[scilab-Users] real time image processing

[Mike Page]

Hi Alain,

Sorry, but I have no experience of SciCos or Xcos at all.

I also tried Sci2C, but it has far too many limitations for me.  Main one is
that I use a lot of 3-dimensional hypermatrices, but Sci2C does not
apparently support this datatype.

Hopefully someone else on the mailing list can answer about linking C with
SciCos.

Mike.


-----Original Message-----
From: alain [mailto:al.bertho at free.fr]
Sent: 19 December 2011 13:08
To: users at lists.scilab.org
Subject: Re: [scilab-Users] real time image processing


Hello,

Thanks, Mike, for your answer.

I've built a C interface (a lib on a linux target) to read image files I
already have in Scilab. It works fine, and encourage me to follow your
way. I think this will lead to a pure C code at the end, when connected
on the hardware, as my loop will be 100 ms.

But I'm still curious about xcos. There are many things I don't cacth
about the C code generation in xcos, maybe because I don't know which
documentation is relevant, the Scicos one or the Xcos one. What I
understand is that you can call Scilab code within an xcos block, and
then, as far as your process is discrete, generate C code... So there's
a scilab C code generator inside xcos? I've read about the atoms sci2c
module, but its limitations don't stand my needs (fixed size variable).
If someone knows entries on this....

Well, maybe the best is to spend a bit of time to try xcos ...


Le 19/12/11 10:35, Mike Page a écrit :
> Hi Alain,
>
> I have successfully used Scilab with a real-time system, though I have not
> used Scicos.  I also use Scilab to prototype my algorithms before
developing
> the "real" system.
>
> Scilab has a built-in system for compiling your own C code into Scilab
> functions, but I don't use it.  I just use MS Visual C++ to write the C
code
> and build it as a DLL.  Then it's quite easy to declare the entry points
in
> Scilab and call the C functions from within your Scilab code.  There is
help
> available in the documentation, but it is a bit fragmented and also not
all
> relevant to the latest version.
>
> Be aware that Scilab can be quite slow, depending on what you are doing,
and
> it may not be able to keep up with your hardware.  However, I am doing
some
> quite intensive signal processing and my code runs with a 250ms loop time.
>
> When you get to the stage where you are integrating your C code and
Scilab,
> just post any specific questions here.  There are lots of people who will
> help you.
>
> Good luck,
> Mike.
>
>
> -----Original Message-----
> From: alain [mailto:al.bertho at free.fr]
> Sent: 17 December 2011 09:33
> To: users at lists.scilab.org
> Subject: [scilab-Users] real time image processing
>
>
> Hello!
>
> I need to real time process a sonar image flow. The goal is to detect
> object and follow them, as the sonar is moving.
> I'm familiar to scilab, and I'd like to use it to prototype the process,
> and if possible, generate C code for the real time software.
> I've never use xcos, but I'd like to.
> As my process is discrete, do you think xcos is adapted for image
> processing? Is it easy to generate C code and then link it to the rest
> of my application?
>
> Thanks.
>
> Alain
>
>