Closed scorpeeon closed 10 years ago
Does background updating, vehicle detection with block-based segmentation, and vehicle tracking with error compensation.
Uses background model. Vehicles are detected based on foreground segmentation and moving-edge extraction. Tracking verification is performed by the fuzzy hybrid information inference mechanism (FHIIM), where compensation is invoked when the comparison result is dissimilar.
Only focuses on lanes on highway roads. But discusses congested scenes, which might be useful. Has good test results.
ANewApproachToVideoBasedTrafficSurveillanceUsingFuzzyHybridInformationInferenceMechanism.pdf