Blue streak in camera image: diagnose reflections and flare

A localized blue streak can resemble a damaged sensor, a dirty optical window, or a software fault. When it disappears after a small camera movement, reflections and flare deserve the first tests—but viewpoint changes can also alter false color on repetitive patterns. For inspection-system builders, the useful question is not simply whether the mark moves. It is which controlled change makes it appear, disappear, or change shape. The tests below separate illumination, optics, acquisition, and processing before you replace hardware.

Start with the most informative explanation

Flare is unwanted light reaching the sensor through reflections or scattering in the optical path. A bright source can produce a colored patch or streak even when that source lies outside the recorded image. Protective windows can contribute additional reflecting surfaces, but an angle-sensitive artifact does not identify which surface caused it.

Use the following sequence as a practical test order rather than a universal ranking of failure rates.

Fine repetitive detail deserves particular attention. Sampling and color reconstruction can create colors absent from the object. The guide to false colors and moiré in camera images explains that separate mechanism.

Establish a repeatable baseline

Do not begin by cleaning, changing software, adjusting exposure, and moving the camera together. Those changes may remove the symptom without identifying its cause.

1. Save an original frame at the camera's native output dimensions. Record the pixel format and whether the image has already undergone color reconstruction.
2. Record exposure, gain, white balance, aperture where adjustable, and any enabled correction functions.
3. Disable automatic adjustments where the camera supports manual control. Keep illumination unchanged while establishing the baseline.
4. Mark the camera position and target position. Perform adjustments with the inspected process safely stopped under the site's normal procedures.
5. Capture a short sequence, such as 10 frames, without moving anything. Check whether the streak is stable or intermittent.
6. Save the exact configuration needed to reproduce the symptom, including the processing sequence.

A display screenshot is useful for communication, but it is not a substitute for the original image file. Screenshots can include scaling, overlays, color management, or compression that the acquisition buffer does not contain.

If automatic controls cannot be disabled, record their reported values for every comparison. Changing iris, exposure, or gain automatically can otherwise conceal the effect of a lighting adjustment.

Test angle and lighting independently

First reproduce the angle sensitivity. Capture the original view, rotate slightly, and return to the marked starting position. A trial change of about 3 degrees is reasonable if the fixture permits it; the exact angle is not a diagnostic threshold.

If the streak disappears and returns reliably, you have established a geometric dependency. You have not yet proved flare: both reflections and sampling artifacts can depend on viewing angle.

Next leave the camera fixed and change only the suspected light path:

• Turn off one controllable light source at a time.
• Block a suspected off-axis source with a matte opaque card, positioned safely outside the camera's view.
• Remove nearby reflective objects individually.
• Repeat the capture after restoring each source or object.

A streak that disappears when one bright source is blocked, then returns when it is restored, strongly supports a stray-light explanation. Verify that the card did not simply darken the entire target and change automatic exposure.

Then replace the striped or textured subject with a plain matte target at approximately the same position. Keep exposure and illumination fixed. If the colored feature requires fine repetitive detail, investigate sampling and color reconstruction before replacing a window.

Do not deliberately aim the camera at the sun, a laser, or another hazardous high-intensity source to provoke flare. Use the existing scene and controlled, ordinary illumination.

Track the artifact in native image coordinates

Record where the streak lies relative to both the image boundary and recognizable scene features. These references answer different questions.

A sensor-related defect may stay at the same native pixel coordinates while the scene moves. A reflection can also occupy a repeatable image region when camera and lighting geometry remain fixed, so coordinate persistence alone is insufficient.

Consider an illustrative image measuring 2400 × 1600 pixels, with a colored region measuring 120 × 80 pixels:

This small affected fraction explains why average brightness can remain nearly unchanged while an inspection region becomes unusable. Measure local channel values and local contrast rather than relying only on a whole-image histogram.

Repeat the coordinate check with these conditions:

• The original target and original illumination.
• A plain matte target under the same illumination.
• The lens fully covered, using otherwise unchanged acquisition settings.

Disappearance with the lens covered supports a light-dependent mechanism, but does not identify the responsible optical surface. Persistence in covered captures points toward acquisition, processing, or sensor behavior requiring further investigation.

For defects that remain fixed across unrelated scenes, use the persistent image-defect isolation procedure. Avoid digital rotation, cropping, or resizing during coordinate comparisons unless you explicitly account for the transformation.

Separate acquisition from downstream processing

Find the earliest image representation in which the streak exists. Compare the saved camera output, the converted color image, the corrected image, and the final displayed image.

Do not call a file raw merely because it was saved before display. Camera output may already include color reconstruction, correction, or other processing. Record what each stage actually contains.

If supported by the acquisition path, save the original sensor mosaic separately from the reconstructed color image. A mosaic requires correct interpretation; displaying it directly as an ordinary color image is not a valid comparison.

Use the same captured frame as input to each processing test. Otherwise, changes in illumination or automatic settings can masquerade as software differences.

Useful checks include:

• Confirm the documented pixel format, channel order, row stride, and buffer dimensions.
• Disable optional resizing, sharpening, overlays, and geometric correction individually.
• Compare saved files in a second viewer capable of interpreting their format correctly.
• Inspect each color channel near the streak for clipping or unusual boundaries.
• Check whether the artifact repeats at image edges or processing-block boundaries.

For an 8-bit unsigned channel, 255 is the maximum code value. A patch reaching that value has lost information in that representation; it does not establish whether clipping occurred in the sensor, camera processing, or later conversion.

A global red/blue interpretation error usually affects the whole scene rather than one isolated patch. If colors are wrong throughout the image, follow the RGB channel-order checks instead of treating every blue region as flare.

Inspect optical surfaces without creating another fault

Inspect accessible front surfaces only after the controlled captures. Look for fingerprints, residue, scratches, or an added protective element that could scatter or reflect light.

Follow the equipment's documented cleaning procedure and use only approved materials. Do not remove a bonded window, open the sensor assembly, or apply an unapproved solvent to test a theory.

A visible scratch is not proof that it causes the colored streak. Its effect depends on location, illumination, aperture, and the rest of the optical path. Compare images before and after an approved cleaning action while restoring the same camera and lighting geometry.

The controlled tests for lens scratches help distinguish a visible surface mark from an image-quality problem that actually affects inspection.

If the evidence implicates an inaccessible internal surface, send the repeatable capture sequence to the equipment's authorized support channel. Include original files and settings rather than only a cropped screenshot.

Choose a corrective action before choosing replacement parts

For confirmed stray light, first change the offending geometry: reposition the source, reduce a nearby reflection, or use an appropriate lens shade. Recheck the inspection target's brightness and contrast after each change.

For pattern-dependent false color, evaluate optical sampling and the processing path. Deliberate defocus may suppress colored patterns, but it can also erase the defect you need to detect. Judge any adjustment against the smallest required feature, not just the appearance of a smooth image.

If the investigation leads to a new optical layout, compare machine vision lenses through the lens finder. Match field coverage, working distance, image circle, mount, and resolution requirements before selecting a part. The sensor-format explanation helps avoid treating nominal format labels as physical sensor dimensions.

The LS-FFL-0679 lens is a C-mount option with a 35 mm focal length. It is a candidate for an independently designed optical setup, not evidence that replacing a lens will cure the streak.

If a camera replacement is justified, compare area scan cameras using the camera finder. Verify the needed acquisition features and optical compatibility rather than selecting solely by resolution.

The LS-CMR-0053 camera is a C-mount, GigE model with a rolling shutter. Its inclusion here does not establish compatibility with an existing integrated imaging assembly.

Shutter behavior matters when the target moves or illumination changes during readout. Changing shutter type, however, is not a general remedy for optical flare.

Verify the fix across the operating envelope

A clean image at one camera angle is not enough if normal part placement or lighting variation can bring the streak back. Repeat the original failure condition after applying the corrective action, then test representative operating extremes.

Use this closing checklist:

• Original condition: Reproduce the original scene, settings, and illumination to confirm the symptom is removed.
• Viewpoint variation: Check the allowed camera and part-position variation, not just the preferred alignment.
• Lighting variation: Test expected source states and nearby reflections without changing several variables together.
• Image stages: Confirm that both saved acquisition data and final inspection images are acceptable.
• Inspection performance: Verify that the smallest required feature remains detectable after optical or processing adjustments.
• Configuration record: Save the final settings, fixture positions, and representative passing and failing images.

If only a narrow viewpoint avoids the problem, decide whether that alignment can be maintained before accepting it as a production fix.

For a revised optical arrangement, start with the lens configuration step. If the cause remains uncertain, use technical support with the original frames, settings, and a short record of which controlled changes affected the streak.