The nonlinear image processing techniques discussed in this chapter can be separated into three categories. In the first category are those filters that operate within a small local window as defined by a filter template similar to the spatial filters discussed in Chapter 4. Many of these filters are based on the graylevel ordering of the pixels included in the filter operation from their minimum to their maximum values. The second area of nonlinear filters discussed includes several of the adaptive filters that are commonly used in electronic image processing. These filters change their filtering characteristics depending on the noise and the image features that are being filtered. The last category discussed is homomorphic filtering, used to remove multiplicative noise and image shading from an image.
Online access to SPIE eBooks is limited to subscribing institutions.