Validate synchronized cameras under burst traffic

What the test resolves

Simultaneous exposure and simultaneous delivery are different requirements. Validate exposure alignment while independently testing whether burst traffic overwhelms the shared link.

This is a vendor-neutral commissioning method. Check the exact camera, lens, sensor or controller manual before changing settings. The numerical worksheet is hypothetical, not a Deyi Vision measurement or a product specification.

Operating prerequisites

Use a common hardware trigger or documented scheduled acquisition. If PTP is used, confirm support and clock stability over time; a locked clock alone does not prove aligned exposures.

Define whether the inspection needs aligned exposure, aligned timestamps or only images associated with the same part. These requirements are different. Establish support for the proposed synchronization method in every camera and network component. A PTP lock indicator alone does not verify exposure alignment. For this test, keep each camera's acquisition sequence identifiable and preserve a path for identifying a group with a missing or repeated member.

Bench and records

Prepare a shared flashing target or documented timing outputs, per-camera IDs/timestamps and network counters. Record topology, packet pacing, frame transmission delays and application queue depth.

Prepare a group record containing part or cycle ID, camera ID, frame ID, exposure settings, trigger method, timestamp clock definition and complete-image status. Document shared links and receiver queues. Provide a common observable event, such as a properly controlled optical target, or documented timing outputs measured on a common instrument. Record how the observable event relates to exposure; software arrival times are not substitutes for that relationship.

Run the test

1. Verify all camera clocks or common input signals before collecting images.
2. Acquire 1000 synchronized groups and measure exposure skew with the optical target or outputs.
3. Run every camera at intended payload and burst rate; log incomplete frames and queue growth.
4. If exposure must remain simultaneous, stagger transmission only with supported controls; repeat after restart and a clock-master change.

Verify each camera independently before starting grouped acquisition. Then acquire the planned groups and check that every group contains exactly the intended camera members. Measure exposure alignment using the common event, while logging transfer completion and queue growth separately. Introduce only supported transmission staggering if bursts overload the receiver; retain the exposure test while changing transport timing so improved reception cannot hide a new synchronization error.

Hypothetical worksheet

Hypothetically four cameras send 2 MB each at 10 fps: average payload is 80 MB/s, but each group produces an 8 MB burst. At an assumed usable 100 MB/s, draining one burst takes at least 80 ms before overhead.

aggregate_payload = cameras * bytes_per_frame * fps
4 * 2 MB * 10 = 80 MB/s
burst_drain_time >= 8 MB / 100 MB/s = 80 ms
clock_offset != exposure_skew != delivery_skew

The average payload calculation answers long-term loading, whereas the burst drain calculation describes a transient shared-link demand. The assumed usable bandwidth must be measured or justified for the actual path. If another burst begins before the previous one drains, queues can grow despite an acceptable average. Include protocol overhead and receiver behavior when interpreting the hypothetical lower bound; the calculated drain time is not a camera timeout setting.

Test sequence

Interpret the evidence

Use the diagram to maintain separate evidence for exposure and delivery. Synchronized cameras may expose together but finish transmitting at different times. Conversely, images arriving together may have been exposed at different times. Plot the observed alignment and transfer completion by group identifier, keeping groups with incomplete members visible. A stable mean is insufficient when an occasional alignment excursion can invalidate a measurement of the moving part.

Troubleshooting

• Good average bandwidth with losses at each trigger suggests burst contention.
• Similar receive timestamps do not prove exposure synchronization; differing clocks can also make timestamp comparisons invalid.

If only grouped acquisition fails, compare receiver statistics with the independent-camera baseline and inspect shared bottlenecks. If timestamps drift, record synchronization state and offsets over the run rather than checking a single startup value. If a camera appears twice in one group, review cycle association before changing network settings. After any reconnect, verify group identifiers and clock state again; a clean link does not prove a clean resumption of synchronized acquisition.

Acceptance and handoff

Specify independent limits for exposure skew, frame completeness and delivery deadline. Require stable clocks and bounded queues during the agreed worst burst pattern; record reconvergence behavior after clock changes.

Archive the raw files, settings, sample identifiers and decision limits together. Discuss the documented requirements with Deyi Vision using the related product and preparation guides below.

Release only when fresh groups satisfy both the agreed exposure-alignment requirement and the complete-delivery requirement under the intended host workload. Include the most demanding image format, trigger pattern and shared traffic within the declared envelope. Verify that a missing camera or incomplete group produces an explicit invalid inspection result. Archive the group ledger, synchronization observations and final transport settings so later bandwidth or camera changes trigger a focused revalidation.