A 2d+3d large aircraft shape defect detection and analysis method based on deep learning

Abstract

The invention discloses a 2D + 3D large aircraft shape defect detection and analysis method based on deep learning, and the method is characterized in that the method comprises the steps of collectinga multi-view 2D image and 3D point cloud data; obtaining a complete aircraft point cloud model through registration; extracting feature points of the image and the point cloud, and performing 2D-3D correspondence according to feature matching; estimating the pose of the camera according to the 2D-3D corresponding relation; according to the camera pose, realizing assignment from the texture colorof the 2D image to the 3D point cloud; determining point cloud semantic segmentation according to the point cloud color and the coordinate information; and detecting and analyzing aircraft appearancedefects according to the point cloud semantic segmentation. The invention discloses the 2D + 3D large aircraft shape defect detection and analysis method based on deep learning. Vision sensor equipment and an optical three-dimensional detection system measurement technology are used for processing and analyzing collected 2D + 3D data, shape defects on a large aircraft can be accurately and effectively detected and extracted, the conception is reasonable, and in practice, automatic application can be achieved in aircraft safety inspection and other scenes.

Description

technical field [0001] The invention relates to the fields of deep learning, computer vision, graphics, etc., and specifically relates to a detection and analysis method for shape defects of large aircraft. Background technique [0002] The traditional non-destructive testing (NDT) technology uses the characteristics of sound, light, magnetism, and electricity to detect whether there are defects or unevenness in the inspected object without damaging or affecting the performance of the inspected object. It can provide information such as the size, location, nature and quantity of defects, and then determine the current technical status of the inspected object (such as whether it is qualified or not, remaining life, etc.), which plays an important role in the detection of large aircraft defects. However, the traditional non-destructive testing has the characteristics of high cost and slow speed, and although its detection rate of volume defects (pores, slag inclusions, tungste...

AI Technical Summary

Problems solved by technology

However, although the 3D point cloud data obtained by the laser tracker has high precision, it also has its own problems, such as relatively sparse points, and the effective perception distance for the algorithm will not exceed 10m

In addition, based directly on the 3D point cloud, the difficulty is that there are no visual features so that tasks such as tracking and positioning are not as direct as in vision.

Images