多尺度信息融合的船舶外板检测三维标志点识别方法

A 3D Mark Point Recognition Method for Ship Outer Plate Inspection

摘要:

针对船舶外板开展数字化检测技术研究对提升船舶外板智能制造水平具有重要意义。为解决船舶外板检测参考点坐标获取问题，提出了一种识别三维标志点的方法，可在船舶外板数字化检测过程中为实测点云和模型点云的匹配提供参考坐标信息。该方法针对船舶外板结构特点和形面测量需求，设计三维标志点的几何结构。此外，结合三维标志点自身多种特征和周边结构差异化特征，将标志点多种尺度信息进行融合，实现对标志点的精准识别。最后通过标志点测量数据计算对应船舶外板上定位点的三维坐标。通过试验验证了在所提出的方法在三维标志点识别方面具有准确性高、定位精准的优点，将大幅提升船舶外板数字化检测的效率。

Abstract:

The research on digital inspection technology for ship outer panels is of great significance for improving the intelligent manufacturing level of ship outer panels. To solve the problem of obtaining the coordinates of reference points for ship outer plate detection, a method for identifying three-dimensional marker points is proposed, which can provide reference coordinate information for the matching of measured point clouds and model point clouds in the digital detection process of ship outer plates. This method designs the geometric structure of three-dimensional marker points based on the structural characteristics and surface measurement requirements of ship outer panels. In addition, by combining the various features of 3D landmark points themselves and the differentiated features of surrounding structures, the multi-scale information of landmark points is fused to achieve accurate recognition of landmark points. Finally, calculate the three-dimensional coordinates of the corresponding positioning points on the outer plate of the ship through the measurement data of the marker points. Through experiments, it has been verified that the proposed method has the advantages of high robustness, accuracy, and speed in 3D marker point recognition, which will greatly improve the efficiency of digital detection of ship outer panels.