3D visual measurement technology allows a video borescope to calculate real dimensions from surfaces inside engines, turbines, pipes, castings, vessels, and other inaccessible components. The two principal methods are 3D Phase Measurement and 3D Stereo Measurement. Both create a three-dimensional point cloud, but they collect surface data differently and respond differently to distance, reflection, movement, surface detail, and probe access.
3D visual measurement converts inspection imagery into X, Y, and Z surface coordinates. The inspector selects a measurement type, places or verifies cursors, checks the point cloud and reference geometry, and compares the result with the applicable maintenance or acceptance limit. The technology is useful when an inspection must establish how long, deep, wide, large, or far from a reference a visible feature is.
This guide explains the complete 3D visual measurement workflow, from choosing a technology to validating and documenting the final result. For the broader decision about whether an inspection needs dimensional data at all, read AIT's guide to when 3D borescope measurement is needed.
The Two Main 3D Borescope Measurement Technologies
Video borescope measurement methods differ in how they obtain scale and three-dimensional position. Choosing a method begins with the inspection problem: the access diameter, viewing angle, feature geometry, surface texture, reflectivity, target distance, expected movement, and required result.
From Inspection Image to 3D Measurement
A valid measurement follows a clear sequence. The probe captures optical information from the target, and calibration data relates that information to the geometry of the camera and tip. Processing then calculates surface coordinates or selected measurement points. The operator chooses the measurement type that matches the dimension being inspected.
-
Locate the indication. Capture enough surrounding geometry to understand its position and reference surfaces.
-
Select the measurement technology. Match structured light, stereo, or comparison methods to the access and surface.
-
Control the capture. Manage distance, focus, lighting, glare, angle, probe motion, and tip cleanliness.
-
Calculate the surface. The system produces point coordinates, a full point cloud, or matched measurement points.
-
Select the measurement type. Choose length, depth, area, profile, plane, angle, clearance, or another supported method.
-
Validate the geometry. Review point-cloud shape, masks, MTD, planes, matches, and cursor placement.
-
Document the result. Save the image, measurement data, location, units, setup, procedure, and relevant limit.
How 3D Stereo Measurement Works
A 3D Stereo tip contains two calibrated optical paths that view the target from slightly different positions. Surface details shift horizontally between the two views. The system matches corresponding features and uses the known tip geometry to triangulate their X, Y, and Z coordinates.
Traditional Stereo calculates coordinates at manually matched cursor positions. 3D Stereo extends the method by generating a full point cloud that can be rotated and inspected. The operator can compare the camera image with the calculated surface, identify incorrect geometry, and refine cursor placement.
3D Stereo needs recognizable surface detail. A uniform surface gives the matching algorithm little information, while repeating patterns can create ambiguous matches. Sharp focus, close working distance, controlled glare, and visible texture improve the result. It can handle some minor movement because both views are captured through the stereo optic rather than through a longer structured-light sequence.
3D Stereo matches visible surface features in two calibrated views to calculate a full point cloud.3D Visual Measurement at a Glance
How 3D Phase Measurement Works
A 3DPM tip contains viewing optics, illumination, and a structured-light projection system. During capture, the tip projects a sequence of line patterns across the target. The lines appear displaced where the surface rises, falls, curves, or breaks. The instrument evaluates those phase shifts together with the calibrated geometry of the tip and camera.
The output is a dense set of calculated surface coordinates. Each usable image location is associated with an X, Y, and Z position. Together, those coordinates form the 3D point cloud used to calculate measurements. The white-light image remains useful for recognizing the indication, but the measurement is calculated in three dimensions rather than by treating the displayed image as a flat plane.
That is why a straight 3D measurement line can appear curved in the 2D image. Perspective and optical distortion change how the line looks on screen, while the result is calculated from the underlying 3D geometry.
A 3DPM tip projects structured-light patterns whose displacement changes with the shape of the surface.What phase shift analysis contributes
3DPM projects line patterns through a series of LEDs and calculates measurements using phase shift analysis with calibrated processing. The system analyzes a controlled optical pattern whose displacement changes with surface geometry. It does not estimate depth from image brightness alone.
Structured light needs a surface that returns enough usable pattern information. Mirror-like reflections can overwhelm parts of the pattern. Deep recesses can hide the bottom from the camera or projector. Shadows can interrupt coverage. These conditions do not always invalidate the entire capture, but they can create areas of reduced or missing 3D data that the operator must identify.
The Measurement-on-Demand Workflow
Traditional stereo workflows may require the inspector to locate a feature with a general viewing tip, withdraw the probe, install a calibrated measurement tip, return through the access route, and find the feature again. 3DPM can create a three-dimensional surface scan with the same compatible tip used to inspect the area. This same-tip workflow remains a defining 3DPM capability.
With a compatible 3DPM setup, the inspector can navigate, inspect, and trigger a 3D capture when a relevant indication appears. This reduces the chance of losing the location, changing the viewing angle, disturbing the component position, or spending additional time retracing a difficult access path.
Measurement on demand still requires a stable capture, suitable tip-to-target distance, usable structured-light coverage, the correct measurement type, and a reviewed point cloud. The feature becomes measurable only after the operator captures and validates the 3D data.
Using the 3D Point Cloud to Validate a Measurement
A 3D point cloud shows the calculated surface behind the measurement. The inspector can rotate it, inspect contours, check noise, review the reference plane, and confirm that each cursor sits on the intended feature. This makes it possible to check the geometry behind the number instead of relying on the displayed value alone.
A cursor can appear to sit on a shroud in the white-light image while resting on the blade in three-dimensional space. Rotating the point cloud exposes the misplaced cursor. Moving it only a few image pixels can change the reported clearance substantially.
The point cloud reveals the geometry behind the displayed measurement and helps expose misplaced cursors.Full image point cloud and measurement point cloud
The full image point cloud helps the inspector understand overall surface geometry, tip orientation, coverage, and broad data anomalies. A measurement point cloud narrows the view to the active measurement and nearby surface. Its depth map can show perpendicular distance from the measurement's reference plane, which makes small pits, raised weld features, or noisy data easier to distinguish.
A point cloud should resemble the component. Unexpected spikes, warped planes, disconnected islands, or surface points bridging a real gap are warning signs. A clean-looking white-light image is not enough if the underlying surface is noisy around the measurement.
MTD and Usable Measurement Range
MTD means maximum target distance. It identifies the distance from the tip to the surface point farthest from the tip among the points used by the active measurement. Treat MTD as a setup indicator rather than an accuracy guarantee. A smaller MTD generally means the target occupies more pixels and the projected patterns contain more spatial information, although feature shape, surface finish, measurement type, and viewing angle still affect the result.
Measurement depth of field is the tip-to-target range over which a given tip can collect useful measurement data under stated conditions. Published depth-of-field values are useful for tip selection, but the working range still changes with feature size, required accuracy, surface finish, and measurement type.
A published measurement range is only a starting point. The capture must contain enough clean 3D data around the feature and reference surface to support the required decision. Move closer when the application safely allows it, then verify the point cloud and measurement setup.
3DPM Tip Selection and Pattern Orientation
Tip selection begins with two questions: from what direction must the feature be viewed, and from what distance can the probe reach it? The access port and target geometry determine whether a forward-view or side-view tip is appropriate. The expected working distance, feature size, and measurement method narrow the optical choice.
For a narrow surface, pattern orientation matters. The Waygate guidelines state that the long dimension of the surface should be perpendicular to the projected pattern lines. Forward-view black, orange, and yellow tips and the green side-view tip project horizontal lines, so narrow targets are best oriented vertically in the image. Blue and red side-view tips project vertical lines, so narrow targets are best oriented horizontally.
Tip and probe optics must be clean, and the tip must be securely attached. Compatible 3DPM tips are automatically recognized by supported systems, but recognition does not replace a verification check or application-specific procedure.
3D Measurement Types and What Each One Measures
Choose the measurement type that matches the inspection question. The cards below mirror the measurement selector used in AIT's 3D borescope guidance.
Depth and Depth Assist
A Depth measurement uses three points to define a reference plane and a fourth point to report perpendicular height or depth. A reference cursor placed on a curved, damaged, or different surface can tilt the plane and produce a plausible but incorrect number. Depth Assist can search nearby surface data and propose the deepest point, highest point, or blade-tip point. The operator must still review the proposed location.
Depth uses three points to establish a reference plane and a fourth point to report perpendicular depth.Area Depth Profile
Area Depth Profile evaluates a series of profile slices across an operator-defined area and identifies the slice containing the highest or lowest point. This is better suited than a single profile line when the operator needs the maximum pit depth across a corrosion field or the maximum height across a weld region. The search area should include the whole feature without pulling unrelated geometry into the analysis.
Area Depth Profile searches a selected region and identifies the profile containing its maximum height or depth.Measurement Plane: extending a valid reference surface
A Measurement Plane is used with another measurement type. Three cursors establish a mathematical plane on a valid surface, allowing measurement cursors to be projected where direct 3D data is missing or too noisy, such as beyond a broken blade edge. The green 3D surface mask shows points close to the plane. Broad green coverage should surround the feature on the relevant local reference surface. Distant curved regions should remain outside that reference.
For missing blade material, an operator can position the plane near the damaged corner, maximize the green mask on the surviving local surface, and use area and projected edge lines to estimate the missing geometry. The fitted plane provides a defined geometric reference for the reconstructed edge. Its validity still depends on correct placement across the surviving blade surface.
The green mask shows surface data close to the plane defined by the three reference cursors.Blade Tip Clearance
Blade Tip Clearance automatically identifies the blade and liner, maps the blade edge, and calculates minimum, maximum, and average clearance across the mapped region. Both 3DPM and 3D Stereo can support the method. A compatible 3DPM tip with a 105-degree field of view can include more of a blade edge from a given distance, while 3D Stereo may work better on shiny compressor blades and liners.
Inspectors should verify that the reference plane aligns with the liner, that the mapped line follows the real blade edge in both 2D and 3D, and that noisy or missing end regions are excluded. Application limits include honeycomb liners and certain irregular or shrouded blade-tip geometries. Advisory messages are measurement warnings and require operator review.
3D Stitching for larger scenes
3D Stitching combines up to ten compatible 3D Phase or 3D Stereo images into a composite image on supported systems. It can help measure a large defect, establish its location relative to fixed component features, or reduce uncertainty by combining close, high-quality captures instead of relying on one distant image. All images in the composite must come from the same probe and tip, and the finished stitch must be checked for alignment artifacts before measurement.
Use at least 50% overlap between images. Avoid rotation steps of about 20 degrees or more and magnification changes of about 30% or more between captures. These capture controls improve registration between images but do not guarantee measurement accuracy.
3D Stitching combines overlapping close-range captures into a larger measurable scene.3D Phase Measurement vs 3D Stereo Measurement
Both methods create a point cloud, but they obtain surface coordinates differently. 3DPM projects structured-light patterns and analyzes their phase shifts. 3D Stereo uses two calibrated optical views and image matching to calculate surface geometry.
Neither method is universally more accurate. Accuracy and minimum measurable feature size depend on the application, and performance changes with feature type, distance, surface, and setup. For a critical application, use representative samples and a Gauge R&R study or equivalent method to determine whether the complete inspection process meets the required accuracy and repeatability.
Video Borescopes for 3D Measurement
Mentor Visual iQ+: 3D Phase Measurement / Real3D Measurement
The Mentor Visual iQ+ VideoProbe is AIT's primary option for inspections that require 3D Phase Measurement (3DPM) and Waygate Real3D measurement. It is a strong fit for measurement-on-demand workflows, wider inspection scenes, irregular damage, and applications where the inspector needs to review a full 3D point cloud before accepting the result.
Mentor Flex+: Real3D Stereo Measurement
The Mentor Flex+ VideoProbe is the AIT option to consider when the inspection calls for Real3D Stereo Measurement. Stereo measurement works well for discrete features such as cracks, edges, gaps, and localized damage, especially when the surface has enough detail for reliable feature matching.
Where Traditional Stereo and Comparison Measurement Fit
Traditional Stereo Measurement
Traditional Stereo provides useful context for 3D Stereo because the two methods use the same left-view and right-view principle. Traditional Stereo calculates geometry only where the operator matches cursors, so it does not produce the full point cloud available with 3D Stereo. The target must remain in focus, the tip calibration must match the stereo optic, and the surrounding texture must confirm that the cursors identify the same physical point in both views.
Comparison Measurement
Comparison Measurement is a two-dimensional method for approximate sizing. It scales an indication against a known-size reference in the same image. The reference and indication must be on the same plane and at the same tip-to-target distance. Comparison does not provide surface depth, a rotatable point cloud, or the geometric checks available with 3D Phase and 3D Stereo.
Selecting the Right 3D Measurement Technology
Use this table as a starting point. Representative application testing should confirm the final method, tip, measurement type, and working distance.
Reading the 3D Data Masks
Yellow data
Yellow highlights lower-quality 3D data, often caused by reflection or shadowing. Avoid placing measurement cursors in yellow areas when possible, especially for Depth, Depth Profile, and Area Depth Profile. A new capture from a different orientation can redirect reflections or improve pattern coverage.
Red data
Red shows locations where the system could not determine 3D coordinates. Direct measurement cannot be taken in those areas. Data near the boundary also deserves close review because noise or interpolation can influence the apparent surface. A Measurement Plane may support certain edge or missing-material measurements, but it does not create direct surface data where none was captured.
Green surface mask
The green 3D surface mask is measurement-specific. It highlights surface points close to the active reference plane and helps the operator judge plane alignment and cursor placement. Useful coverage follows the relevant local reference surface. Green masking across unrelated curved or stepped geometry can indicate a poorly defined plane.
Common 3D Phase Measurement Errors
Common 3D Stereo Measurement Errors
3D Stereo depends on accurate feature matching between two calibrated views. A sharp image can still produce weak surface data when the target has little texture, a repeating pattern, heavy glare, or details that are difficult to match between the two optical paths.
3D Stereo needs sharp focus and recognizable surface detail so the system can match points between the two calibrated views.Field Checklist for Validating 3D Measurement Results
-
Define the required dimension. Know whether the decision depends on length, perpendicular depth, planar area, maximum depth across an area, clearance, or reconstructed missing geometry.
-
Confirm operator qualification and procedure. Follow the organization's approved measurement process and application criteria.
-
Select the viewing direction and working range. Match forward or side view, pattern orientation, feature size, surface, and access route.
-
Clean and secure the optics. Inspect the probe lens and tip, then confirm the tip is fully attached.
-
Verify the system. Use the applicable traceable verification block and prescribed before-and-after checks.
-
Move close enough. Fill the image with the feature and its reference surface without losing context.
-
Control reflection and shadow. An off-perpendicular view can improve 3DPM depth data on shiny surfaces by directing mirror reflections away from the camera.
-
Hold the probe still. Stabilize the tip throughout the structured-light capture sequence.
-
Review masks before placing cursors. Avoid yellow data where possible and do not place direct measurements in red areas.
-
Rotate the point cloud. Confirm the shape resembles the component, the plane is aligned, and no cursor sits on the wrong surface.
-
Use the depth map for small height or depth results. The feature should stand out from the local 3D noise.
-
Inspect automated results. Check Depth Assist, Auto Area, Blade Tip Clearance, and other proposed geometry before acceptance.
-
Save the evidence. Retain the measurement image, result, cursor placement, 3D data, location, procedure reference, and reviewer notes required by the inspection program.
Accuracy: What Can and Cannot Be Claimed
A single accuracy number cannot describe every 3DPM application. Accuracy depends on target distance, feature size, surface reflectivity, surface texture, capture angle, pattern coverage, probe stability, measurement type, cursor placement, and the quality of the reference surface.
Under good conditions, Real3D measurement technologies are often capable of resolving feature depths around 0.125 mm (0.005 in.) and achieving errors around plus or minus 0.05 mm (plus or minus 0.002 in.) or better. These are conditional capability figures, not guaranteed performance for every component. Use a Gauge R&R study with representative known samples to establish application-specific accuracy and precision.
Do not treat extra decimal places on the display as extra measurement confidence. A result with several decimal places can still be wrong if the reference plane is tilted, the cursor touches noise, or the feature sits outside usable 3D data.
Practical Applications
Corrosion, erosion, and pitting
Depth can measure an isolated pit when a sound local reference surface surrounds it. Area Depth Profile is better when the deepest point could lie anywhere within a larger damaged field. Area or Auto Area can quantify affected surface coverage, but the boundary must be tied to a documented inspection criterion.
Weld and wear-groove assessment
Depth measures a selected high or low point relative to a plane. Depth Profile shows one cross-section. Area Depth Profile searches multiple slices to find the maximum result across a region. The right choice depends on whether the procedure calls for one specified location or the worst condition in an area.
Blade edge and missing-corner damage
Point to Line can measure damage from a surviving reference edge. A Measurement Plane can reconstruct the local plane through a region where the original edge is missing. Missing Corner tools on supported systems can report area and edge lengths, but the plane and projected edge extensions must match the surviving blade geometry.
Blade-to-shroud clearance
Automated Blade Tip Clearance can replace repeated manual depth placements across an edge, but automation shifts the operator's task from placing every point to validating the liner plane, blade-edge map, excluded regions, and advisory messages.
Documentation and Offline Review
A useful inspection record needs more than the final number. Save the original measurement data when possible, because supported desktop software can reopen measurement images, move or add cursors, clear measurements, and review the point cloud. The record should also identify the component location, measurement type, units, tip and probe configuration, inspection procedure, relevant limit, and any conditions that affected confidence.
Re-measurement does not repair a weak capture. If reflection, motion, distance, or missing structured-light coverage corrupted the original data, the correct action is a new field capture.
Choosing a Video Borescope with 3D Measurement
Define the inspection conditions before comparing instrument features. Record the access diameter and length, viewing direction, target distance, surface finish, required measurement types, reporting workflow, and the smallest feature that can change the maintenance decision.
Use those requirements to confirm the functions available on the proposed instrument, probe, tip, software version, and license package.
For 3D measurement work, start with the measurement method and access requirements. The Mentor Visual iQ+ video borescope is the primary AIT option for 3D Phase Measurement / Real3D workflows, while the Mentor Flex+ VideoProbe is the option to consider for Real3D Stereo Measurement. Final capability still depends on the probe, optical tip, software version, and license package, so match the configuration to the access route, component geometry, surface condition, feature size, and required measurement.
Waygate Technologies measurement video