The Conventional Compressive Sensing (CSS) which uses non-adaptive projection for the representation of natural images shows inefficient compression performance when compared to JPEG and JPEG 2000 which are the classical image compression standards. This paper investigates one of the methods of Block Compressive Sensing (BCS) called as coefficient random permutation (CRP). The effectiveness of the CRP method lies in balancing the sparsity of the sampled vectors the images DCT domain and in improving the efficiency of Compressive Sensing Sampling. The analysis shows that the proposed method is effective in reducing the BCS based image representation dimension and to improve the quality of the recovered image. The proposed method replaces the robust image compression and applications of the encrypted image compressions.