The Propagator Method (PM) replaces the eigenvalue decomposition with linear operation, in which the computational complexity is reduced but traversing search within the whole space spectrum is still needed, and the estimation performance degrades seriously under low SNR. In this paper, an Image Compression PM (IC-PM) algorithm is presented on the basis of the PM algorithm. The IC-PM algorithm uses the orthogonal characteristics of the propagator and guide vector, and constructs the real source in one dimensional space image of the source, which is equivalent to the entire spatial spectrum in the compression. Theoretical analysis and simulation experiments show that the IC-PM algorithm can reduce the computational complexity to half of the PM algorithm, and increase the estimation accuracy under the condition of low SNR.