基于结构特征的大型飞机整体油箱点云初始配准方法OA

To meet the requirement for generating digital models of large aircraft integral fuel tanks,proposes a coarse registration method for measured point clouds based on structural features.Traditional coarse registration methods exhibit low discriminative power for features on integral tank panels,which easily leads to mismatching.This issue primarily stems from the regular distribution of numerous fasteners(such as rivets,bolts,etc.)on the tank's inner surface,creating repetitive structural patterns in local regions.Such repetitive features significantly degrade the accuracy of coarse registration,consequently affecting the quality of subsequent 3D model reconstruction.Therefore,constructs a feature descriptor capable of uniquely characterizing the structural configuration of integral fuel tank panels.This descriptor aims to enhance the accuracy and precision of coarse registration in the presence of repetitive structures,thus providing a favorable initial point cloud position for subsequent steps like fine registration and model reconstruction.