The goal of the present master’s thesis was to analyze object recognition for the automatic finding and picking up of objects, the so-called “pick and place” procedure, and to show possibilities for integration with industrial production processes. In chapter 1 of this work, the theoretical background is explained. In particular, the conclusions of this part were that an integration of such a procedure with an orchestrable manufacturing process emerges a huge potential for the industry, but no interoperable and reasonably priced solutions are currently o?ered. The remainder of this thesis was therefore dedicated to the goal of finding such a solution and proving that an integration is possible. In chapter 3 di?erent concepts for multi-object recognition are explained and tested in order to try if they are suitable for the above-mentioned mission. The application of the neural network 3D-R2N2, the stereo reconstruction approach, as well as the use of a Microsoft Kinect camera turned out to be inappropriate. The use of an Intel RealSense camera in combination with an object recognition algorithm for point clouds, on the other hand, proved to be promising. As a consequence, two object recognition implementations, part of the Point Cloud Library, were tried out. Both of them could find the searched-for objects in less than one minute of calculation time, but the found positions were drifted apart from the actual locations with a squared distance error of 0.60 and 0.42 mm2, respectively. The error and calculation duration behavior depending on the di?erent algorithm parameter was shown using a comprehensive experiment including statistical evaluations. That is why a tailored object recognition algorithm was developed as part of this thesis, which takes advantage of the constraints specific to the thesis’ task in order to achieve precise results in less time. The algorithm was afterwards integrated into a process engine and tested in a field test using a Universal Robot. This way, the thesis could be concluded with proof that the integration is possible. The software’s source code and a large number of photographs of the practical field test are present within this thesis.