Abstract : In this paper, we focus on the global optimization of the satellite imaging chain. The theoretical analysis of the satellite imaging chain optimization is a difficult problem that needs lot of approximations. In order to consider the complex real satellite imaging chain, we propose to address this problem numerically and we present, based on numerical experiments, techniques to optimize the quality of the reconstructed final image. We first focus on the common question of the position of the restoration step in the imaging chain, that is on-board before coding or on-ground after coding. Then, we present several methods to remove the coding artifacts inherent in wavelet based coder schemes. From these numerical results we propose a new satellite imaging chain and we show visual and rate-distortion results on a real satellite image.