We demonstrate an efficient approach for correcting spatially varying (anisoplanatic) aberrations in digital holographic imaging by leveraging a Zernike-Fourier domain representation. The imaging operator was modelled in a matrix form as a combination of Fourier basis functions and Zernike decomposed field-dependent wavefront aberrations. The single-step matrix multiplication greatly reduces computational complexity compared to traditionally relying on explicit matrix inversion or point-wise convolution.