Questions about upcoming RT1060 OpenMV Module

OpenMV Boards
1

Hello,

I’ve been skimming this forum and you guys are providing amazing support for your dev boards and software.

I’m hoping to deploy your code on the RT1170 dev platform.

  1. Do you think it’s realistic that solely the M4 core could handle barcode reading/ simple blob detection?

  2. Do you have an ETA on when the RT1060 codebase is going to be released?

  3. Any reasons using the 1170 might be a mistake vs 1060?

2

Hi!

  1. Yes, we plan to support the RT1170. However, cash flow wise, I can’t comment on when we have the ability to do that. As for the M4. No point, the M7 runs circles around it. Note that our software runs on the STM32F427… which was our original OpenMV Cam. So, running code on the M4 is not a problem per-say.
  2. When the product comes out. Ibrahim is not interested in fielding help support questions about people trying to run the code on IMX dev boards at the moment.
  3. RT1170 chip prices are $22 a chip. It’s no joke to build anything with them. It’s more than 2x what the 1060 is. So… that makes it hard to design around it. Otherwise, it’s fine.
3

Not sure how complicated this follow up is, but now I’m using an RT1170 dev kit.

I was wondering how hard it would be to extract the source code that does barcode reading from the OpenMV code?

4

Hi, Ibrahim released the first copy of the port to IMX online this week.

However, I want to make this clear… we will provide NO SUPPORT for the dev kits. You are on your own fighting issues making those work.

We do have tentative plans to make something based on the RT1170. However, we have to get the RT1060 out the door first, better camera modules, and new shields before I can think about that chip.

As for your question:

I was wondering how hard it would be to extract the source code that does barcode reading from the OpenMV code?

You’ll need to extract most of the image library code from our code base. So, it’s going to be quite a few C files. It’s not hard per say. But, it will take a day of grabbing code, compiling, grabbing more code to fix compile errors, and etc. until you’ve extracted everything.

5

hi, as nxp is a sponsor of National College Student Intelligent Car Competition(全国大学生智能汽车竞赛) in china, there already are some projects using nxp rt106x running openmv, like OpenART mini by seekfree and micropython-rocky by an nxp engineer. is it to be welcomed that port then to openmv mainline?

6

Unfortunately… that split off as a port more than 4 years ago. So, it’s code base is pretty incompatible.

We worked with Rocky back then, but, NXP didn’t invest in having him port MicroPython to the CPU so we couldn’t support it. What allowed us to target the RT1060 now is that MicroPython supports it.

Anyway, now that we having a working image for the RT1060 we could definitely modify our image to support the PCB boards. OpenART would then need to submit PR’s to our repo to enable the firmware on their boards. Since we’re not selling those units we won’t spend time on supporting them. However, we’re happy to take PRs that enable us to support them in our firmware mainline and in the IDE.

As for the OpenART project itself. I don’t know what should be done with the code base. Probably the best thing to do would be for them to send PRs enabling their hardware with the mainline.

7

ok, i would start this work after about a half year or more to avoid effecting the sales of openmv5, if no one will have done this at that time.

1 Like