A Small Target Detection Method based on Generalized S-Transform

A novel small target detection method based on 2-D Generalized S-transform is proposed in this paper. A four dimensional array was obtained after the 2-D Generalized S-transform of an image and the energy sum of nonzero frequency components in the S-transform domain can be computed. Comparing the energy sum of nonzero frequency components with a predefined threshold at each point, the position of suspected targets were obtained. The double window technique is chosen to recognize real small target in spatial domain. The experiments of real images show that this method can detect the small target in low signal-noise-ratio image effectively.