Simple object recognition code

Compilation:

• mkdir build
• cd build
• cmake ..
• make

Usage example:

object_recognition --step 1 --dataFile data.txt --scene cloud.pcd

--step: specify how many model views to skip at each iteration (e.g. if it is 2 it takes one model each 2 and so on)

--dir: specify the directory where to find the data.txt file of a specific model

--scene: specify the .pcd point cloud where the model has to be identified

Output:

The program will print out the following informations:

• Current model view information (prefix filename and camera transform)
• A point represents the fact that an alignment was found but it was not a best one
• An x represents the fact that an alignment was found and it was better than the previous best one
• Final resulting transform (camera transform + alignment transform)