the rapid acquisition of underground space information,it still faces important problems: In the blind environment of underground space,there is no GNSS signal,and the position and attitude of the system are completely calculated by the navigation position of IMU/ otometer.Even the high-precision IMU will be unable to provide reliable position and pose in a very short time due to the excessive drift,and the accuracy of otometer will be seriously decreased in complex irregular motion.Therefore,continuous and effective control of pose is the core problem to solve the underground space mobile LiDAR measurement.This study has carried out targeted research on the above problems,and the main achievements are as follows:(1)The error introduction mechanism and error propagation law of mobile LiDAR measurement in underground space are deeply studied.In cooperation with the research team of Chinese Academy of Surveying and Mapping,UMLS-0,a portable mobile measuring system for underground space measurement,was developed,and the SSW-3 mobile measuring system was adapted.The error introduction mechanism and propagation rule are studied from the aspects of sensor system performance,multi-sensor system integration,external environment,measurement datum,data calculation and so on.The correlation of each error source is analyzed,and the main influencing factors of underground space mobile LiDAR measurement and the change status of the main influencing factors with the change of time,travel and external environment are clarified.(2)A 3D combined target system is developed for mobile LiDAR measurement in underground space.Three-dimensional combined target system adopts intensive design mode,which is composed of prism target combination system,graphic target combination system and rigid frame.It can effectively improve the convenience of position and pose transmission in underground space.The precision calibration of the3 D combined target system is realized based on the spatial 3D network adjustment theory.The robust plane target center positioning algorithm based on the proposed M estimate realizes the accurate positioning of the point cloud target center.The ICP matching algorithm based on the robust estimation realizes the seamless connection between the observation data,the prism combination system,the graphic target system and the measuring point cloud.The reliability of position and pose transmission is guaranteed.(3)The construction method of spatio-temporal reference datum of underground space based on three-dimensional composite target system is proposed.Under the guidance of spatio-temporal maintenance mechanism,a spatio-temporal maintenance reference base of underground space with 3D combined target system as the core was constructed.The reference base of spatio-temporal maintenance of underground space was jointly constructed by integrating external coordinates,3D combined target system,local revisit observation feature information,pose conversion and correction model and other elements.The reliability of the time-space maintenance datum is demonstrated by combining theory with practical measurement.The essence of the construction of the spatio-temporal reference datum of underground space is to introduce the error mechanism and the law of error propagation,weaken and eliminate the error from the mechanism,control the error within a limited range continuously and effectively,and provide the mechanism support for the precise calculation of the final observation data.(4)Multi-source information fusion extended Kalman filter positioning model is proposed.Based on the extended Kalman filter theory of autonomous positioning and some ideas of SLAM theory,automatic positioning and auxiliary positioning are combined to realize the precise pose calculation of internal and external integration.Model parameters include state vectors such as IMU,3D control target and local revisit feature,and the measurement residuals contain position and attitude information at the same time,which changes the traditional model which only relies on point position reference,better matches the parameter types,increases the constraint conditions,and improves the reliability of calculation data.(5)In order to verify the theoretical methods in this paper,summarize experience from experiments and improve relevant methods,a series of experiments are designed: 1)When the cross target is used as the reference base in the subway tunnel,the median error of the corrected point cloud coordinates and reference coordinates in the X,Y and Z directions is 0.051 m,0.162 m and 0.096 m,respectively,and the median error of the point position is 0.195 m.2)Under the condition of GNSS lock-out simulation on the ground,the calculated trajectory is compared with the real trajectory solely by relying on IMU,and the planar graphic target is used as the reference.When the target interval is 50 meters,the error in point position can be controlled within 0.2m.3)The point cloud error can be controlled within 0.3m when19 repeated scans are carried out in the corridor and the 3D combined target system and local revisit feature are used as constraints.And the constraint feature spacing is60-100 m.4)In outdoor comprehensive comparison experiment,3D combined targeting system and local revisiting features are used as constraints,and extended Kalman filtering with fusion constraint features is adopted to correct the pose of the trajectory.When the average distance between constraint features is 80 meters,the point cloud error can be controlled within 0.3m.The effectiveness of the proposed algorithm is also verified by comparison results with point clouds of terrain 3D laser scanner.5)Part of the verification indicators in Experiment 4 are further verified in the underground parking lot.With the support of a small number of 3D combined targets and local revisit features,the point cloud error with 0.3m can be achieved.51 figures,12 tables,176 references. |
PDF Full Text Request