Odometry with RGBDSLAM for mapping white walls

edit

Hi guys!

I'm trying to map a room with some white walls spaces (no texture) with turtlebot + kinect, and the problem is that the number of keypoints on it makes the robot lost the reference everytime it moves a bit.

I saw this question related to my problem but couldn't see any differences with ICP activated when mapping the wall. I was thinking if that's possible to "merge" two packages like RGBDSLAM and Robot_pose_ekf to build the map with the odometry information.

If the information about the position from the robot induce the matching of point clouds in rgbdslam, i guess it could perfectly map the big white wall.

Since i'm not an expert at ros packages and subscribed topics, i'd like some advice from who know or tried something like this before.

Is there another method could i use to do this?

EDIT:

The loop problem was related to wrong parameters at robot_odometry.launch, specifically at odom_frame_name and base_frame_name.

I'm able to create a map using the odometry information but there's still a problem, the floor contains so many "holes".

.

I'd like to know how can i capture the whole floor without raising the max_depth parameter (raising it would bring me some problems, because up to 3.8 meters at max_depth, the kinect captures ceiling). Is that related to ground_truth algorithm? How it works in rgbdslam?

This is my robot_odometry.launch file:

  <node pkg="rgbdslam" type="rgbdslam" name="rgbdslam" cwd="node" required="true" output="screen"> 
<!-- Input data settings-->
<param name="config/topic_image_mono"              value="/camera/rgb/image_mono"/> <!--could also be color -->
<param name="config/topic_image_depth"             value="/camera/depth_registered/sw_registered/image_rect_raw"/>


<param name="config/feature_extractor_type"        value="ORB"/><!-- also available: SIFT, SIFTGPU, SURF, SURF128 (extended SURF), ORB. -->
<param name="config/feature_detector_type"         value="ORB"/><!-- also available: SIFT, SURF, GFTT (good features to track), ORB. -->
<param name="config/detector_grid_resolution"      value="10"/><!-- detect on a 3x3 grid (to spread ORB keypoints and parallelize SIFT and SURF) -->
<param name="config/max_keypoints"                 value="600"/><!-- Extract no more than this many keypoints -->
<param name="config/max_matches"                   value="300"/><!-- Keep the best n matches (important for ORB to set lower than max_keypoints) -->

<param name="config/min_sampled_candidates"        value="4"/><!-- Frame-to-frame comparisons to random frames (big loop closures) -->
<param name="config/predecessor_candidates"        value="4"/><!-- Frame-to-frame comparisons to sequential frames-->
<param name="config/neighbor_candidates"           value="4"/><!-- Frame-to-frame comparisons to graph neighbor frames-->
<param name="config/ransac_iterations"             value="100"/>
<param name="config/cloud_creation_skip_step"      value="5"/><!-- subsample the images' pixels (in both, width and height), when creating the cloud (and therefore reduce memory consumption) -->

<param name="config/cloud_display_type"            value="POINTS"/><!-- Show pointclouds as points (as opposed to TRIANGLE_STRIP) -->
<param name="config/pose_relative_to"              value="largest_loop"/><!-- optimize only a subset of the graph: "largest_loop" = Everything from the earliest matched frame to the current one. Use "first" to optimize the full graph, "inaffected" to optimize only the frames that were matched (not those inbetween for loops) -->
<param name="config/backend_solver"                value="pcg"/><!-- pcg is faster and good for continuous ...