Stream tracks

StreamTracks emits the set of currently tracked objects as it updates. By default you receive the filtered track set (the configured operational filter); set unfiltered to receive the raw fused set.

Before you begin

You need an authenticated channel — see Authenticate.

stream_tracks.py

from greenroom.perception.v1 import perception_service_pb2, perception_service_pb2_grpc
from greenroom.perception.v1 import classification_pb2

client = perception_service_pb2_grpc.PerceptionServiceStub(channel)

request = perception_service_pb2.StreamTracksRequest(vessel_id="vessel_1")
for response in client.StreamTracks(request):
    for obj in response.tracks.objects:
        cls = obj.classification[0].classification if obj.classification else None
        name = classification_pb2.ObjectClass.Name(cls) if cls is not None else "—"
        print(f"{obj.id} {name} range={obj.range:.0f}m bearing={obj.bearing:.0f}°")

Filtering on size and class

Classification and physical extent (obj.shape) let you filter for the objects you care about; the same criteria are expressed declaratively in ObjectsFilter, used by object alerts.

Optional fields may be absent

bearing and range are declared optional — check presence (obj.HasField("range")) before reading them, rather than trusting a 0.

What's next

• Configure object alerts — react to objects appearing or being lost.
• Streaming and reconnection — keep the stream alive.