Stereo camera trigger synchronization for paired image capture

Two cameras receiving capture commands together do not necessarily expose together. For stereo imaging of moving objects, that difference can shift features between the views and distort the reconstructed geometry. You need a timing requirement, a documented trigger path, and evidence that each image pair meets the requirement.

This procedure focuses on exposure timing rather than software arrival order. It explains when to use a shared trigger, why an exposure-output connection introduces additional delay, and how to verify image pairing. For the basic connection sequence, start with triggering two USB cameras together.

1. Choose the timing architecture before buying cables

Prefer a common hardware pulse delivered to both cameras' documented trigger inputs. This removes the intentional sequence in which one camera must begin exposing before it signals the other. It does not eliminate differences in input circuitry, trigger latency, or sensor behavior.

An alternative is to configure one camera's documented exposure-status output to trigger the second camera. That arrangement can work when the permitted timing error is larger than the measured delay. However, the second camera responds after the first camera generates an output and the receiving camera recognizes it. An exposure-status output is not automatically a zero-delay synchronization signal.

For moving stereo, use the shared-input arrangement unless testing demonstrates that another architecture meets your timing budget. The distinction between requests and physical exposure is explained in trigger calls versus exposure timing.

Before proceeding, obtain the manufacturer's documentation for trigger inputs, output functions, electrical limits, and acquisition modes. Catalog interface and shutter information alone cannot establish these capabilities.

2. Turn scene motion into a timing requirement

Specify how much motion-related mismatch you can tolerate between views. Use projected image velocity at the feature being measured, not simply vehicle speed or conveyor speed.

For approximately constant image velocity during the timing offset:

Suppose a feature moves at 10,000 pixels/s and the allowed temporal displacement is 0.25 pixel:

This is an application estimate, not a camera specification. Image velocity changes with scene depth, viewing direction, and rotation. Use the fastest relevant motion within the measurement region, with margin for uncertainty.

Exposure duration also matters. Suppose the right exposure starts 10 microseconds later, while the left and right durations are 100 and 140 microseconds:

That exceeds the example limit even though the exposure starts differ by only 10 microseconds. Matching exposure durations removes this particular midpoint contribution, but does not correct the start-time offset.

These exposures also produce approximately 1 and 1.4 pixels of motion blur, respectively. Synchronization cannot recover detail already smeared during integration. For the wider relationship between timing and depth quality, use stereo camera synchronization for reliable depth maps.

3. Check shutter behavior and connection compatibility

Global shutter generally simplifies moving-scene stereo because the image does not have a row-dependent exposure start. Rolling shutter requires additional analysis: matching frame triggers does not by itself prove that corresponding image rows observed the scene together.

Within area scan cameras, the MV-CH120-11UC is a USB3 camera with a global shutter. Treat it as a candidate for further evaluation, not proof of a supported synchronization arrangement.

Use the camera finder to narrow interface and shutter choices. Then obtain the timing and I/O documentation for the exact candidate before committing to the architecture.

Check the following separately:

• Trigger input: supported electrical interface, polarity, threshold, permitted pulse width, and input filtering.
• Timing behavior: trigger latency, jitter, trigger acceptance conditions, and available exposure-status signals.
• Output circuit: whether it sources, sinks, or switches an isolated circuit, plus its load limits.
• Connector: exact pin assignment, signal reference, cable orientation, and compatible mating connector.
• Acquisition behavior: whether external triggering works in the required continuous acquisition mode.

A matching connector shape does not establish electrical compatibility. For the wiring review, follow camera GPIO wiring checks.

4. Build and verify the trigger distribution

1. Disconnect power before changing I/O wiring. Identify every connection from the documented pinout rather than wire color or connector appearance.
2. Check whether the trigger source can drive the combined input load. Use a suitable distribution interface when direct fan-out does not meet the documented electrical requirements.
3. Connect the required signal references exactly as documented. Do not bridge isolated circuits or grounded equipment without checking the permitted arrangement.
4. Route and secure the interconnects, providing strain relief and clearance from moving parts.
5. Measure the pulse at each receiving input using suitable instrumentation. Check amplitude, polarity, pulse width, ringing, and relative edge timing.

Equal cable lengths can reduce one source of asymmetry, but they cannot cancel different camera input delays. The measurement point should be the receiving connection, not only the pulse generator output.

Choose trigger interconnects and data cables as separate parts of the design. The LS-CBL-0219 is a USB3.0 camera data cable; it is not the electrical trigger connection described here.

The accessory finder helps organize cable selection. Confirm connector fit and interface suitability independently for every connection.

5. Configure both cameras before releasing pulses

Configuration names vary between devices. In GenICam-based tools, common controls include TriggerSelector, TriggerMode, TriggerSource, and TriggerActivation. Use only the settings and dependencies documented for your camera.

1. Stop acquisition before changing trigger configuration, unless the documentation explicitly permits the change while streaming.
2. Select the appropriate frame-start trigger function. Confirm that the selected event starts the intended image rather than an acquisition sequence or burst.
3. Select the documented physical input and matching edge polarity on both cameras. Follow the trigger source and line configuration checks if the expected input is unavailable.
4. Set trigger delays deliberately. Begin without intentional delay where supported. Record any unavoidable or configured differences.
5. Set fixed exposure durations and gain. Disable automatic changes during initial timing validation so the exposure relationship remains stable.
6. Select the documented ongoing acquisition mode. A single-frame mode may accept one event and then stop; see hardware triggering that stops after one frame.
7. Start and arm both acquisition paths before enabling the pulse source. Confirm readiness using the documented mechanism where available.

If one camera drives the other, additionally select the documented output function and confirm its active polarity. Do not substitute a general acquisition-active signal for exposure-active without understanding their different meanings.

Read back the effective settings after acquisition starts. Automatic controls or mode dependencies may change requested values. Where supported, inspect per-frame exposure and gain metadata rather than assuming every image used the initial request.

6. Prove exposure alignment, not just pulse alignment

A shared edge proves a shared input event. It does not prove that the sensors began integrating together.

Where documented exposure-status outputs are available, observe them alongside the input pulse. Measure both the difference between exposure starts and the exposure durations. Confirm what each status signal actually represents; an output transition can itself have delay relative to sensor integration.

Add an optical test that both cameras observe. For example, use a controlled illuminated target with a time-varying pattern and independently characterized timing. A single long flash that appears in both images proves only that the exposure windows overlapped the flash. It does not establish a small start-time difference.

Use a moving target as an application-level check. Compare corresponding feature positions while varying speed and trigger phase. Do not interpret all residual stereo mismatch as timing error: calibration, lens distortion, blur, and correspondence failures can produce similar symptoms. Establish geometric performance with paired stereo calibration images before assigning the remaining error to synchronization.

For an initial validation plan, record 1,000 trigger events at a low rate, then repeat at the intended operating rate. Report the largest observed timing difference, not only the average. This sample does not guarantee future timing, but it can expose rate-dependent delays and intermittent missed triggers.

7. Preserve image pairs through transport and processing

USB arrival order is not exposure order. Buffering, transfer scheduling, and processing can make correctly synchronized images arrive at different times.

Maintain a pairing record that associates the images with their physical trigger event. Use documented event identifiers or counters where available. Independently increasing frame counters are not necessarily a shared event identity: after a missed trigger, pairing the next received image from each camera can silently combine different events.

For each accepted pair, retain:

• Camera identity and stream identity.
• Available frame counters and trigger-event information.
• Exposure duration and relevant acquisition status.
• Timestamp meaning, units, clock domain, and any established clock relationship.
• Whether the pair passed timing and completeness checks.

Do not subtract timestamps from unrelated clocks as though they share an origin. Also distinguish exposure timestamps from host delivery times, as described in interpreting camera buffer timestamps.

Deliberately test a missing-image case. Confirm that the application rejects an incomplete pair instead of shifting all subsequent pairings. For asynchronous processing, use frame matching with timestamps to keep delayed results attached to their original acquisition records.

8. Check rate limits and document acceptance

A correct low-rate setup can fail when the trigger interval becomes shorter than the camera's accepted cycle time. Exposure, readout, acquisition mode, and transfer constraints all matter; do not assume they simply add together because some devices overlap operations.

At 50 trigger events/s, the interval is:

Whether that interval is acceptable depends on the documented mode and verified behavior. Watch for rejected triggers, incomplete frames, and steadily growing queues. Use external-trigger frame-rate checks when event counts and image counts disagree.

Before ordering or accepting the station, complete this checklist:

• Timing budget written: permitted displacement is converted into an exposure-alignment requirement.
• Trigger architecture documented: common-input or camera-output topology is explicit.
• Electrical compatibility confirmed: pinouts, references, loading, and pulse requirements match.
• Exposure behavior measured: shutter type, durations, latency difference, and observed jitter are recorded.
• Pairing failures tested: missing and delayed images cannot silently create false stereo pairs.
• Operating rate validated: measurements cover the intended pulse rate and processing load.
• Restart behavior checked: both cameras restore the intended settings and arm before pulses resume.

The usual shared-trigger answer is incomplete when rolling-shutter timing is undocumented, independent exposure control changes integration windows, or the application cannot identify missing events. Resolve those limitations before relying on the depth output. This procedure validates imaging behavior, not a machine safety function.

Next, use the camera configurator to organize candidate parts. For a timing and wiring review, send the proposed topology, electrical documentation, motion estimate, and required timing tolerance through application support.