I am interested in doing some prototyping using Wolfram Language and the Pi. In general, I need to be able to capture hardware-triggered single images on the camera board and then perform some sophisticated image processing. I would prefer being able to work from a Mathematica front end. I can think of a few possibilities for connecting all this, but I have never used the Pi and would really appreciate some advice.
1) Connect a PC (Windows 7 x64) to the Pi by USB, Bluetooth or wireless. Run the Mathematica front end on the windows PC and the kernel on the Pi. If I understand things correctly, the kernel will have access to the Pi hardware, but the PC will have a Mathematica front end with all the advantages of the notebook front end. The disadvantage is that the kernel on the Pi is performing all the computations, and is far less powerful than the PC.
2) Use the same setup as in #1 to develop the hardware-triggered image capture on the Pi in Wolfram Language, but then allow the Pi to run that Wolfram code as a headless process. Somehow connect Mathematica running both kernel and front end on the PC to the Pi, using the Pi only to trigger and perform the capture, notify the PC, and participate in the transfer of image data to Mathematica on the PC, where the PC resident kernel can perform the analysis.
3) Use a scheme similar to #2, except program the Pi in Python. This still needs a method to communicate with the Pi from Mathematica running on the PC, preferably by USB, Bluetooth or wireless.
One additional issue is that I will want to remove the lens from the camera module to allow a custom optic to be used.
Any and all comments on this are very welcome.