一种用于标定工业CT设备点云提取精度的配准及评价算法
A Registration and Evaluation Algorithm for the Calibration of Point Extraction Precision Based on ICT Equipment
-
摘要: 基于工业CT测量数据的逆向建模技术广泛应用于军事、汽车制造、航空、航天等领域。目前,在基于工业CT点云数据的逆向重构精度验证和标定上,国内外缺少相关研究。本文提出了一种用于标定工业CT设备点云提取精度的配准及评价算法。基于CT扫描和三维光学扫描点云数据,在研究多种配准算法的基础上,总结手动选点粗配准和自动全局精配准相结合的配准算法:手动选择三个匹配点,利用坐标变换完成粗配准;用ICP算法完成自动精配准,并利用修正的Hausdorff距离为目标函数来评价配准结果。对基于七参数法和ICP算法的计算步骤做了详细说明,并赋以实例。最后,以自行设计的标准件为例,获得CT测量设备点云提取精度标定值,验证配准流程的适用性。Abstract: Reverse modeling based on CT measuring data is widely used in military, automotive, aviation, aerospace and other fields. At present, there is a lack of relevant research to verify the accuracy of the points based on industrial ICT equipment. A kind of registration and evaluation algorithm for the calibration of point extraction precision based on ICT equipment is presented. On the basis of many registration algorithms, a method that contains manual original registration and auto accurate registration is described. Two set of data for registration are obtained from ICT equipment and 3D optical scanner. The Modified Hausdorff distance between each two correlative points is used to verify the algorithm accuracy. The calculation steps of seven parameters and ICP(Iterative Closest Point) algorithms are presented detailed to demonstrate how to complete the original registration and accurate registration. The evaluation value for point-cloud data extraction accuracy based on CT equipment is calculated and the applicability of registration process is validated as self-designed standard parts for example.