Research On Pose Optimization Method Of Mobile Laser Survey System In Underground Space

With the gradual improvement of the level of urbanization in China,the contradiction between population growth and the increasing shortage of land resources has prompted people to pay attention to underground space.Due to the problems of insufficient light and lack of GNSS signals in the underground environment,it is particularly difficult to obtain information on large areas of underground space.The mobile laser surveying system has the advantages of wide adaptability,fast scanning efficiency and good data quality.However,the mobile laser surveying system relies heavily on the navigation equipment to provide a position reference.Therefore,the effective solution to the positioning problem of the mobile laser surveying system in the GNSS signal loss scenario is the key to realizing the rapid acquisition of underground spatial data.This paper studies the pose calculation problem of mobile laser measurement system in the absence of GNSS for a long time.The main research results are as follows:(1)Firstly,the composition of the underground mobile measurement system is introduced,including the carrier platform,laser scanner and POS system,and then several commonly used coordinate systems in the mobile measurement system are introduced,and the conversion model between the coordinate systems is deduced.Finally,the error source of the underground mobile measurement system is analyzed and researched.(2)According to the characteristics of 3D information acquisition in underground space and the need to achieve the consistency of indoor and outdoor spaces,a 3D comprehensive target system was developed.In order to realize the precise calibration of the target system,the problem of accurate extraction of the center coordinates of the plane target is studied in depth,and a robust algorithm for extracting the center coordinates of the point cloud plane target center is proposed.The filtering algorithm extracts the initial target point cloud;then uses RANSAC and the overall least squares algorithm to perform optimal plane fitting on the initial target point cloud;then transforms the target point cloud and uses a dynamic ellipse for the projected point cloud.The convex hull algorithm is used to extract the boundary points,and finally,the RANSAC algorithm is used to extract the target center coordinates for the boundary points.Through experimental analysis,it is verified that the algorithm has the characteristics of high fitting accuracy and strong antiinterference ability.(3)The key problem of optimizing the trajectory error by revisiting the point cloud is the accurate solution of the transformation parameters.This paper studies the problem of accurate registration of the point cloud,and proposes a feature extraction algorithm based on ISS and an improved rotational projection statistical feature description This method first extracts some ISS feature points from the original point cloud,and then uses the rotational projection statistical feature descriptor to describe these points.Here,a local reference coordinate established based on covariance is used.The system replaces the local reference coordinate system in the original algorithm to improve the generation speed of feature descriptors,and then extracts the corresponding point set according to the similarity between the descriptors and uses the RANSAC algorithm to eliminate erroneous corresponding points.Finally,an improved iterative closest point algorithm based on k-d tree is used for fine registration.The experimental results show that the method has the advantages of strong anti-interference ability,high registration accuracy and fast calculation speed,and can meet the registration requirements in practical engineering.(4)Aiming at the problem of pose and attitude calculation of mobile measurement system without GNSS signal,an initial trajectory correction and secondary optimization method relying on 3D comprehensive target and revisit point cloud registration is proposed.Attitude estimation model of mobile laser measurement system based on filtering algorithm and trajectory optimization model based on cubic spline interpolation.The experimental results show that the method in this paper constrains the checkpoint residual within 0.15 m,which effectively solves the serious deviation problem caused by the IMU drift of the mobile laser measurement system in the underground environment.