Automated single view 3D Texture Mapping and Defect Localisation of Thermography Measurement on Fuselage utilising an industrial robot and laser system

Abstract

Thermography measurements of large objects are always difficult and very expensive to automate. The main reason is that most large objects are bigger than measuring field. Furthermore, continuous measurements are prevented due to lateral thermal diffusion. Another unhandy and less intuitive method is that certified examiners evaluate every 2D thermography image per measuring field one by one. Defect analysis from 2D images restricts to demine the accurate defect size and 3D defect propagation. 3D evaluation is complex but gives complete 3D defect propagation knowledge. Moreover absolute robot accuracy is major challenge for robot based automated thermography. Robot Tool-Centre-Point (TCP) positional accuracy depends mostly on robot pose and their orientation. Many factors affect the absolute positioning accuracy of created robot TCP positions. 0.1° TCP inaccuracy can cause for 1 meter working distance up to 3mm position error. In order to overcome these limitations and retrieve a 3D thermography texture for the observed object, DLR Augsburg has developed both hardware and software components. The essential information required to generate 3D thermography are objects geometry, location and orientation of object with reference to Camera. DLR-AQP department is working on two different methods to calculate 2D-3D camera projection parameters. Method one projects predefined 3D points per measuring field by using laser projector and applies 2D-3D point correspondence algorithm. The second Method uses the robot pose and a kinematic-coordinaterelationships algorithm is applied to the images. For both methods single thermal image is captured by using thermography camera, which is also used for NDT. The first method is suitable for research and development as well as for production inspection with or without using a robot for camera positioning while the second method requires a robot. The first method enables also contactless localisation of detected defects. About accuracy, reproducibility and flexibility of both methods have been compared and discussed in this paper. The 3D results are to be imported into the CAD environment for further structure analysis. Furthermore thermography results are visualised in accordance to their depth in 3D-Thermo-Thermography model.