I use image() a lot for quickly getting an overview of what a 2-d array contains. I have searched through the Volume visualization guides without finding anything similar for 3-d arrays, so I have made one myself.

的方法is very simple: make a "box" out of 6 patches centered at each datapoint, and multiply each walls opacity by the corresponding array elements numeric value.

Caveats:
1. As the number of elements grows, the memory usage and plotting system cpu usage quickly gets too large
2. This is a crude approximation of "voxels", but the amount of blocked light is proportional to the number of wall transitions, not the real distance.