Optimized Image Processing Based on CUDA in a Combined Measurement Technique of PIV and Shadowgraph

Abstract

The issue of ventilation concerning Air Quality Index (AQI) has been widely investigated theoretically and experimentally. In this work, an optimized image processing based on CUDA is applied to an experimental method of simultaneous flow visualization with temperature variation. With the use of GPU, the processing time of an image sample (of resolution 2048 × 2048) is reduced to about 0.2 seconds. It is about 10X speedup ratio calculation on GPU than that on CPU. By this means, we achieved a 4FPS (frames per second) real-time experimental faculty to get the velocity field and the temperature variation of the entire flow zone during the experiment