Camera Commissioning and Acceptance Testing

No single federal mandate governs camera acceptance testing on buses; the compliance bar here is 49 CFR 396 maintenance record requirements and district policy, which means the acceptance process itself is entirely the fleet's own quality control mechanism. Installation confirms the hardware is physically mounted and wired. Acceptance testing confirms it actually does its job — the field of view is correct, the timestamp is accurate, footage is legible at the distances that matter, and data reaches wherever it's supposed to reach. Treating these as one combined step, rather than two distinct ones with separate sign-offs, is how a mounting error or a wiring fault goes unnoticed until an incident review months later.

A camera can record clear, well-composed footage while still failing acceptance on either of these two points. Timestamp accuracy needs to be checked against a known reference — a phone clock, a synced dispatch system — rather than assumed correct because the camera displays a plausible-looking date and time on screen; a drifted clock produces footage that looks fine but can't be reliably placed on the correct day or hour later. Plate legibility needs an actual field test with a real vehicle at a realistic stop-arm distance, not a bench test in a shop bay, since resolution, frame rate, and lens quality all interact differently at that specific range than they do up close.

Acceptance testing needs to confirm that footage or event data actually reaches its intended destination — whether that's a cloud upload triggered by an event, a bulk sync completed overnight on depot Wi-Fi, or a manual download process for a simpler setup. Confirming this at commissioning, rather than assuming it based on the vendor's product description, catches configuration errors before the camera has been silently failing to upload for weeks. Health reporting should be confirmed the same way: the device's status needs to be visible in whatever monitoring system the fleet uses, verified at commissioning rather than trusted to "just work" from the point of install onward.

Here's the objection worth naming directly: install positions chosen for convenience rather than field of view, leaving the loading door uncovered, is exactly the kind of mistake that acceptance testing is designed to catch before a bus ever runs a route. A camera mounted at an angle optimized for general side visibility can pass a casual glance — "yes, there's a camera there, it's recording" — while still missing the specific zone that mattered most. The acceptance checklist's field-of-view step should require someone to actually pull up the recorded frame and compare it against the named coverage zone, not just confirm the unit is powered and blinking.

A district running 45 buses had previously treated a technician's verbal "yeah, it's installed" as sufficient sign-off for camera work. After adopting a formal per-bus acceptance checklist with individual sign-off, a routine commissioning test on one bus revealed the forward camera's timestamp had drifted nearly a full day off from the actual date — a factory default clock setting that had never been corrected during install. Because the checklist specifically required a timestamp accuracy check against a known reference, the issue was caught and fixed before that bus ran a single route, rather than being discovered later during an incident review where the wrong date on a clip could have created real confusion.