An efficient nonlinear contrast source inversion scheme for electromagnetic imaging of sparse two-dimensional investigation domains is proposed. To avoid generating a sequence of linear sparse optimization problems, the non-linearity is directly tackled using the nonlinear Landweber (NLW) iterations. A self-adaptive projected accelerated steepest descent (A-PASD) algorithm is incorporated to enhance the efficiency of the NLW iterations. The algorithm enforces the sparsity constraint by projecting the result of each steepest descent iteration into the L 1 -norm ball and selects the largest-possible iteration step without sacrificing from convergence. Numerical results, which demonstrate the proposed scheme's accuracy, efficiency, and applicability, are presented.INDEX TERMS Contrast source inversion, electromagnetic imaging, inverse problems, Landweber iterations, nonlinear inverse scattering, sparse reconstruction, steepest descent algorithm.