A Color Images Steganography Method by Multiple Embedding Strategy Based on Sobel Operator

In this paper, we propose a color images steganography. Firstly, because modification of pixel of image edges is less sensitive to human eye, the edge or texture in one plane is detected by Sobel edge detection operator and the largest certain number of gradients are selected as embedding positions. Secondly, because of strong relevance in gradient among R, G, and B planes, the corresponding LSB (least significant bit) of pixel values to hide the secret data in other planes is modified. Finally, combine the stego-planes and restore the color image. Furthermore, multi-times embedding can be adapted to obtain high data capacity. Experimental results show that stego-images keep good quality and high capacity. For example, there is no noticeable degradation after three times hiding process, the average PSNR is about 45 db and data capacity is 6.3 bpp. The method also can extract the secret data accurately.