Argon Design: Difference between revisions
(Created page with "Contact: 'Steve Barlow' <steve.barlow@argondesign.com> I’m thinking of a project around image processing operations on an FPGA. Something like a corner detector or stereo d...") |
No edit summary |
||
| Line 1: | Line 1: | ||
Contact: 'Steve Barlow' <steve.barlow@argondesign.com> | Contact: 'Steve Barlow' <steve.barlow@argondesign.com> | ||
I’m thinking of a project | I confirm we’d like to put forward a brief for a project for next year (Lent 2018). | ||
My delay in replying was because I was chewing over ideas for possible projects. I’ve still not had an idea I’m really happy with so I thought I’d better reply now and then keep thinking. | |||
An FPGA project is nice because it’s different technology. The downside is it takes more work to develop compared with software and so might not do something as obviously impressive. The reasons you would use an FPGA are if you ultimately want to implement something in hardwired logic for lower power or you need throughput or low-latency that is beyond a software solution. My current thought is a hardware corner detector that could provide input to a software SLAM algorithm. | |||
I know Simon Moore from a number of years ago, when he was working on asynchronous processors. What FPGA boards do you have? I might alternatively provide a board if that is an easier way of connecting a camera and not having too many FPGA resource constraints. | |||
Another possible non-FPGA project direction is “slow cameras”. If you only capture an image every hour, you can spend a long time doing image processing on it on a microcontroller. You don’t need a powerful system. So what could you do to create smart sensors using this? | |||
I’ll continue to think over the next couple of months. | |||
Revision as of 11:33, 21 October 2017
Contact: 'Steve Barlow' <steve.barlow@argondesign.com>
I confirm we’d like to put forward a brief for a project for next year (Lent 2018).
My delay in replying was because I was chewing over ideas for possible projects. I’ve still not had an idea I’m really happy with so I thought I’d better reply now and then keep thinking.
An FPGA project is nice because it’s different technology. The downside is it takes more work to develop compared with software and so might not do something as obviously impressive. The reasons you would use an FPGA are if you ultimately want to implement something in hardwired logic for lower power or you need throughput or low-latency that is beyond a software solution. My current thought is a hardware corner detector that could provide input to a software SLAM algorithm.
I know Simon Moore from a number of years ago, when he was working on asynchronous processors. What FPGA boards do you have? I might alternatively provide a board if that is an easier way of connecting a camera and not having too many FPGA resource constraints.
Another possible non-FPGA project direction is “slow cameras”. If you only capture an image every hour, you can spend a long time doing image processing on it on a microcontroller. You don’t need a powerful system. So what could you do to create smart sensors using this?
I’ll continue to think over the next couple of months.
