Deconvolution Algorithm Based on Generalized Inverse

Deconvolution with frequency-domain division is ill posed when the kernel contains zero(s) in its Fourier spectrum. An improved deconvolution algorithm through time-domain cyclic convolution based on the concept of generalized inverse is proposed to circumvent such difficulty. Its frequency-domain equivalent approach is also introduced. It is easy to be implemented by programming and its computation complexity is just proportional to the fast Fourier transform (FFT).