In this work SAR/ISAR (Inverse Synthetic Aperture Radar/Inverse Synthetic Aperture Radar) parametric image reconstruction concepts are discussed. First, an image reconstruction procedure based on l0 norm optimization is developed and applied over reduced number of measurements defined by randomly generated azimuth and range sensing matrices. Second, Kalman algorithm is applied for ISAR image extraction. Vector measurement and state equations are derived. Approximation functions based on LFM signal are defined. Results of numerical experiments are presented.