Image dimension for Hyperspectral data cube

As far as I have understood, hyperspectral cube consists of various images stacked one over the other. The dimensions of each of these image is 2D (say m X n). Whenever I try to take some pictures to stack them in a cubical manner, each of the pictures are represented by a 3D matrix (mxnx3). What can I do to convert these 3D matrix into a 2D matrix and still view the converted 2D matrix in colour ?