-2
\$\begingroup\$

I'm building an Indoor navigation robot which can be used in offices and factories especially in closed spaces with good lighting. I've already found out the algorithm i need to use for Indoor navigation from an expert Check the comments of the blog if you like to know. Basically, I need to connect a stereoscopic camera or a single webcamera to a microprocessor or FPGA. Now the question is which one is better? I know FPGA is good for it's parallel programming(sounding noob) capabilities but Microprocessor is easier to use and implement OpenCV in Linux distro< in emdebian or simply Ubuntu > (OpenCV in Python or anyother language). But, i'm an ECE Senior and i really donno why i should see FPGA vs Microprocessor for implementing this program(i know a bit of HDL coding especially in Verilog i.e., i can do coding at a behavioral level to a good extent but not logical level). Or should i go for combined FPGA and Microprocessor kit for doing this. Microprocessor i will be using is ARM based boards like Raspberry Pi or Beagleboard or Pandaboard depending on the requirements.

Please help me sort out the right tool. And also suggest me which language is better Python or C++ or others(please mention).

Thanks.

Edit 1: This is an Embedded system.

\$\endgroup\$
3
  • 2
    \$\begingroup\$ IMHO doing computer vision on a FPGA is more complicated than doing it with a microprocessor. I'd go with the Raspberry Pi + OpenCV, both of which have very good support for Python. \$\endgroup\$
    – m.Alin
    Aug 17, 2012 at 17:39
  • \$\begingroup\$ If you want help, you would do well to fix capitalization and put links to the referenced software. And ask the main question, not tacking on Python vs. C++. \$\endgroup\$ Aug 17, 2012 at 18:26
  • \$\begingroup\$ Start with software on a processor, namely the PC you use for development. If you figure out exactly the algorithm you need to run, and can't do it as fast as you want, then would be the time to consider an FPGA. If you find that your algorithm works in software with the PC loafing along, then you can consider using a lower powered embedded board. \$\endgroup\$ Aug 17, 2012 at 22:59

2 Answers 2

2
\$\begingroup\$

Based on your question, I believe this is over your head. Unless you get a pre-made kit that includes the camera, there is a good deal of work to connect it to a FPGA. And computer vision is a difficult topic in and of itself.

\$\endgroup\$
1
  • 2
    \$\begingroup\$ +1 Image processing is not a trivial subject. \$\endgroup\$
    – Matt Young
    Aug 17, 2012 at 17:41
0
\$\begingroup\$

If that is factory navigation , you could simply do this with a technology like LIDAR.

http://en.wikipedia.org/wiki/LIDAR

I think FPGA system is a overkill for such a thing.Because domain is inside a factory.

\$\endgroup\$
1
  • 1
    \$\begingroup\$ you only need something like bird eye aren't you? \$\endgroup\$ Aug 17, 2012 at 17:47

Not the answer you're looking for? Browse other questions tagged or ask your own question.