A Strong Scattering Solution with Applications to Remote Sensing Systems Modelling

Although not always fully appreciated as such, the linear stationary convolution model for a signal and/or image, generated by the interaction of an incident wave field with an inhomogeneous medium, is based on the 'weak scattering condition', commonly referred to as the 'Born approximation'. This approximation forms the basis for signal processing and analysis associated with applications over a broad range of frequencies. The Born approximation assumes that a scattering interaction is based on single scattering events alone and that the scattered field is therefore a weak field in comparison to the incident field. The physical limitations of such a model are usually overcome by introducing an additional stochastic field term which takes into account all effects that do not conform to the Born approximation, coupled with background 'system noise'. In this paper, a solution to the 'scattering problem' for scalar fields is presented, which does not explicitly rely on the weak scattering condition. It is shown how, for side-band systems, this solution can be cast in terms of a linear stationary model, thereby allowing standard signal analysis and processing methods to be adopted. A case study is presented based on the application of a self-affine model for the scattering function to simulate Synthetic Aperture Radar images.