Aperture synthesis imaging has been proved to be attractive in surveillance and detection applications. Such an imaging process is inevitably subject to aberrations introduced by instrument defects and/or turbulent media. Redundant spacing calibration (RSC) technique allows continuous calibration of these errors at any electromagnetic wavelength. However, it is based on specially designed array, in which just enough redundancy is included to permit the successful implementation of RSC. A new design criterion for linear RSC array is described, which introduces coverage efficiency and redundancy efficiency factors, aiming to find the perfect configurations, which have as complete uv -plane coverage as possible while containing required redundancy. Optimum linear arrays for N (number of subapertures) up to 10 are listed based on simulated annealing algorithm. The comparisons with existing linear RSC arrays with equivalent subaperture number are implemented. Results show that the optimized arrays have better performance of both optical transfer function, point spread function, and object reconstruction with reasonable value of the matrix condition number. After that, linear arrays are used to construct two-dimensional (2-D) pseudo-Y-shaped RSC arrays, which give a way to design 2-D RSC arrays without exhaustive searches.