Research Of Displacement Computing Based On Point Cloud Of Characteristics

In the high slope, buildings and Bridges monitoring and so on various aspects are mostly based on sensor monitoring and the monitoring of3d laserscanner. Based on the monitoring of sensors has the advantage of quick,real-time transmission is detected morphology change of the object, isdeficiency in this way can only reflect the sensor monitoring by that point, andcan’t overall change of the observed objects. Monitoring based on3d laserscanner can monitor the object’s3d changes, so as to calculate thedeformation displacement of buildings, to the structure of the healthmonitoring.This paper proposes a displacement calculation method based on thecharacteristics of point cloud, the parameterized method is first of all, thewhole point cloud data; Then for the next to get point cloud data to establish aparameterize the same coordinate system, such two get point cloud data can becompared under the same coordinate, if somewhere of these changes, find themethod of point cloud characteristics is used to calculate building deformationdisplacement. Point cloud characteristics include gaussian curvature extremummethod to extract the feature points, the discount method to extract thecharacteristic line and region growing method to extract features from threeaspects, according to the different characteristic values of point cloud, canscan the object coordinate system is established.In this paper, the main work is as follows:(1)Using3d laser scanner to get point cloud data, to get point cloud datapreprocessing, denoising method using point cloud border to denoising, data sampling method to streamline of point cloud data, based on point cloud dataand the main direction of legal automatic initial registration.(2)Since three-dimensional point cloud data by using data obtained fromafter the initial registration does not meet the requirements of accuracy, inorder to further establish deformation displacement calculation model,accurate three-dimensional point cloud data registration, the original ICPregistration algorithm was improved, implements the more precisethree-dimensional point cloud data registration, so that the displacement ofcalculation is more accurate. Using the algorithm of point cloud dataregistration of a building, not only has the good effect of registration, andachieve the high efficiency of registration.(3)From point cloud of feature point and feature line, characteristics ofsurface concrete analysis in three aspects: the theory and method of thedisplacement monitoring, respectively the three features are applied to thedisplacement calculation of experiment, different experiment results areobtained. Contrast experiment results and theoretical analysis, found that thedisplacement calculation based on the feature points can be a single point, theeffect is not ideal, and based on the characteristics of the displacementcalculation is too complex, only displacement calculation based on thecharacteristic line can achieve more ideal effect, can reflect the overall objectof change; And analyzes the possible error and how to reduce experimentalerror.