Ask Your Question

Revision history [back]

Do you have a (valid) camera_info along with the images from the Kinect?

If so then https://answers.ros.org/question//project-point-cloud-to-image-coordinates-using-camerainfo/.

I don't know if there is a complete end to end example anywhere but looking up project3dToPixel produces a lot of results. If I get around to it I'll update this answer with one or link to another answer that does (or anyone else is welcome to do that...)

If the projected point output is within the resolution of the camera (between 0,0 and image width/height in pixels) then it should be visible, probably an additional check is required to make sure the object is in front of the camera and not behind if- if the z value is positive or not.

If you still are interesting in kinect angles-of-view you can calculate them from the fx/fy parameters in the camera_info. https://stackoverflow.com/questions/39992968/how-to-calculate-field-of-view-of-the-camera-from-camera-intrinsic-matrix

Do you have a (valid) camera_info along with the images from the Kinect?

If so then https://answers.ros.org/question//project-point-cloud-to-image-coordinates-using-camerainfo/.

I don't know if there is a complete end to end example anywhere but looking up project3dToPixel produces a lot of results. If I get around to it I'll update this answer with one or link to another answer that does (or anyone else is welcome to do that...)

If the projected point output is within the resolution of the camera (between 0,0 and image width/height in pixels) then it should be visible, probably an additional check is required to make sure the object is in front of the camera and not behind if- if the z value from the transformed point is positive or not.

If you still are interesting in kinect angles-of-view you can calculate them from the fx/fy parameters in the camera_info. https://stackoverflow.com/questions/39992968/how-to-calculate-field-of-view-of-the-camera-from-camera-intrinsic-matrix

Do you have a (valid) camera_info along with the images from the Kinect?

If so then https://answers.ros.org/question//project-point-cloud-to-image-coordinates-using-camerainfo/.answer.ros: "project pointcloud to image".

I don't know if there is a complete end to end example anywhere but looking up project3dToPixel produces a lot of results. If I get around to it I'll update this answer with one or link to another answer that does (or anyone else is welcome to do that...)

If the projected point output is within the resolution of the camera (between 0,0 and image width/height in pixels) then it should be visible, probably an additional check is required to make sure the object is in front of the camera and not behind if- if the z value from the transformed point is positive or not.

If you still are interesting in kinect angles-of-view you can calculate them from the fx/fy parameters in the camera_info. https://stackoverflow.com/questions/39992968/how-to-calculate-field-of-view-of-the-camera-from-camera-intrinsic-matrix