PSI - Issue 73

Marek Kawulok et al. / Procedia Structural Integrity 73 (2025) 51–57 Marek Kawulok et al. / Structural Integrity Procedia 00 (20 2 5) 000 – 000

53 3

moving. The steel base structure is versatile and has previously been used in other experimental measurements, including studies involving pendulums with viscous dampers (Pospíšil et al., 2014). The second component of the experimental setup is a support structure composed of two wooden plates. The shape of the plates was modified using a laser so that the edges form a rolling track for the ball with a curvature radius of 0.2 m. The edge of the plate that guides the rolling motion was modified to allow the installation of rubber coated tubes, which were added to the structure to improve friction and maintain consistent rolling behaviour. However, the installation of the radius of tubing alters the curvature of the rolling track. The plates are connected by threaded rods, with their precise spacing maintained by 3D-printed spacers. The gap between the plates can be adjusted by varying the spacer length. The entire structure is mounted on a movable trolley and fixed in place with screws. In the experiments presented in this contribution, the spacing between the tracks was set to 0.04 m. The last component of the setup was a red ball with a diameter of 0.0615 m and a mass of 0.219 kg. The experimental setup can be seen in Fig.1. Experimental measurements were recorded using a mobile phone camera at a resolution of 1920×1080 displayed pixels, with a frame rate of 120 frames per second (fps). To ensure consistency and reproducibility, the camera was securely mounted on a tripod to maintain a fixed position throughout all experiments.

Fig. 1. Experimental setup.

3. Determining the position of the ball from video footage To determine the position of the centre of the ball, an algorithm was used to obtain information about the position of the tracked circular points by post-processing the video footage. This algorithm is applied not only to the ball itself, but also to reference points, whose function will be explained in more detail later in the text. Video processing is carried out in several steps. First, the video is divided into individual frames from which distortion caused by optical distortion is removed. For this purpose, tools from the MATLAB Toolboxes for Image Processing and Computer Vision (MathWorks, 2025) were used. To remove distortion, the lens parameters must be known. These were determined before the experiment began. The camera calibration was accomplished using images of the calibration checkerboard. Subsequently, a colour filter is applied to the undistorted images to remove any distracting background. The filter works on the principle of colour contrast between the tracked object and a uniform background, where pixels corresponding to the selected colour are displayed white after filtering, while other areas are suppressed (blackened). The modified images are further analysed using the Hough Circle Transform (Duda and Hart, 1972), which detects the centre of the spherical object in each frame. However, the resulting coordinates are expressed in pixels and, therefore, must be converted into physical units. The conversion is accomplished via the three previously mentioned reference points, which are tracked together with the ball in each frame. These points are positioned at known distances along two perpendicular axes. Based on