search

rock-core / base-types

C/C++ and Ruby definition of the base types in Rock
6 stars 40 forks source link

add SonarVisualization for single beam sonar #126

Closed alcantara09 closed 5 years ago

alcantara09 commented 5 years ago

Add visualization for Scanning Sonar visualization using the new data Type. It has the same functionality of SonarBeamVisualization plus the option to visualize only the current beam or the full scan

doudou commented 5 years ago

Is that expected ?

image

alcantara09 commented 5 years ago

yes

doudou commented 5 years ago

.... could you change the message to describe what should be expected ? I was put off by the fact that the beam is not centered. I would do it myself, but I can't since I don't know whether the colors matter (if they do, describe them).

saarnold commented 5 years ago

Could the motor step angle not be computed by successive scans? Currently the default value is zero and in case of single scans needs to be set by the user.

doudou commented 5 years ago

Currently the default value is zero and in case of single scans needs to be set by the user.

This is because the "compute with successive scans" method is not robust at all if for whatever reason the widget is not getting updated fast enough. We've had a lot of trouble with this setting in all the sonar visualizations.

saarnold commented 5 years ago

This is because the "compute with successive scans" method is not robust at all if for whatever reason the widget is not getting updated fast enough.

Maybe we could keep it as an option, deactivated by default.

What about using the beam width as initial setting? That would probably lead to gabs or overlapping between beams but would at least correlate with the area the signal comes from.

doudou commented 5 years ago

What about using the beam width as initial setting? That would probably lead to gabs or overlapping between beams but would at least correlate with the area the signal comes from.

Well ... duh. It would not be only better, it is the right thing to do (and it's simpler). The width of the beams should definitely be ... the beam_width. Thanks !

alcantara09 commented 5 years ago

@saarnold, I stopped to think about the objective of the sonar visualization for a while. The full scan mode generates a really hard to interpret visualization, given that the robot's movement moves its position and the only reading that is aligned with the world is the current one. In the SonarWidget visualization, the scan is always in relation to the origin of the sensor, what does not hold in the 3d visualization.

Given that we have the SonarWidget, I suggest we keep only the single and multibeam visualization for the current beam and do not support full scan mode. What do you think about it?

saarnold commented 5 years ago

Given that we have the SonarWidget, I suggest we keep only the single and multibeam visualization for the current beam and do not support full scan mode. What do you think about it?

I would be OK with that. The keep old data flag is another way to visualize a batch of successive samples.