Gazebo naive object recognition based on laser sensor [closed]
Hello everyone, I'm trying to implement in gazebo a faked camera sensor plugin for object recognition purposes. My first idea is to build a sensor which cast a number of rays. From each ray it takes the id of the nearest object it has intersected. Finally, if the number of times that an object is intersected is higher than a threshold, the objects is meant to be identified.
I plan to modify the ROS laser plugin and the gazebo Rayplugin which the former inherits from. I know that the Rayplugin uses a MultiRayShape structure, which can provides information about the detected fiducial value detected by each ray. I think that values are related with what I need, but I can't find information about what they really are.
In case that that fiducial values are not the id of the nearest object intersected, I've seen that the RayShape class has a method GetIntersection which seems to return that information (though I'm not sure).
As the Rayplugin use the MultiRayShape as base, it would be harder to modify in case of needing to change to the RayShape, so it'd be grand if I could use the fiducial information for my purposes. Could someone clarify me what are that fiducial values?. Any other idea about how to undertake the design of the sensor would be welcome! thanks in advance.
