Content or Square Detection ?

Feb 2, 2011 at 2:21 PM

Hello,

I was trying to understand the SLARToolkit. It seems that the Engine is detecting only the square boarders, not the REAL CONTENT inside the marker(black borders).

I tried to use different PAT files (example: a marker with Alphabet '1' inside the black borders). In this case the SLARToolkit is detecting the markers & adjusting the image to the respective projection. 

But The Problem is, if i use the default SLAR Marker PDF and any new marker (with Alphabet '1' inside the black borders) PAT file, THEN ALSO the SLARToolkit is detecting the markers & adjusting the image to the respective projection !!!

And my understanding is that the engine detects the square boarders not the content inside the black boarders.

So, my question is that; Is SLARToolkit is meant for detecting only SQUARE BOARDERS? not to detect the content inside the black borders??

Thanks alot for your time

-Arun

Coordinator
Feb 2, 2011 at 3:19 PM

It detects all black rectangles and then provides the results, which contain the patterns that were detected. You have to check the Confidence property of the result.

This was discussed a couple of weeks ago. Please read this Thread for details. http://slartoolkit.codeplex.com/Thread/View.aspx?ThreadId=243160

 

- Rene Schulte

Feb 2, 2011 at 5:52 PM

Hey Rane,

Thanks man that thread helped me alot to understand the SLARToolkit... :) 

Thanks for your immediate response !

-Arun

 

Feb 9, 2011 at 3:56 PM

Hello Rene,

I tested with up to 90 markers with proper (25%) border thickness and SLARToolkit worked perfectly.

 

From this thread (http://slartoolkit.codeplex.com/Thread/View.aspx?ThreadId=228342) it is clear that the SLARToolkit only supports square markers with the black border.

I was doing some r&d over that.

SLARToolkit was able to detect markers even though the border thickness was reduced up to 3.125 %. 

But when i increased the distance between the camera and the marker, it could not detect the marker like it did for the normal marker with proper (25%) border thickness.

Will you please suggest me a solution, so that the markers with very thin borders can also be detected properly.

Thanks for your time,

-Arun