ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange
Ask Your Question
2

can't detect object model

asked 2012-11-25 02:04:03 -0500

salma gravatar image

updated 2012-11-25 16:36:37 -0500

Hi all, i made 3D object model with roboearth stack like in this tutorial http://www.ros.org/wiki/roboearth/Tutorials/Record%20an%20object%20model

when i tried to detect this object using

rosrun re_kinect_object_detector re_kinect

and

rosrun re_object_detector_gui detect

it begin to read the model then received cloud points and open the detect gui

but No detection happened! No error just receiving point clouds for a long time !! No red mask on the object like in this picture :

image description

edit retag flag offensive close merge delete

Comments

salma gravatar image salma  ( 2012-11-25 16:33:11 -0500 )edit

1 Answer

Sort by ยป oldest newest most voted
0

answered 2012-11-25 21:08:31 -0500

ddimarco gravatar image

Make sure that the camera is ~60 cm away from the object, from your screenshot it looks a bit too close. To get some visual debug information, you can uncomment the "#define DEBUG_VIS" line in re_kinect_object_detector/src/recognitionmodel.h. This will show you the images from the model, the SURF features and possible matches from the model as lines.

edit flag offensive delete link more

Comments

I have uncommented the "#define DEBUG_VIS" line in re_kinect_object_detector/src/recognitionmodel.h. and the camera now is 60 cm away from the object, but still have the same problem .

salma gravatar image salma  ( 2012-11-25 23:06:50 -0500 )edit

DEBUG_VIS does not really change anything, it is just there to help you find problems. Please note that we use 2D features only, so maybe the coke bottle does not provide enough features. What happens if you try it with another, more textured object?

ddimarco gravatar image ddimarco  ( 2012-11-25 23:12:22 -0500 )edit

I already have tried my own object model but no detection occurred ,so i have downloaded the object model of the coke bottle from roboearth.org but the kinect didn't detect it too ,so i think that the problem isn't with the object model and i can't find object models of products in egypt

salma gravatar image salma  ( 2012-11-26 01:12:17 -0500 )edit
1

Ok, it is hard to figure out the problem like this. Well textured objects like the one at http://api.roboearth.org/data/objects/beverage/icetea/preview.png should work very well. So all advice I can offer is to record an object model for something like that and try it again.

ddimarco gravatar image ddimarco  ( 2012-11-26 01:23:46 -0500 )edit

@ddimarco thanks alot it works ! ,, but i would ask about objects that have One color or little textured How can i detect it ? ? and i thought that cloud points features are width and height and color but it detected a similar object Not exactly it !!?

salma gravatar image salma  ( 2012-11-26 23:49:12 -0500 )edit
1

As I said, the software only uses 2D image features. You'll need to use another package if you require to detect single-colored objects.

ddimarco gravatar image ddimarco  ( 2012-11-27 00:21:00 -0500 )edit

Ok ,, thanks :)

salma gravatar image salma  ( 2012-11-27 00:23:06 -0500 )edit

Question Tools

1 follower

Stats

Asked: 2012-11-25 02:04:03 -0500

Seen: 270 times

Last updated: Nov 25 '12