Pierre R. Mai

Even in the case of data that a sensor might detect, this would have potentially different semantics and should either be added as new fields to VehicleClassification or to DetectedMovingObject...

Note that a wand is something wizards wield, not something commonly found on roads, so I'm rather hoping that ADI does not use this term... In any case this is...

Note that nearly all of the proposed additions are semantic road signs in OSI, hence the list of wanted enums would have to be aligned with the code / sub-code...

I think it is somehow a bit more complex, because some Ids should be unique across classes (e.g. for detected stuff this really makes sense, the sensor id stuff also),...

I think the sentence stating that the parent frame is the sensor frame was added erroneously during a clean up (2740f2d), which assumed that since it is in SensorData it...

This seems to be a hold-over from the initial OSI commit, which already did not contain a color with index 4. We should likely fix this at some time; for...

BTW it would be nice to consult with ISO 23150 and AUTOSAR groups whether tricycle should include e.g. motorcycle with sidecar. In general the definitions need to be much more...

I'm a bit confused why TYPE_VERTICAL_STRUCTURE and TYPE_RECTANGULAR_STRUCTURE occur in this list, since they are verbatim already part of OSI's definition. Similarly TYPE_LAMP_POST, TYPE_OVERHEAD_OBJECT, TYPE_REFLECTOR, and TYPE_BEACON are already present...

Looking like a very good start to me, will form the basis of our further performance work going forward.

CCB 2022-05-02: The use case for this in regards to OSI SensorView (rather than native camera output formats, which are not OSI-based) is not very clear yet, and should be...