I've intergrated skimap with ptam.
I like this better because you do not have to do a mod since ptam published pose msg.
I have one issue: it works find when I feed it tiago_bar.bag but if I feed it a bag from other slam dataset
rgbd_dataset_freiburg3_long_office_household_validation.bag even when adjusted for camera parameters the point cloud and map are way off. see pictures below. Any idea why this would be. Is there different about the camera set up . I strip out all but the image and depth messages in both bag files.
My integrations launch file is as follows. note; ptam requires that you set the plane using its gui which will pop up once set it will forward pose.
I've intergrated skimap with ptam. I like this better because you do not have to do a mod since ptam published pose msg. I have one issue: it works find when I feed it tiago_bar.bag but if I feed it a bag from other slam dataset rgbd_dataset_freiburg3_long_office_household_validation.bag even when adjusted for camera parameters the point cloud and map are way off. see pictures below. Any idea why this would be. Is there different about the camera set up . I strip out all but the image and depth messages in both bag files.
My integrations launch file is as follows. note; ptam requires that you set the plane using its gui which will pop up once set it will forward pose.
pose_tf.cpp : code for node
I combined the Parameter files as follows: