OpenGR How to set parameters more reasonably when using the data collected by Kinect

Hello. When I use the data collected by Kinect in the OpenGR data set, the matching effect is always unsatisfactory. How do you generally set the parameters when you use the corresponding data?

Sep 03 '21 03:09 edwinlu-hdu

Hi, Did you follow these recommendations ? https://storm-irit.github.io/OpenGR/a00012.html

Sep 03 '21 08:09 nmellado

Hi， I follow these recommendations ，but when I transform the first frame and the second frame in Kinect/Guy, the best situation is as shown in the figure below: 2021-09-03 11-52-58 screenshot

Sep 03 '21 08:09 edwinlu-hdu

The parameters at this time are as follows : -o 1 -d 0.01 -n 10000

Sep 03 '21 08:09 edwinlu-hdu

Ok, I’ll test and see if I can reproduce this behavior. Can your please detail your configuration ? OS, compiler.

Nov 04 '21 17:11 nmellado

The OS is Ubuntu18.04 ,and the compiler is gcc 7.5.0.
sorry, I don't known what configuration will effect this algorithm. I originally thought that as long as there is a set of suitable parameters, I can register.

Nov 05 '21 07:11 edwinlu-hdu

No problem, I'll try to have a look. Just to be sure, did you use the point cloud available at http://geometry.cs.ucl.ac.uk/projects/2014/super4PCS/ ?

Nov 08 '21 10:11 nmellado