The previous chapters have included examples of filtered images and a discussion of the implementation of morphological and logic-based filters. This chapter presents a case study showing how a morphological filter may be designed for a specific type of noise in images, namely astronomical images.
Imaging instrumentation is widely used in space-based astronomy and solar physics where it has the potential to produce excellent pictures. However, these are frequently degraded by bursts of cosmic ray ions that saturate the charge-coupled devices (CCDs) and produce an overlaid speckle. This is a source of frustration to observers and can obscure vital detail. In this chapter it will be shown that the speckle may be removed from the image using a type of nonlinear filter known as a soft morphological filter.
Soft morphological filters comprise a branch of nonlinear image processing that is particularly effective for noise removal. They originate from the field of mathematical morphology but their operations are less harsh since the structuring elements used are designed to have soft boundaries. The implementation of such filters makes extensive use of rank-ordering operations.
The chapter will describe how a training set may be created for the images and how the optimal filters may be derived using genetic algorithms. The results of processing the images with the optimal filters will be presented. Finally, experiences of implementing the filters in programmable hardware will be given.
Online access to SPIE eBooks is limited to subscribing institutions.