Abstract: To improve the range resolution of a synthetic aperture radar (SAR) system, one of the effective technological approaches is to increase signal bandwidths via multi-channel synthesis. In view of the issue of channel phase mismatch among multi-channels, a method based on the principle of entropy minimization is proposed to estimate and compensate for channel phase error, which is modeled as a polynomial in the frequency domain. By adopting the image information entropy of a range compressed pulse as the objective function and taking the polynomial coefficients as the estimated variables, an optimal estimation model for channel phase errors is established based on the principle of entropy minimization. Characterized by independence from the surface feature of the imaging scene, this method effectively makes up for the insufficiency of internal or external calibration; meanwhile, it just requires a small amount of raw data as the sample for error estimation, which results in the advantage of no loss of SNR, few storage consumption and fast convergence speed. The validity and robustness of the method are demonstrated by the experimental results of eight-channel raw data processing from different scenes.

Key words: multi-channel SAR system, phase error, entropy minimization, estimation, compensation