A method for testing the bonding quality of complex curved surfaces

Abstract

This invention proposes a method for detecting the bonding quality of complex curved surfaces, comprising: defining a turntable coordinate system and a robot base coordinate system and calibrating their poses; acquiring multi-view point cloud data of the inner workpiece, the outer workpiece, and the assembled assembly; using a TrICP algorithm based on curvature similarity to perform point cloud registration, generating three-dimensional surface models of the inner layer, outer layer, and assembled assembly respectively; combining the inner and outer layer point clouds into the same coordinate system and ensuring their axes are collinear to form a combined point cloud; based on the principle of non-contact measurement, solving for the normal direction of the surface point cloud of the bonding area, calculating the distances from the component surface, inner layer surface, and outer layer surface to the axis along the normal direction, and calculating the adhesive layer thickness through projection difference; and using the adhesive layer thickness to reflect the bonding quality of the curved surface. This invention overcomes the accuracy limitations of traditional contact measurement, achieving high-precision, non-contact, and automated adhesive layer thickness detection, and has significant engineering application value and broad promotion potential.

Images