This paper deals with the optimization of sensing matrices and sparsifying dictionaries for compressed sensing systems. A gradient-based method with a new measurement strategy denoted as real mutual coherence is proposed. Further more, the sensing matrix is optimized by minimizing an objective function in which the target Gram is selected as Ψ Ψ, this choice has advantages to reconstruct real images more accurately and efficiently. Experiments are carried out and the results show that the sensing matrix obtained using the proposed approach outperforms the existing ones in terms of signal reconstruction accuracy for real images.