3D transforms added

[gestom]

Set axis now saves 3D transform information and whycon uses it:).

[protobits]

Hi!
Nice, I'll try it out. I see it is a big change, however there are lots of whitespace changes which I guess was introduced by our text editor.

From what I understand, this would still require four circles during calibration when only three could be used for 3D case, right? Maybe later this could be split into two cases to simplify calibration.

[protobits]

any particular reason for this? c++11 is supposed to be more standard. Moreover, I read gnu++0x flag is deprecated in favor of c++11.

[gestom]

Hi Matias,
actually, I would leave the four circles - the overall 3D transform can be composed of four 3D ones, that use three circles each. Calculating marker position from 4 transformations is a bit more precise + you can get position uncertainty.
Regarding c++11: I got a compiler error. When I removed the c++11 flag, catkin proposed to add gnu++0x one. After adding it, compilation went fine. Petr Cizek, who tested the software, was reporting the same issue.
Cheers,
Tom

[protobits]

Ah, ok. But is that what is implemented now? Or that could be later added?

Ok, let's leave that flag then. I suspect it is due to older compilers. May
I ask which gcc version have you tried that has this problem?

On Mon, Sep 28, 2015 at 8:49 AM, Tom Krajnik [email protected]
wrote:

Hi Matias,
actually, I would leave the four circles - the overall 3D transform can be
composed of four 3D ones, that use three circles each. Calculating marker
position from 4 transformations is a bit more precise + you can get
position uncertainty.
Regarding c++11: I got a compiler error. When I removed the c++11 flag,
catkin proposed to add gnu++0x one. After adding it, compilation went fine.
Petr Cizek, who tested the software, was reporting the same issue.
Cheers,
Tom

—
Reply to this email directly or view it on GitHub
#6 (comment).

[Praneet1997]

there is no service as stated in wiki set_axis/set
anyone can solve the problem