Dantzig selector based compressive sensing for radar image enhancement

Compressive sensing (CS) is the technique for acquiring and reconstructing a signal utilizing the apriori knowledge that it is sparse in a certain domain. This paper investigates the application of this technique to radar imaging. Present radar systems operate on high bandwidths and demands high sample rates following the Nyquist-Shannon theorem. Compressive Sensing can prove to be a good alternative to reduce data handling, complexity, weight, power demands and costs of the existing radar systems. There are two major novelties in this work. First of all we have used Dantzig selector based CS which gives better result when applied on radar images than that using the conventional ℓ₁-norm based CS. Secondly, we also show that Dantzig selector based CS supresses speckle noise in radar images. We demonstrate the results on both simulated and real radar images.