Lens Replacement Qualification

This plan evaluates an industrial imaging lens against an EOL incumbent or qualifies a lens for a new machine-vision system. Execute it under the common controls in Machine-Vision Camera and Lens EOL Qualification. Test the lens first on a calibrated reference camera or optical bench, then on the proposed production camera.

Objectives and release criteria

The lens evaluation shall establish:

• mechanical fit, sensor coverage, mount/flange compatibility, retention, and production adjustability;

• correct focal length, FOV, magnification, working distance, entrance pupil, and image/object-space geometry;

• focus position, field curvature, depth of field (DOF), and stability;

• spatial-frequency response across field, wavelength, aperture, object distance, and manufacturing samples;

• lateral and longitudinal chromatic aberration;

• distortion, telecentricity or perspective, principal point, and calibration residual;

• relative illumination, vignetting, spectral transmission, colour shading, flare, ghosting, and stray-light behaviour;

• environmental, vibration, focus/iris retention, unit-to-unit, and lot consistency;

• acceptable application decisions and measurements with the production camera, illumination, mechanics, and algorithm.

Unless a test is explicitly marked not applicable in the approved protocol, lens qualification requires PASS for L01–L14. L01–L07, L10, L14, and every false-accept or measurement-accuracy metric are critical. The remaining items are release requirements unless the system owner and quality owner approve a documented deviation.

Required equipment and fixtures

In addition to the common bench in Machine-Vision Camera and Lens EOL Qualification, prepare:

• an optical MTF bench or calibrated reference camera whose sampling, cover-glass stack, active area, and noise do not mask the lens behaviour;

• the baseline and proposed production cameras;

• calibrated slanted-edge, Siemens-star or other MTF targets covering the full object field, plus a traceable dot/grid target;

• a collimator and reticle for infinity-focus lenses, or a rigid finite- conjugate target stage for machine-vision lenses;

• narrow-band sources or filters at every critical production wavelength;

• a characterized uniform source and a spectroradiometer or calibrated detector for relative illumination/transmission;

• axial translation with adequate resolution for through-focus MTF, flange distance, longitudinal colour, and DOF measurements;

• object-height/depth fixtures and a precision rotary/translation stage for telecentricity, chief-ray, and distortion tests;

• an external bright source, black target, masks, and angular stage for flare and ghost testing;

• production mounts, adapters, filters, retaining rings, locking screws, torque tools, cable routing, and heat sources representative of the machine.

If a camera is used as the MTF detector, keep it and its processing identical for baseline/candidate lens comparisons. Its Nyquist frequency and SNR must exceed the highest required lens frequency with margin. Report the measured result as camera-lens system MTF unless a validated detector MTF correction has been applied; never label an undersampled system result as lens-only MTF.

Lens test summary

ID	Evaluation	Primary data	Required result
L01	Identity, documentation, compliance, lifecycle	Part/revision/lot matrix, supplier and coating evidence	Exact identity and all mandatory evidence accepted
L02	Mechanical fit, mount, coverage, retention	Inspection, tolerance stack, image-circle/CRA checks	All fit, coverage, safety, and retention limits pass
L03	Focal length, FOV, magnification, working distance	Grid fit, scale/FOV versus distance, pupil data	All first-order geometry CTQs pass
L04	Focus, flange/back focus, field curvature, DOF	Through-focus MTF surfaces and best-focus maps	Fixed-plane focus and DOF limits pass
L05	MTF/resolution across field	Tangential/sagittal MTF versus field/frequency	Every field/band/direction/application-frequency gate passes
L06	Longitudinal/lateral chromatic aberration	Best-focus shift and colour-displacement vector fields	Chromatic focus, MTF, and registration limits pass
L07	Distortion, telecentricity, perspective, calibration	Mapping residuals, magnification versus field/depth	Raw and calibrated geometry budgets pass
L08	Relative illumination, vignetting, chief-ray compatibility	Flat-field maps versus aperture/band/focus	Coverage and uniformity limits pass
L09	Spectral transmission and colour balance	Transmission curves and exposure/channel ratios	In-band signal and out-of-band rejection pass
L10	Flare, ghosting, scatter, and high dynamic range	Source-angle maps, veiling glare, ghost catalogue	No false feature; contrast/flare limits pass
L11	Aperture/focus adjustment and lock repeatability	Setting cycles, torque, scale, focus, MTF distributions	Every cycle returns within setting and imaging limits
L12	Thermal, vibration, shock, and contamination stability	Before/during/after optical and physical results	CTQs pass throughout environment with no damage
L13	Unit/lot consistency and incoming control	Per-unit distributions and incoming test specification	Variation fits budget; incoming controls accepted
L14	Integrated application validation and pilot	Blind decisions, measurement-system analysis, pilot log	All production decision and measurement gates pass

Detailed experiments and pass/fail rules

L01 – identity, documentation, compliance, and lifecycle

Objective. Confirm that the exact lens configuration is traceable, controlled, and supportable for the required machine lifetime.

Procedure and conditions. Record manufacturer, order code, serial, manufacturing lot, optical revision, mount, focal-length designation, aperture type/range, focus range, coating band, filter or protective window, glass/material declarations, and country of origin. Review the full specification, drawing, environmental limits, cleaning restrictions, regulatory/material declarations, warranty, PCN policy, lifecycle commitment, repair/calibration support, and whether coating, glass, cement, mechanics, or supplier may change under the same order code.

Collect and output. Produce a signed identity/compliance matrix, supplier evidence pack, lifecycle risk assessment, and controlled difference list versus the baseline lens.

PASS. Delivered identity and coating match the order code; every mandatory compliance and lifecycle requirement has acceptable evidence; serial/lot and revision are traceable; and optically consequential changes require adequate notification and requalification.

FAIL. Any mandatory evidence is absent, coating/glass/revision is ambiguous, a sample differs from its declaration, or uncontrolled material/design changes can enter production.

L02 – mechanical fit, mount, sensor coverage, and retention

Objective. Demonstrate that the lens installs safely, covers the sensor, and maintains alignment without interference or mount-induced variation.

Procedure and conditions. Inspect outside envelope, mass, centre of gravity, mount thread/flange, flange focal distance, shoulder squareness, thread engagement, locating datum, rear-element protrusion, sensor-window clearance, filter thread, iris/focus ring access, lock locations, cable/machine clearance, and required support. Calculate the tolerance stack over focus travel and production adjustment.

Measure usable image circle and relative illumination on the complete active sensor, including worst manufacturing decentre and mount tolerance. Review exit-pupil/chief-ray angle (CRA) against the sensor microlens and cover-glass requirements. Fit the lens in a representative machine using the released adapter, torque, locking method, and strain relief. Apply the required orientation and gravity directions.

Collect and output. Produce dimensional inspection, CAD/envelope overlay, tolerance stack, image-circle/CRA evidence, mount runout/tilt, fit photographs, torque record, and any support-bracket calculation.

PASS. All hard dimensions, clearances, thread engagement, loads, flange/datum limits, active-sensor coverage, CRA, retention, access, and safety margins pass for every unit and orientation. No unapproved adapter modification, rear-element collision, mount sag, vignetting, or adjustment obstruction exists.

FAIL. Any hard fit/coverage/retention limit fails, the camera mount carries an unapproved load, or an undocumented assembly selection is needed to obtain alignment or focus.

L03 – focal length, FOV, magnification, working distance, and pupils

Objective. Verify the first-order optical geometry across required conjugates.

Procedure and conditions. Image a traceable grid at the nominal and extreme object distances using the reference camera. Fit object-to-image mapping and measure effective focal length where the method supports it, transverse magnification, horizontal and vertical FOV, working distance from controlled datums, principal point, entrance/exit pupil location or working f-number where relevant, and focus travel/margin. For zoom, liquid, or motorized lenses, test every released setting and approach each setpoint from both directions.

Report results in physical sensor/object units. Do not infer focal length only from the engraving, and do not compensate a wrong FOV by image resizing.

Collect and output. Produce calibrated grid images, scale/FOV versus distance plots, fitted first-order parameters, focus/adjustment margin, and baseline/candidate differences.

PASS. Required FOV is covered with margin at every working distance; magnification, scale, aspect, focal-length-equivalent behaviour, pupil/working-f-number, and adjustment range meet their guard-banded absolute limits; and no metric regresses beyond its approved baseline margin.

FAIL. Any required field is cropped, scale or working distance exceeds its budget, focus cannot be reached with assembly tolerance, or electronic resampling is needed to conceal a geometry failure.

L04 – focus position, flange/back focus, field curvature, and DOF

Objective. Determine whether the complete field and object-height range remain sharp at one lockable production focus.

Procedure and conditions. At every required aperture, object distance, wavelength, and temperature, sweep the detector or target axially through focus using steps fine enough to resolve the permitted focus error. At centre, mid-field, corners, and application-critical points, calculate MTF at the application frequency in tangential and sagittal directions versus axial position.

Record best-focus position per field, through-focus width at the required MTF, field curvature, astigmatic focus separation, flange/back-focus margin, focus sensitivity, and object-side near/far DOF. Determine the single production sensor plane/focus setting by the pre-approved optimization rule, then report all field points at that common plane. Also repeat after remounting and locking.

Collect and output. Produce through-focus curves, field-versus-focus MTF surface, best-focus and astigmatism maps, selected common plane, usable DOF interval, remount repeatability, and mechanical focus margin.

PASS. At the locked common production plane, every required field, direction, wavelength, object height, and unit meets the guard-banded MTF limit; near/far DOF contains the entire object range with tolerance margin; flange/focus travel permits assembly; and remount/refocus variation fits its allocation.

FAIL. Only independently refocusing each field produces a pass, usable DOF misses a required object height, production focus lies at an adjustment stop, or field/thermal/remount shift violates the fixed-plane MTF limit.

L05 – MTF and resolution across field

Objective. Measure spatial contrast over the full field and required spectral/aperture range at the actual production focus.

Procedure and conditions. Using an optical MTF bench or adequately sampled reference camera, measure tangential and sagittal MTF at centre, multiple mid-field points, four corners, and application-critical locations. Test every production wavelength/band, aperture, conjugate, focus setting, and orientation. Sample frequency densely through and beyond the highest application frequency, within the detector’s valid range. Repeat after independent mounting and on all qualification units.

Report both:

• best-focus MTF, useful for diagnosing intrinsic lens performance; and

• common-production-plane MTF, which is the release metric.

When using captured targets, correct for documented target and detector limits or retain the result as end-to-end system SFR. Check phase/alias sensitivity and distinguish lens anisotropy from sensor-row/column response. The KrakenOS captured USAF workflow is documented in Captured USAF-1951 MTF.

Collect and output. Produce native images or bench data, tangential/sagittal MTF curves, MTF at application frequencies, MTF50/MTF10 where stable, field heatmaps, azimuth asymmetry/decentration indicators, repeatability, and baseline/candidate overlays.

PASS. At the locked production plane, every field, azimuth, direction, band, aperture, conjugate, and unit meets the absolute MTF requirement at each application frequency and the approved non-inferiority margin. Azimuth asymmetry and remount variation fit their budgets, and no pass depends on detector sharpening or undersampling.

FAIL. Any required point/direction/band misses its guard-banded limit, only best-focus rather than common-plane data pass, unit decentre/tilt exceeds its limit, or the detector/target cannot support the claimed measurement.

L06 – longitudinal and lateral chromatic aberration

Objective. Measure wavelength-dependent focus and magnification/position changes that can reduce colour contrast or separate channels.

Procedure and conditions. Use narrow-band sources near the short, centre, and long production wavelengths; add intermediate or out-of-band points when material spectra require them. For each wavelength and field point, run a focus sweep and record best-focus position and through-focus MTF. At a common production focus, measure MTF in every band.

Image a high-accuracy dot grid, edge, or pinhole field at the same mechanical geometry. Fit a common reference coordinate system, remove only approved global translation/rotation, and calculate field-dependent lateral displacement and magnification difference for every wavelength pair. Report the vectors in sensor micrometres, pixels of each proposed camera, object-space units, and normalized field height.

Run valid BB/CB/BC/CC combinations and the reference lens/camera isolation experiment in Machine-Vision Camera and Lens EOL Qualification. Measure before and after the production colour correction; do not let a high-order calibration obscure an unstable lens.

Collect and output. Produce longitudinal best-focus-versus-wavelength plots, per-band common-plane MTF, lateral-colour vector maps, radial/tangential components, object-space error, correction model/residual, and repeatability across remounts and units.

PASS. All per-band common-plane MTF limits pass; best-focus spread is no more than 25 % of usable application DOF by default; residual lateral displacement is no more than 25 % of the smallest positional tolerance and no more than 0.5 production-camera pixel where channel registration is assumed; and no chromatic metric exceeds its approved baseline regression. Application- specific limits replace these defaults when pre-approved.

FAIL. Any required wavelength fails MTF/focus/position limits, correction residual is unstable across unit/focus/temperature, or chromatic error is reported only in a unit that hides its impact (for example micrometres without pixels and object-space error).

L07 – distortion, telecentricity, perspective, and calibration residual

Objective. Qualify raw image geometry and the residual after a controlled production calibration.

Procedure and conditions. Image a traceable grid covering the FOV at nominal focus and object depths spanning the URS. Measure radial and tangential distortion, local magnification, aspect/orthogonality, principal point, field-dependent position error, and decentre. For object-space telecentric lenses, translate the target through depth and measure scale, centroid, and perspective change. For image-space telecentric requirements, assess exit pupil/CRA using an approved method.

First report uncorrected geometry. Then fit the exact calibration model and number of parameters allowed in production on a training capture. Validate residuals on independent positions, depths, orientations, remounts, temperatures, and lens units. Do not increase model order after viewing the validation result.

Collect and output. Produce distortion vector/percentage maps, local scale and telecentricity versus depth, calibration coefficients, independent residual maps and quantiles, extrapolation mask, and stability results.

PASS. Raw distortion/telecentricity meet any uncorrected limits; guard-banded independent calibration residual, scale drift, perspective, and principal- point stability each fit their allocated measurement budget over the entire qualified volume; all units pass with the released model/order; and relative regression is within margin.

FAIL. Any usable location/depth exceeds its geometry budget, a pass requires training/validation leakage or a different model per image, calibration is unstable after remount/temperature, or uncorrected geometry violates an algorithm assumption.

L08 – relative illumination, vignetting, and chief-ray compatibility

Objective. Verify usable image-circle coverage and stable brightness/colour across the sensor.

Procedure and conditions. Using a characterized uniform source, acquire dark-subtracted flat fields at each production aperture, focus/conjugate, wavelength/channel, and camera. Avoid source clipping and saturation. Rotate the lens/camera or use a calibrated source map to separate source, sensor PRNU, and lens shading. Inspect mechanical and optical vignetting, cat-eye pupil clipping, corner colour shift, dust sensitivity, and interaction with the candidate sensor’s microlens/CRA.

Measure relative illumination versus field, corner/centre ratio, asymmetry, colour shading, usable image circle, and stability with focus/aperture. Validate the proposed flat-field correction on independent captures, temperatures, and lens units.

Collect and output. Produce per-band raw/corrected illumination maps, radial and azimuth profiles, corner/centre ratios, image-circle boundary, colour-shading maps, calibration files, and independent residuals.

PASS. No active or measurement-critical region is mechanically clipped; raw relative illumination and colour shading meet their absolute/relative limits; corrected residual meets the application uniformity budget on every unit and condition; and correction does not create clipping or excessive noise. In the absence of an application value, raw corner/centre illumination of at least 70 % and corrected uniformity within \(\pm 5\) % are initial screening gates only, not universal qualification limits.

FAIL. Any required field is clipped, asymmetry indicates excessive decentre, raw/corrected uniformity or colour shading exceeds its limit, or a unit- specific correction is unavailable where one is required.

L09 – spectral transmission and colour balance

Objective. Confirm adequate in-band throughput, colour balance, and rejection of unwanted wavelengths.

Procedure and conditions. Measure lens transmission with a spectrophotometer or compare a calibrated detector/reference path over the production spectrum. Include coating-band edges, illumination lines, sensor-sensitive near-IR/UV, filters, and incidence angles relevant to the field. At the production camera, measure exposure needed to reach the operating signal, per-channel ratios, and SNR on actual materials at the motion-safe exposure limit.

Account for aperture and effective f-number at finite conjugates. Separate spectral transmission from relative illumination and sensor response.

Collect and output. Produce absolute or relative transmission spectra with uncertainty, in-band/out-of-band integrals, exposure/channel-ratio comparison, SNR by material/band, and baseline/candidate delta.

PASS. Every required band/material reaches its signal and SNR limit within the allowed illumination, aperture, gain, and exposure; prohibited out-of-band response is below its limit; colour/channel balance is calibratable within range; and transmission does not regress beyond the approved margin.

FAIL. A production band lacks throughput, motion-safe exposure cannot reach SNR, out-of-band leakage creates a false response, or a result relies on saturated channels or excessive gain.

L10 – flare, ghosting, scatter, and high-dynamic-range behaviour

Objective. Ensure bright sources, reflective parts, and off-axis illumination do not hide defects or create false ones.

Procedure and conditions. Image a dark field containing a controlled bright source at multiple intensities and angular positions: within the FOV, just outside each edge and corner, and at every known machine source/reflection direction. Repeat at production aperture, focus, filters, wavelengths, sensor, and exposure. Include glossy, curved, and highly reflective application parts.

Measure veiling-glare level, black-region rise, local contrast loss, ghost position/area/peak/energy, flare symmetry, saturation bloom/smear, and recovery after a bright frame. Compare with the baseline at matched object radiance and matched useful signal. Inspect lens barrels, spacers, filter surfaces, and adapters for responsible paths.

Collect and output. Produce annotated source-angle image sets, veiling-glare and contrast curves, ghost maps/catalogue, recovery timeline, worst-case application images, and ray-path hypothesis or KrakenOS model where useful.

PASS. No source angle or intensity produces a false feature, hidden required feature, saturation path, or unsafe decision; guard-banded veiling glare, ghost peak/energy, contrast loss, and recovery meet absolute limits; and none regresses beyond the approved baseline margin. Where no value exists, use no more than baseline as the screening gate until the application contrast budget is approved.

FAIL. Any prohibited false/hidden feature occurs, a flare metric exceeds its limit, or testing omits a credible production source angle/material.

L11 – aperture/focus adjustment, lock, and repeatability

Objective. Verify that production assembly and maintenance can set and retain the qualified optical state.

Procedure and conditions. For manual lenses, cycle focus and iris across their ranges, approach the production marks from both directions, apply the released setting method and lock torque, then measure aperture/effective f-number, focus position, FOV, principal point, and MTF. Include at least 30 set/lock/release cycles per unit for engineering screening. For motorized/tunable lenses, command repeated positions from both directions and after restart; measure backlash, repeatability, settling, temperature dependence, and position feedback.

Apply the production cable/hosing loads and camera orientation. Mark witness lines or tamper evidence where used. Check that locking focus or iris does not shift tilt, decentre, or the other adjustment.

Collect and output. Produce cycle-level settings, torque, focus/FOV/MTF distributions, hysteresis and backlash plots, lock-induced image shift, settling time, and work- instruction capability result.

PASS. Every setting/lock cycle returns within allocated focus, aperture, FOV, principal-point, and MTF limits; no lock-induced shift or cross-coupling exceeds its budget; motorized settling fits cycle time; and the documented method is repeatable across operators/units.

FAIL. Any cycle misses an imaging limit, engraved scales or feedback are insufficient for repeatable setup, lock torque damages/shifts the lens, or performance depends on unrecorded operator judgement.

L12 – thermal, vibration, shock, and contamination stability

Objective. Verify optical alignment and materials throughout handling and operation.

Procedure and conditions. Using the production mount/support, test at the qualified temperature and humidity extremes after soak. Measure focus, FOV/scale, principal point, distortion/calibration residual, MTF, relative illumination, and chromatic metrics at each plateau and during transition if the machine operates then. Apply required vibration and shock in the relevant orientations, then repeat optical tests. Perform approved dust, oil mist, humidity, cleaning-agent, coating-durability, or ingress tests when credible.

Inspect for focus/iris slip, loose elements, cement/coating change, condensation, contamination, scratches, fungus risk, thread damage, and permanent optical shift. Return to room condition and repeat the baseline sequence.

Collect and output. Produce environmental profiles, before/during/after optical tables and maps, focus/scale drift coefficients, physical inspection, photographs, and failure/event log.

PASS. Every CTQ remains within its guard-banded absolute limit throughout the claimed operating envelope; no relative regression exceeds its margin; vibration/shock/cleaning cause no slip, damage, contamination, or permanent change; and any required refocus/calibration interval is compatible with the maintenance plan.

FAIL. Any CTQ leaves its limit, the lens changes permanently, a lock slips, optical surfaces/materials are incompatible with the environment, or a pass depends on an unplanned manual adjustment.

L13 – unit/lot consistency and incoming inspection

Objective. Ensure evaluated performance is representative of routine supply.

Procedure and conditions. Test at least five candidate lenses from at least two lots where available. Compare L02–L12 results by serial/lot, emphasizing MTF field asymmetry, best-focus/field curvature, lateral colour, distortion, relative illumination, transmission, focus/iris torque, and cosmetic/defect inspection. Investigate outliers rather than deleting them.

Define incoming identity/cosmetic inspection plus a short optical test using the production or reference camera: grid/FOV, fixed-plane centre/corner MTF, flat field, and an application golden sample. Define storage, handling, cleaning, calibration, quarantine, and PCN requalification triggers.

Collect and output. Produce per-unit/lot interval plots, outlier investigations, estimated variation, approved serial register, incoming-inspection procedure and limits, golden images/data, and requalification matrix.

PASS. Every unit meets every individual hard limit; observed unit/lot variation fits inside the guard band; no unexplained asymmetry or bimodal population exists; and incoming tests can reject wrong, damaged, contaminated, misassembled, or optically degraded units before production.

FAIL. A unit passes only through group averaging or selected orientation, observed variation consumes the application margin, an outlier remains unexplained, or incoming control cannot detect a consequential change.

L14 – integrated application validation and production pilot

Objective. Prove that the candidate lens and camera-lens pair perform the real task.

Procedure and conditions. Lock the production focus, aperture, illumination, camera settings, calibration, and algorithm before final scoring. Execute configuration BC where valid and the proposed CC configuration. Use the blinded representative part set, nuisance variation, measurement-system analysis, and pilot procedure specified by C15 in Camera Replacement Qualification.

Challenge optical failure modes deliberately: smallest defects at centre and corners, object height/tilt limits, wavelength/colour extremes, gloss, reflective edges, flare-source angles, illumination drift, focus/temperature limits, motion direction, and remount after maintenance. Preserve native images and link every decision to a physical sample and condition.

Collect and output. Produce the locked recipe, part/image provenance, confusion matrices or measurement-system analysis, results by field/height/material/condition, worst-case images, production-rate pilot log, availability/cycle-time result, and failure review.

PASS. Every pre-registered class and condition meets its false-accept, false-reject, detection, measurement bias, repeatability, reproducibility, calibration residual, cycle-time, and availability limit with the approved confidence method; no critical pilot event occurs; and the result uses the same camera-lens-processing configuration intended for release.

FAIL. Any critical application or measurement limit fails, a field/colour/height stratum is hidden by an aggregate result, the final data influenced tuning, or the pilot differs from the locked production configuration.

Lens report checklist

The lens report is complete only when it includes:

• a one-page L01–L14 result matrix with worst unit and evidence links;

• baseline/candidate identity, serials/lots, camera/detector, aperture, focus, conjugates, wavelengths, and exact BB/BC/CC configurations;

• native target images or bench data and analysis settings sufficient to reproduce every derived metric;

• through-focus and fixed-plane MTF, chromatic, geometric, illumination, transmission, flare, adjustment, environmental, and application evidence;

• results in image micrometres, production-camera pixels, object-space units, and normalized field where relevant;

• explicit separation of lens-only/bench claims from camera-lens system measurements;

• every failure, retest, exclusion, deviation, corrective action, and unit/lot variation result;

• the production mounting/focus/aperture procedure, calibration, incoming test, maintenance limits, and signed G1–G4 decisions from Machine-Vision Camera and Lens EOL Qualification.