Decentration of a lens within an optical system can be used to redirect the path of the axial principal ray, shifting the entire image. This will introduce ‘decentred aberrations’ into the system. This paper will investigate the decentred aberrations of an image intensified system, which uses a decentred lens to achieve image shift to correct boresight errors at assembly. The analysis will be conducted using a decentred aberration theory developed at Canon. This system will be compared to a patent lens with image stabilising capability. It will be shown that an acceptable level of lens shuffle insensitivity can still be achieved in a relatively simple design form. The decentred aberration theory will then be used to suggest alternative design forms.