[scilab-Users] Quiver on an image (SIVP toolbox)

Stephane Mottelet stephane.mottelet at utc.fr
Wed Jul 14 11:14:14 CEST 2010


Peter von Niederhaeusern <pvonnied at gmail.com> a écrit :

> hold on (or _hold on) uses the "default" plotting context of Scilab
> and plots subsequent plots
> onto the same "canvas"...
>
> But what about different contexts? imshow from the SIVP toolbox uses
> Tk to show images.

I missed that point

> I don't think one can overlay plots from the Scilab context onto Tk, right?
>
> Are there any alternatives for mixing?

no. Scilab should implement truecolor canvases...

S.

>
> I'd like to draw the quiver field onto the image to show the meaning
> of the difference vectors of
> a tracked object (the image).
>
> On Tue, Jul 13, 2010 at 8:13 PM, Stephane Mottelet
> <stephane.mottelet at utc.fr> wrote:
>> Peter von Niederhaeusern <pvonnied at gmail.com> a écrit :
>>
>>> Yes. I've found out the cause for this: the scale factor ;-)
>>> Now the quiver plot looks exactly as in the Matlab example.
>>>
>>> Remaining problem: plot the quiver on the image...
>>
>> use "hold on"
>>
>>>
>>> On Tue, Jul 13, 2010 at 2:30 PM, Stephane Mottelet
>>> <stephane.mottelet at utc.fr> wrote:
>>>>
>>>> Peter von Niederhaeusern <pvonnied at gmail.com> a écrit :
>>>>
>>>>> Hi all,
>>>>>
>>>>> We are to convert scripts from Matlab to Scilab 5.2.2.
>>>>>
>>>>> One task of this conversion is to plot a quiver (we use plotlib)
>>>>> on an image of a moving car (difference image, displacement, object
>>>>> tracking...).
>>>>>
>>>>> The SIVP toolkit uses a Tk widget to display an image and quiver uses
>>>>> internal means of Scilab: is it possible to combine the two?
>>>>>
>>>>> BTW: are there any significant differences between the quiver command of
>>>>> Matlab
>>>>> and the _quiver command of plotlib? Our results differ heavily...
>>>>
>>>> Hi. the graphical results should only differ concerning arrow
>>>> size, not direction. Is it the case ?
>>>>
>>>> S.
>>>>
>>>>>
>>>>> Merci,
>>>>> Peter
>>>>>
>>>>
>>>>
>>>>
>>>
>>
>>
>>
>





More information about the users mailing list