A new approach to simulate the dynamic behavior of particulate matter using a canny edge detector embedded PIV algorithm

Over the last few years, the study of characteristics of particles present in the environment becomes an interesting research area for scientists. Simulation of physical and dynamic characteristics of particulate matter (PM) is a prominent area for researchers. For the implementation of particle image velocimetry (PIV) for complex particulate matter with overlapping boundaries, it is necessary to remove non-physical measurements. These non-physical measurements such as unsteady surface and inaccurate edge detection lead to the spurious velocity of particles. In this note, a Canny edge detector is employed to identify the edge of particles. For unsteady surfaces, a special process is followed as follows: (a) find the gradient magnitude in the particles image velocimetry frame from the Canny edge detected frame to optimize particle detection, (b) classify a large high-intensity area in view to extract the particles, (c) find out the rough surface area which contains these particles with their reflections, (d) finally eliminate these particle's reflections. Finally, after this, PIV is implemented on these extracted processed frames from the video dataset to measure the motion of the particles. In this paper, a Canny edge detector with particle image velocimetry (PIV) algorithm is proposed to simulate the dynamic behavior of particulate matter present in particles stock video footage (PSVF) dataset of particles. The proposed model is trained to estimate the motion of particles, and the result showed an accuracy of 92.97% for the particles stock video footage dataset over the other existing methods.