Noise suppression in scanning Auger Images — comparison of various digital filters

Summary

Due to the small number of electrons collected for each pixel of the image, images from Scanning Auger Microscopy (SAM) are typically noisy. Therefore it is necessary to apply noise suppression techniques both to enhance the optical impression of the picture and to support further image processing methods, such as the analysis of scatter diagrams or segmentation, which are the basis for qualitative and quantitative investigations. A variety of techniques have been proposed to reduce noise in digital images. These can be judged according to their capability of smoothing noise inside homogeneous regions, preservation of edges, sensitivity to outlier contamination, ease of implementation and performance. During the segregation of sulphur in polycrystalline high purity iron, the SAM-images of sulphur show regions of different, homogeneous sulphur coverage [1], which correspond to grains and are sharply separated by grain boundaries. Several noise suppression techniques were applied to these images and the results were judged with respect to the above mentioned characteristics.