0
$\begingroup$

Rosanswers logo

I plan to try RGBDSlam on a pre recorded data. For that I have written a publisher which reads my data from text files and publishes. However the result that I get is not very good and also I have seen that depth image seen on RGBDSlam is only black and white.

I am publishing:

  • /camera/rgb/image_mono which is a grayscale image sent through cvbridge

  • /camera/depth/image, which is a IPL_DEPTH_32F IplImage, filled with depth values (in meters) and again sent through cvbridge

  • /camera/rgb/points which is generated as a vector of PointXYZRGB

The point cloud and mono image are ok, I can see them clearly in RGBDSlam, however seeing depth image as I have mentioned makes me suspect that I am missing something. Am I?

Thanks..


Originally posted by paul simmons on ROS Answers with karma: 11 on 2011-09-02

Post score: 0

$\endgroup$

1 Answer 1

0
$\begingroup$

Rosanswers logo

The depth image is only black and white; the grayscale values map to depths. RGBDSlam also works with the grayscale version of the RGB image (because that's what the visual feature library needs).

See this question for some more details of what RGBDSlam expects as input.

Out of curiosity, why are you storing your data in text files, rather than using rosbag?


Originally posted by Mac with karma: 4119 on 2011-09-02

This answer was ACCEPTED on the original site

Post score: 1

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.