The recently proposed super resolution radar imaging algorithms

which are based on the 2-D sinusoid signalmodel

often suffer from the motion through resolution cell error（MTRC） and failed completely. In this paper

an algorithm isproposed based on the 2-D approximate radar imaging model. By minimizing a nonlinear least-squares cost function

the algorithm combines the parameter estimation with the compensation of MTRC errors . The Cramer-Rao bounds are derived andsimulation results are also presented to demonstrate the performance of the algorithm.