A method for detecting bonding quality of a complex curved surface

Abstract

The application provides a complex curved surface bonding quality detection method, which comprises the following steps: defining a rotary table coordinate system and a robot base coordinate system and calibrating a pose, collecting multi-view point cloud data of an inner workpiece, an outer workpiece and an assembled component; adopting a TrICP algorithm based on curvature similarity improvement to perform point cloud registration, and generating three-dimensional surface models of the inner workpiece, the outer workpiece and the assembled component respectively; combining the inner and outer point clouds to the same coordinate system and making the axes collinear to form a combined point cloud; according to a non-contact measurement principle, solving the normal direction of the surface point cloud of the bonding area, calculating the distance from the surface of the component, the surface of the inner layer and the surface of the outer layer to the axis along the normal direction respectively, and measuring the glue layer thickness through a projection difference; and reflecting the curved surface bonding quality according to the glue layer thickness. The application breaks through the precision limitation of the traditional contact measurement, realizes high-precision, non-contact and automatic glue layer thickness detection, and has significant engineering application value and wide popularization potential.

Images