Quantitative Evaluation of Gray Level Difference and Successive Abandonment Methods Using Artificial Visualized Images

Pattern tracking algorithms using the gray level difference (GD) as a similarity index of correlation area and the successive abandonment (SA) method were evaluated quantitatively through the application to the artificially generated images visualized with particulate or smoke tracers for a three-dimensional impinging jet flow. The performance of both algorithms for smoke images is inferior in general compared with the performance for particulate images. However, the performance for smoke images with high spatial frequency of concentration change is expected to approach the performance for particulate images. The introduction of the safety factor of 2 or 3 to the threshold, which is used in SA and obtained with our statistical model, improves the performance of SA.