No abstract

Digital image correlation (DIC) is a highly accurate image-based deformation measurement method achieving a repeatability in the range of   σ =   10−5 relative to the field-of-view. The method is well accepted in material testing for non-contact strain measurement. However, the correlation makes it computationally slow on conventional, CPU-based computers. Recently, there have been DIC implementations based on graphics processing units (GPU) for strain-field evaluations with numerous templates per image at rather low image rates, but there are no real-time implementations for fast strain measurements with sampling rates above 1 kHz. In this article, a GPU-based 2D-DIC system is described achieving a strain sampling rate of 1.2 kHz with a latency of less than 2 milliseconds. In addition, the system uses the incidental, characteristic microstructure of the specimen surface for marker-free correlation, without need for any surface preparation—even on polished hourglass specimen. The system generates an elongation signal for standard PID-controllers of testing machines so that it directly replaces mechanical extensometers. Strain-controlled LCF measurements of steel, aluminum, and nickel-based superalloys at temperatures of up to 1000 °C are reported and the performance is compared to other path-dependent and path-independent DIC systems. According to our knowledge, this is one of the first GPU-based image processing systems for real-time closed-loop applications.

No abstract

In this paper the results obtained by the use of new CNN based visual algorithms for the control of welding processes are described. The growing number of laser welding applications from automobile production to micro mechanics requires fast systems to create closed loop control for error prevention and correction. Nowadays the image processing frame rates of conventional architectures [1] are not sufficient to control high speed laser welding processes due to the fast fluctuation of the full penetration hole [3]. This paper focuses the attention on new strategies obtained by the use of the Eye-RIS system v1.2 which includes a pixel parallel Cellular Neural Network (CNN) based architecture called Q-Eye [2].In particular, new algorithms for the full penetration hole detection with frame rates up to 24 kHz will be presented. Finally, the results obtained performing real time control of welding processes by the use of these algorithms will be discussed.

No abstract