Search Results

Industrial robot manipulators are widely used for different automated applications where flexibility is necessary or where repetitive tasks are present. These systems are designed for optimum repeatability and little emphasis is given to the accuracy aspect. In recent years, we see an increase of interest in the aerospace industry for using industrial robots in applications where more precision is necessary. For this purpose, many solutions were proposed in the last years for dynamically tracking the robot TCP (tool center point). These systems are very interesting when the workspace and the robot cell are large enough to accommodate the minimum stand-off of the tracking system. In this work, we present two alternatives for increasing the accuracy of industrial robots. The first approach uses vision-based servoing of the TCP in order to accomplish specific tasks in small workspaces. One and two cameras scenarios are presented.

3D digitalization and modeling is very popular in industrial applications such as metrology, geometric dimensioning and tolerancing (GD&T) and tridimensional object inspection. The 3D captured data can be used for the inspection of object surfaces. However, hidden defects cannot be detected with this technology. In order to get this kind of information, NDT/E (Non-Destructive Testing/Examination) approaches are used (ultrasounds, infrared imaging, etc.). These two types of information are very important in aerospace products inspection (e.g. composite materials). The two modalities (3D Vision and NDT) are still used separately. In this work, we present a new multi-modal fusion framework for combining these two data types. The fusion scheme works on 3D image data and temporal thermal images in order to obtain a hybrid model for simultaneously inspecting the 3D surface and the hidden sub-surface defects.