Figure 5

Algorithm to detect areas with diffuse defects. The Outlier Image is obtained as for the analysis of the compact defects. A circular kernel is then applied to each pixel in the image to detect areas in which the observed number of outliers exceeds the expected number, based on a binomial test. The defected areas thus determined undergo a round of cluster detection and size threshold, in order to eliminate small areas. The final step involves a dilation and erosion of the defects, in order to better outline the areas.