ABSTRACT

The design of 2-D digital filters, like the design of 1-D filters, encompasses four different steps in general, as follows: Approximation, Realization, Implementation and Study of quantization effects. The implementation of the filter depends on the application and may be in terms of software or hardware. This chapter is concerned with the solution of the approximation problem in the case of non-recursive filters. One of the successful methods for the design of 1-D non-recursive filters is based on the application of the Fourier series. The amplitudes of Gibbs’ oscillations can be reduced by using discrete window functions. An alternative window function which has the attractive property that its ripple ratio can be adjusted continuously from the low value of the Blackman window to the high value of the rectangular window by changing a window parameter is the Kaiser window function. The chapter considers the design of circularly symmetric lowpass, highpass, bandpass and bandstop filters.