The super resolution method is investigated for use in instantaneous invers
e synthetic aperture radar (ISAR) imaging. The time-varying Doppler spectru
m is modelled as that of superimposed LFM chirp signals in a short time int
erval. A parametric relaxation-based algorithm is proposed for estimating t
he parameters of the signals. The effectiveness of the algorithm is verifie
d by applying it to ISAR imaging.