Abstract: For 1-bit quantization signal, conventional linear processing method can not suppress the harmonic component caused by nonlinear quantization. A joint method is proposed to perform harmonic suppression and target reconstruction in spatial domain based on linear and nonlinear combination. Firstly, based on linear preprocessing and predetection, the main components of the received signal are obtained and the dimension reduction observation model is established. Secondly, based on the generalized approximate message passing (GAMP), dimension reduction signal model is solved so as to suppress harmonic components and reconstruct true targets. To address the grid mismatch problem, the dimension reduction model is modified by estimating the parameters of predetected targets. Simulation results show that the proposed method can reconstruct true targets and effectively suppress the spatial harmonics caused by 1-bit quantization.