In this paper we address the problems of flux and rotor resistance estimation of induction motors. We propose a radically new approach that combines the recently introduced techniques of parameter estimation based observers (PEBO) with the dynamic regression extension and mixing (DREM) parameter adaptation. The PEBO framework is used to recast the flux observation task as a parameter estimation problem, for which the DREM technique is utilised. The resulting flux observer is then combined with a standard gradient estimator for the rotor resistance. Simulation results of an adaptive implementation of the classical field oriented controller demonstrate the effectiveness of the proposed flux observer and rotor resistance estimator even in closed-loop operation.