A new fast discrete Radon transform for enhancing linear features in noisy images

Noise contributions along lines in an image tend to cancel because of the integration process, whereas contributions derived from a linear feature tend to be accentuated. Consequently, bright lines, which may be obscured by the noise in an image, appear as bright patches in Radon feature space, and conversely, dark lines appear as dark patches in Radon feature space. This process is particularly beneficial for linear feature identification in synthetic aperture radar (SAR) images which are inherently prone to speckle (noise). A method of reducing the computational requirements of the Radon transform is derived from the use of the frequency domain representation of the image function, f (x,y). The Fourier Slice Theorem lies at the heart of the method as it relates the one-dimensional Fourier transform of a projection of a function f(x,y) to the two-dimensional Fourier transform of f(x,y). The authors show that the Radon transform involves the repeated use of a one-dimensional FFT computation