When using a laser scanner to scan the aircraft skin

the cross surface hole defects may occur due to the occlusion caused by the aircraft frame

which makes it difficult for reverse software to repair and affects the accuracy of the 3D reconstruction. In order to repair the cross surface hole defects

a method based on feature region classification for repairing cross surface hole defects in aircraft skin is proposed. Firstly

feature regions are defined based on the feature values to classify cross surface hole types and detect seed point clouds of cross surface holes. Then

the point cloud region growth method is used to segment the cross surface holes to obtain small area holes. Finally

the least squares hole filling method is used to repair the cross surface holes. The test results of the cross surface hole repair on public datasets show that the algorithm can effectively restore the missing regions. Based on the above verification

the experiment is conducted on the actual data of cross surface holes in the engineering example of aircraft skin. By comparing with the repair effects of reverse soft-ware and radial basis function algorithm

the experimental results show that the present method has the highest overlap degree of 98.52% with the original model in the repair area

maintains the curvature characteristics of the holes

and achieves good repair effect

providing a technical basis for detect the damaged regions on airplane.