| ARToolKit | Mailing List Archive |
|
| From: | "Victor Ng-Thow-Hing" <victorngthowhing@y ........> | Received: | Jan 30, 2007 |
| To | <artoolkit@h ..................> | ||
| Subject: | TrackerMultiMarker with non-coplanar markers? | ||
This is a multi-part message in MIME format. ------=_NextPart_000_0006_01C74480.D78B6D10 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Hi, I'm interested in tracking 3D cubes with markers on all sides of the = cube and using TrackerMultiMarker to robustly track the cube. In all the examples I've seen, the multiple markers are always arranged = in a tiles on the same plane. I noticed we can specify a 4x4 matrix for each marker in the set. Will the ARToolKitPlus library actually handle markers which may be at translational and rotational offsets to each other? or will this = multiple marker tracker only work with translational offsets? =20 Regards, Victor. ------=_NextPart_000_0006_01C74480.D78B6D10 Content-Type: text/html; charset="us-ascii" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; = charset=3Dus-ascii"> <TITLE>Message</TITLE> <META content=3D"MSHTML 6.00.5730.11" name=3DGENERATOR></HEAD> <BODY> <DIV><FONT face=3DArial size=3D2><SPAN=20 class=3D421020823-30012007>Hi,</SPAN></FONT></DIV> <DIV><FONT face=3DArial size=3D2><SPAN class=3D421020823-30012007>I'm = interested in=20 tracking 3D cubes with markers on all sides of the cube and using=20 TrackerMultiMarker to robustly track the cube.</SPAN></FONT></DIV> <DIV><FONT face=3DArial size=3D2><SPAN class=3D421020823-30012007>In all = the examples=20 I've seen, the multiple markers are always arranged in a tiles on the = same=20 plane. I noticed we can specify a 4x4 matrix for each marker in the=20 set.</SPAN></FONT></DIV> <DIV><FONT face=3DArial size=3D2><SPAN class=3D421020823-30012007>Will = the=20 ARToolKitPlus library actually handle markers which may be at = translational and=20 rotational offsets to each other? or will this multiple marker tracker = only work=20 with translational offsets?</SPAN></FONT></DIV> <DIV><FONT face=3DArial size=3D2><SPAN=20 class=3D421020823-30012007></SPAN></FONT> </DIV> <DIV><FONT face=3DArial size=3D2><SPAN=20 class=3D421020823-30012007>Regards,</SPAN></FONT></DIV> <DIV><FONT face=3DArial size=3D2><SPAN=20 class=3D421020823-30012007>Victor.</SPAN></FONT></DIV></BODY></HTML> ------=_NextPart_000_0006_01C74480.D78B6D10-- |
|||
| From: | Daniel Wagner <daniel@i ...............> | Received: | Jan 31, 2007 |
| To | ARToolKit Mailinglist <artoolkit@h ..................> | ||
| Subject: | Re: TrackerMultiMarker with non-coplanar markers? | ||
Victor Ng-Thow-Hing wrote: > Will the ARToolKitPlus library actually handle markers which may be at > translational and rotational offsets to each other? or will this > multiple marker tracker only work with translational offsets? The algorithm ARToolKit(Plus) uses is restricted to planar targets. To my knowledge this also includes the multi-marker part. Daniel |
|||
| From: | Daniel Wagner <daniel@i ...............> | Received: | Jan 31, 2007 |
| To | ARToolKit Mailinglist <artoolkit@h ..................> | ||
| Subject: | Re: TrackerMultiMarker with non-coplanar markers? | ||
Debadeepta Dey wrote: > I wanted to ask whether the multimarker tracking in ARToolkitPlus > works the same as in ARToolkit? Or have any extra optimisations been > done for this purpose? ARToolKitPlus has an additional pose estimator which is often more robust as the original one. Still that pose estimator is limited to planar targets too. Daniel |
|||
| From: | Hartmut Seichter <hartmut@t ................> | Received: | Jan 31, 2007 |
| To | ARToolKit Mailinglist <artoolkit@h ..................> | ||
| Subject: | Re: TrackerMultiMarker with non-coplanar markers? | ||
Yes, I think the case Victor is describing should work for ARToolkit and ARToolkitPlus taking granted none of the markers will bent. You can provide for each marker within a multimarker configuration file a 4x3 matrix describing translation and rotation (see also /bin/multi/ marker.dat) /Hartmut On 31/01/2007, at 9:10 PM, Debadeepta Dey wrote: > I have read the RPP paper and have tested it for a tracking > application I am developing . It indeed is more robust than > ARToolkit's incumbent algorithm. But as i understand that the only > planarity constraint on RPP as well as on ARToolkit's algo is that a > single marker should be coplanar ( ie it should not be warped so that > the 4 marker vertices do not lie on the same plane) but I do not see > any constraint on having a global coordinate system ( such as for the > cube tracking which Victor has with different markers being at both > rotational and translational offsets to ech other) and markers strewn > across arbitrarily in space but referenced by the same global > coordinate system. > > Please correct me if i am wrong but I beleive that the > marker_480-499.cfg can be easily modified for rotational offsets too . > > > > On 1/31/07, Daniel Wagner <daniel@i ...............> wrote: >> Debadeepta Dey wrote: >> > I wanted to ask whether the multimarker tracking in ARToolkitPlus >> > works the same as in ARToolkit? Or have any extra optimisations >> been >> > done for this purpose? >> ARToolKitPlus has an additional pose estimator which is often more >> robust as the original one. >> Still that pose estimator is limited to planar targets too. >> >> Daniel >> >> > > > -- > Debadeepta Dey, > Research Intern, > Hitech Robotics Systemz, > India. |
|||
| From: | "Debadeepta Dey" <3dey.dce@g ........> | Received: | Jan 31, 2007 |
| To | "Daniel Wagner" <daniel@i ...............> | ||
| Subject: | Re: TrackerMultiMarker with non-coplanar markers? | ||
I wanted to ask whether the multimarker tracking in ARToolkitPlus works the same as in ARToolkit? Or have any extra optimisations been done for this purpose? On 1/31/07, Daniel Wagner <daniel@i ...............> wrote: > Victor Ng-Thow-Hing wrote: > > Will the ARToolKitPlus library actually handle markers which may be at > > translational and rotational offsets to each other? or will this > > multiple marker tracker only work with translational offsets? > The algorithm ARToolKit(Plus) uses is restricted to planar targets. > To my knowledge this also includes the multi-marker part. > > Daniel > > -- Debadeepta Dey, Research Intern, Hitech Robotics Systemz, India. |
|||
| From: | "Debadeepta Dey" <3dey.dce@g ........> | Received: | Jan 31, 2007 |
| To | "Daniel Wagner" <daniel@i ...............> | ||
| Subject: | Re: TrackerMultiMarker with non-coplanar markers? | ||
I have read the RPP paper and have tested it for a tracking application I am developing . It indeed is more robust than ARToolkit's incumbent algorithm. But as i understand that the only planarity constraint on RPP as well as on ARToolkit's algo is that a single marker should be coplanar ( ie it should not be warped so that the 4 marker vertices do not lie on the same plane) but I do not see any constraint on having a global coordinate system ( such as for the cube tracking which Victor has with different markers being at both rotational and translational offsets to ech other) and markers strewn across arbitrarily in space but referenced by the same global coordinate system. Please correct me if i am wrong but I beleive that the marker_480-499.cfg can be easily modified for rotational offsets too . On 1/31/07, Daniel Wagner <daniel@i ...............> wrote: > Debadeepta Dey wrote: > > I wanted to ask whether the multimarker tracking in ARToolkitPlus > > works the same as in ARToolkit? Or have any extra optimisations been > > done for this purpose? > ARToolKitPlus has an additional pose estimator which is often more > robust as the original one. > Still that pose estimator is limited to planar targets too. > > Daniel > > -- Debadeepta Dey, Research Intern, Hitech Robotics Systemz, India. |
|||
| From: | Tijs de Kler <tijs@k ............> | Received: | Jan 31, 2007 |
| To | ARToolKit Mailinglist <artoolkit@h ..................> | ||
| Subject: | Re: TrackerMultiMarker with non-coplanar markers? | ||
Hey, I have done some testing with a cube in ARTK+. (cant remember is it was RPPP or oringinal) It can track the position and orientation of the cube's centre, but sometimes it does however give wrong estimations, where it estimates the cube centre outside of the cube. My guess is that it is something buggy in the algorithm. But about 95% of the frames it gives a good pose of the cube. regards, Tijs On Wed, 2007-01-31 at 09:29 +0100, Hartmut Seichter wrote: > Yes, I think the case Victor is describing should work for ARToolkit > and ARToolkitPlus taking granted none of the markers will bent. You > can provide for each marker within a multimarker configuration file a > 4x3 matrix describing translation and rotation (see also /bin/multi/ > marker.dat) > > /Hartmut > > > > > On 31/01/2007, at 9:10 PM, Debadeepta Dey wrote: > > > I have read the RPP paper and have tested it for a tracking > > application I am developing . It indeed is more robust than > > ARToolkit's incumbent algorithm. But as i understand that the only > > planarity constraint on RPP as well as on ARToolkit's algo is that a > > single marker should be coplanar ( ie it should not be warped so that > > the 4 marker vertices do not lie on the same plane) but I do not see > > any constraint on having a global coordinate system ( such as for the > > cube tracking which Victor has with different markers being at both > > rotational and translational offsets to ech other) and markers strewn > > across arbitrarily in space but referenced by the same global > > coordinate system. > > > > Please correct me if i am wrong but I beleive that the > > marker_480-499.cfg can be easily modified for rotational offsets too . > > > > > > > > On 1/31/07, Daniel Wagner <daniel@i ...............> wrote: > >> Debadeepta Dey wrote: > >> > I wanted to ask whether the multimarker tracking in ARToolkitPlus > >> > works the same as in ARToolkit? Or have any extra optimisations > >> been > >> > done for this purpose? > >> ARToolKitPlus has an additional pose estimator which is often more > >> robust as the original one. > >> Still that pose estimator is limited to planar targets too. > >> > >> Daniel > >> > >> > > > > > > -- > > Debadeepta Dey, > > Research Intern, > > Hitech Robotics Systemz, > > India. > |
|||