train classifier database matching implemented in perception_pcl ? [closed]
Hallo everyone,
I am trying to do 3D object recognition with global recognition pipeline (segmentation, vfh descriptor, matching, alignment.. like proposed in this paper . I installed the perception_pcl stack in ROS and I started with segmentation of my scene point cloud (recorded by Xtion Pro Live) and then estimating VFH descriptor. But unfortunately I couldn't get how to train a classifier and create a database with different views of my CAD model (I converted it to PCD file). Isn't this implemented in the perception_pcl stack? Do I need another package? Which package should I use? I'm working on ROS Kinetic.
I would be so happy if anyone could help me
Best regards
gre721
This is really a PCL specific question. Could you please ask that over at the PCL forum? We try to keep ROS Answers on-topic as much as possible.
Thanks.