非対称多時刻粒子軌跡を用いた気流速度分布計測

抄録

We propose a PIV method which uses digital single-lens reflex (SLR) camera and diode laser which illuminates several pulses for obtaining several pairs of images from single image. This method can improve the resolution of the visualized images and it aim at the application to high-speed airflow measurement and acceleration measurement. Each pairs of trajectory of each particle are recognized in visualized image and the distance between each center of weight of each pulse are evaluated to generate velocity vectors. In this paper, in order to measure the velocity of the swirling flow in the system of transferring the semiconductor wafer, we developed a trajectory detection algorithm based on the relative shape of the trajectory. The system is the non-contact convey method which uses the balance of negative pressure by the velocity in the chamber and positive pressure by blowing out the air. Flow rate of the air is set to 5.0 L/min and the total time of a pair of pulse is 5ms. Visualized area was illuminated by DPSS laser (1000mW) and visualized images are taken by digital SLR camera (6000 × 4000 pixels). About 40% of the trajectories were detected by combining several results which were obtained from one visualization image. The mismatched vector generated in the detection process was removed manually. It is necessary to improve the recognition of trajectories by improving the algorithm of image processing and visualization method.