Re: Further Processing of camera images & RTAB data -- best way to do it?

Posted by ibd on
URL: http://official-rtab-map-forum.206.s1.nabble.com/Further-Processing-of-camera-images-RTAB-data-best-way-to-do-it-tp3015p3031.html

Hi again!

You might also be interested in YOLO 9000, which can detect many object classes! However there is currently no ROS node available for it.

1. & 4. Yes, in fact I am on the same tango setup as described in my other thread. Thank you, I will look into this optimization once I need it.

2. Since I am using tango, I can cache the TF as you described. Since the delay is rather small (<1s), I'm just assuming that there is usually no loop closure that would change a TF in the past. If it is necessary, I will check out your interpolation example.

3. Yes, exactly. Thank you for your example. I think in the end it will turn out to be an exercise in projection between image and camera coordinates, and in putting together the correct transformations to go between the camera frame to the map frame.

Does RTAB-map offer functionality to take the depth image and the "regular" image (used for odometry & loop closure) from different cameras? Asking because tango offers a fisheye camera with a very wide field of view, it might be nicer for loop closure than the color camera (which is the same as the depth camera, by using the infrared spectrum).