fix(harness): switch VO backend to ORBVisualOdometry

SequentialVisualOdometry uses MockInferenceEngine (random keypoints) in
dev/CI, so RANSAC on random point pairs finds ≈0 geometric inliers and
vo_success is always False. ORBVisualOdometry uses real OpenCV ORB
features and achieves 99/100 tracking on EuRoC MH_01.

ESKF still never initialises (no start_gps call in harness) — that is
the next layer to address.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

src/gps_denied/testing/harness.py

@@ -29,7 +29,7 @@ from gps_denied.core.metric import MetricRefinement
 from gps_denied.core.models import ModelManager
 from gps_denied.core.processor import FlightProcessor
 from gps_denied.core.recovery import FailureRecoveryCoordinator
-from gps_denied.core.vo import SequentialVisualOdometry
+from gps_denied.core.vo import ORBVisualOdometry
 from gps_denied.schemas.graph import FactorGraphConfig
 from gps_denied.testing.datasets.base import (
     DatasetAdapter,
@@ -175,7 +175,7 @@ class E2EHarness:
         streamer.push_event = AsyncMock()
         proc = FlightProcessor(repo, streamer)
         mm = ModelManager()
-        vo = SequentialVisualOdometry(mm)
+        vo = ORBVisualOdometry()
         gpr = GlobalPlaceRecognition(mm)
         gpr.load_index(self._flight_id, "dummy")
         metric = MetricRefinement(mm)