论文部分内容阅读
两次扫描间扫描单点具有不能重叠和不可辨识的特点,目前同名变形点的获取一般会针对监测对象的点云特点采用不同的方法,如重心法、靶标拟合法和模型求差法等,具体应用时受到很多限制,或得到的同名变形点数太少或要求监测表面较连续光滑。文中提出一种基于两期点云直接计算变形的新方法,即通过对扫描点云进行特征分割、坐标变换与投影、格网划分、最短坐标距离计算与排序、中位值段取平均以及逆变换等一系列处理过程后,最终得到整个变形体高密度监测点的变形值。室内外实验的结果表明,该方法采用无接触测量,不需要对原始点云进行预处理,具有很强的抗噪声能力,能形成密集的同名变形监测点,适合于任意表面,在野外150 m范围可达到mm级变形监测精度。
At present, the acquisition of the deformation point with the same name generally adopts different methods for the point cloud characteristics of the monitoring object, such as the center of gravity method, the target fitting method and the model difference method, etc. Therefore, Specific applications by a lot of restrictions, or get the same name deformation points too little or require continuous monitoring of the surface smooth. This paper presents a new method based on direct calculation of two-point cloud deformation, that is, by performing feature segmentation, coordinate transformation and projection, grid partition, shortest coordinate distance calculation and sorting, Transformation and a series of processing, the final deformable body densification of high density monitoring point. The results of indoor and outdoor experiments show that this method uses non-contact measurement and does not need to preprocess the original point cloud. It has strong anti-noise ability and can form dense dense deformation monitoring points with the same name, which is suitable for any surface. In the field of 150 m Range can reach mm level deformation monitoring accuracy.