Gradient is an important operator in signal processing. A gradient operation always is sensitive to noisy environment. In this paper, a new differential method is designed in the field of one-dimensional signal processing which is robust in low signal to noise ratio. The proposed scheme is suitable for signal differentiation in analog and digital signal processing applications. For low frequencies, this scheme performs the signal derivative, but for high frequencies recasts to a constant gain. This prevents differentiation of high frequency components of added noise. The proposed differential operator is applied to sound and image signals. Results show superiority of the proposed method.

Keywords