Computer Vision

Additive manufacturing is an increasingly important manufacturing technology with huge economical potential. However, its popularity is accompanied by high material and time losses, as defects are often detected at a very late stage. One solution for a more sustainable production is the automated detection of manufacturing defects using artificial intelligence. This article describes the digitization of the defect detection process in additive manufacturing using a system based on a neural network. In addition to the steps for automated defect detection, system performance is also discussed.

Optical systems are a popular choice for quality inspection because they are not only contactless and precise but also comparably fast. Particularly in cases where a 100% quality inspection is required low measurement and evaluation time is of essential importance. With high production rates of several parts per second, manual inspection is not feasible anymore and the evaluation needs to be automatically carried out by algorithms. In this article we propose a fast method for anomaly detection in image data based on principal component analysis and filtering. The method shows competitive performance on a data set of challenging surface inspection.

Progress in the areas of 3D sensor systems and artificial intelligence provides new opportunities for the development of flexible robotic systems that are applicable in scenarios without predefined and constant environmental conditions or standardized processes. An example from the field of logistics is the automatic unloading of containers. The development of a suitable robotic system on the one hand requires applicable gripping technologies, on the other hand it requires a reliable object recognition method in order to recognize and localize differently shaped logistic goods within a packaging scenario. This paper presents an object recognition method for logistic goods from three different shape classes using point clouds acquired by a laser scanner. The method is evaluated with real packaging scenarios.