Definition
Road Triage does not rely on a single image to assess pavement distress. The Perception Engine observes every distress across multiple frames as the survey vehicle passes — the same crack may appear in 5 to 15 consecutive images from varying distances and angles. This multi-observation approach builds a measurement profile for each distress instance, cross-validates measurements between frames, and identifies the optimal observation distance for the most reliable results. Distress dimensions (length, width, area) are derived from pixel-precise AI segmentation masks with AI-powered camera calibration. Crack widths are classified into Low, Medium, and High severity per ASTM D6433 and AASHTO R 55. Every measurement carries an internal confidence score — uncertain classifications are flagged for quality review rather than guessed.
In context
Because the same crack appears in several frames as the vehicle passes, multi-frame measurement can cross-check its size.