Modern manufacturing increasingly requires the measurement of complex surfaces, miniature features, free-form geometry and assemblies that cannot be inspected efficiently with conventional contact instruments. Optical, coordinate and non-contact metrology provide measurement methods in which geometry is inferred from light, imaging, position sensing or controlled probing. These methods are central to aerospace, automotive, biomedical, semiconductor, additive-manufacturing and precision-tool industries.

The term metrology denotes the science of measurement. In this chapter, measurement is treated not merely as obtaining a number, but as establishing a traceable relationship between a measured indication and the quantity intended to be measured. A reliable result therefore depends on instrument capability, calibration, datum selection, environmental control, workpiece preparation, software, operator decisions and uncertainty evaluation.

Learning Objectives

1. Measurement Foundations for Advanced Inspection

1.1 Contact and non-contact measurement

A contact instrument physically touches the surface, as in a micrometer, dial indicator or touch-trigger CMM probe. Contact methods are often highly accurate and comparatively insensitive to surface appearance, but they may deform soft parts, damage delicate coatings, require access to the measured feature and collect data relatively slowly.

Non-contact systems obtain information without mechanical contact. They include optical comparators, laser triangulation, interferometry, photogrammetry, structured light and machine vision. Their advantages include high data density, rapid acquisition and suitability for fragile or inaccessible surfaces. Their limitations include sensitivity to reflectivity, transparency, ambient light, occlusion, surface texture and optical alignment.

A hybrid instrument may use optical sensing for rapid scanning and a tactile probe for datum establishment or verification. Method selection should be based on the functional requirement, not on the apparent sophistication of the instrument.

1.2 Accuracy, precision and resolution

Accuracy describes closeness to the measurand's true or reference value. Precision describes the closeness of repeated results to one another. Resolution is the smallest displayed or detectable increment. A high-resolution instrument may still be inaccurate if it is poorly calibrated or incorrectly aligned.

where x̄ is the estimated measured value and U is the expanded uncertainty, generally obtained by multiplying a combined standard uncertainty by a coverage factor. The result is meaningful only when the measurand, method, units, environmental conditions and confidence convention are stated.

For independent standard uncertainty components u_i, the combined standard uncertainty is often estimated by the root-sum-square relationship:

This expression assumes that the components are independent and represented as standard uncertainties. Correlated components require covariance terms. An expanded uncertainty may be written U = k u_c, where k is a stated coverage factor, often approximately 2 for a near-95% coverage probability under suitable distributional assumptions.

2. Interferometry

2.1 Principle

Interferometry uses the superposition of coherent light waves. A beam is divided into a reference path and a measurement path; after reflection and recombination, the optical path difference produces bright and dark fringes. Movement of a surface changes the phase of the returning wave and therefore shifts the fringes.

For a displacement measured in reflection, one complete fringe corresponds to a path change of one wavelength, but the physical surface displacement is one-half wavelength because the light travels to the surface and back. Thus, for fringe order change N and wavelength λ,

where ΔL is surface displacement. This relation assumes a reflective measurement path, stable wavelength and correctly identified fringe order. In transmission or other optical arrangements, the path factor must be derived from the geometry rather than assumed.

where φ is phase difference in radians and Δ is optical path difference. The equation applies to monochromatic or sufficiently narrow-band coherent radiation in a medium where λ is the wavelength in that medium.

2.2 Types and applications

A Michelson interferometer divides and recombines a beam using a beam splitter and mirrors. It is used for displacement calibration, flatness and wavelength-related experiments. Laser interferometers are widely used to calibrate CMM axes, machine-tool positioning and precision stages. White-light or low-coherence interferometers determine surface height by locating the position of maximum fringe contrast; they are particularly useful for surface topography and step-height measurement.

Interferometers can provide nanometre-scale displacement sensitivity, but sensitivity is not equivalent to total accuracy. Air refractive-index variation, vibration, thermal expansion, beam misalignment, surface reflectivity and phase-unwrapping errors can dominate the uncertainty.

3. Optical Comparators and Machine Vision

3.1 Optical comparators

An optical comparator projects a magnified silhouette of a component onto a screen or camera. The projected profile is compared with a drawing overlay, master profile or digital CAD boundary. Typical uses include checking angles, radii, thread forms, stamped profiles and small machined components.

where M is magnification, h_i is image size and h_o is object size. If a feature measures 48 mm on the screen at 20× magnification, its corresponding object dimension is 2.4 mm. This simple relationship assumes linear, calibrated magnification and negligible distortion over the measurement region.

Errors arise from incorrect focus, parallax, lens distortion, edge blur, poor lighting, incorrect screen calibration and positioning the part away from the intended measurement plane. Digital comparators improve repeatability by using a camera and edge-detection software, but the edge algorithm remains dependent on illumination and surface condition.

3.2 Machine-vision systems

A vision system usually contains illumination, optics, an image sensor, processing software, fixturing and an output interface. Its workflow is:

Pixel-to-length calibration is not necessarily a single constant. Perspective, lens distortion and telecentric errors can make scale position-dependent. A calibration artefact with known features should therefore be measured across the field of view. Telecentric optics reduce apparent size changes caused by object-height variation and are valuable for dimensional inspection.

Vision systems are excellent for high-volume presence, orientation, surface-defect and two-dimensional dimensional checks. They are less suitable when the important characteristic is hidden, highly reflective without controlled lighting, or intrinsically three-dimensional unless multiple views or depth sensing are added.

4. Laser Measurement Systems

4.1 Laser triangulation

A laser line or spot is projected onto a surface and imaged by a camera at a known angle. Surface height changes the position of the image on the sensor. After calibration, image position is converted into distance. A scanning stage or moving workpiece converts profiles into a three-dimensional surface.

Triangulation accuracy depends on baseline length, angle, focal geometry, sensor resolution, calibration quality and surface interaction. Shiny surfaces may saturate the detector; dark surfaces may return insufficient light; translucent surfaces may produce ambiguous signals. Speckle, occlusion and shadowing can also create missing or noisy data.

4.2 Laser interferometric and scanning systems

Laser interferometers measure displacement through phase or fringe counting and are commonly used for calibration. Laser scanners, in contrast, may use triangulation or time-of-flight principles to collect dense surface data. Time-of-flight distance is based on propagation time, but practical performance is affected by timing resolution, reflectivity and atmospheric conditions.

The appropriate distinction is therefore between the sensing principle and the application: a laser may serve as a calibration reference, a line scanner, a displacement sensor or a position encoder.

5. Coordinate Measuring Machines

5.1 Construction and operation

A CMM determines the coordinates of points on a workpiece relative to a machine coordinate system. The principal architectures are bridge, gantry, cantilever, horizontal-arm and articulated-arm machines. A probe may be touch-trigger, scanning, optical or laser-based.

A coordinate point is represented as P = (x,y,z). A measured distance between points P₁ and P₂ is:

This is the Euclidean distance in a Cartesian coordinate system. It assumes both points are expressed in the same orthogonal coordinate frame and that the intended characteristic is point-to-point distance.

5.2 Alignment and datums

An alignment transforms machine coordinates into a part coordinate system. A common three-dimensional datum alignment uses a primary plane to constrain translation and rotation, a secondary feature to constrain additional degrees of freedom, and a tertiary feature to establish the remaining direction or origin. The purpose is not simply to make the part visually level; it is to reproduce the design-intent datum reference frame.

Poor alignment can create a false failure. For example, a hole pattern may appear displaced if the part is aligned to an arbitrary edge rather than to the functional datum structure specified by the drawing. Datum simulators, fixture repeatability and datum mobility must be considered when interpreting geometric tolerances.

5.3 Feature fitting

Measured points are fitted to ideal geometric entities. A circle may be fitted by least squares, minimum zone, maximum inscribed or minimum circumscribed criteria, depending on the inspection requirement. These criteria do not produce identical results. The fitting method must therefore be specified, especially for form and size decisions.

For a set of measured points, least-squares fitting minimizes the sum of squared residuals. It is useful for estimating a central trend but does not necessarily represent the worst-case form error. Geometric product specifications may require a minimum-zone evaluation instead.

5.4 CMM errors and verification

Sources of error include scale error, squareness error, probe qualification, probe-lobing, thermal expansion, machine vibration, fixture deformation, stylus bending, controller interpolation and software settings. CMM performance is verified with calibrated artefacts such as gauge blocks, ring gauges, ball bars and step gauges. Verification should be distinguished from calibration: verification demonstrates conformity to specified performance, whereas calibration establishes the relationship between indication and reference values.

6. Structured Light and Three-Dimensional Scanning

Structured-light systems project coded patterns—stripes, grids or phase-shifted fringes—onto an object and use one or more cameras to reconstruct surface coordinates. Because the projector and camera have known geometry, correspondence between projected and observed patterns allows triangulation.

A 3D scanner may produce a point cloud, a collection of points (x_i,y_i,z_i). The cloud can be filtered, aligned, meshed and compared with nominal CAD geometry. Alignment methods include datum alignment, best-fit alignment and iterative closest point (ICP). Best-fit and ICP minimize an overall discrepancy, but they can conceal a functionally important datum shift. They should not replace drawing-specified datum alignment when acceptance depends on functional references.

Scanning quality is affected by point spacing, field of view, calibration, motion, occlusion, surface reflectance, ambient light and registration error. A dense cloud is not automatically an accurate cloud. Data density, spatial resolution, trueness and uncertainty are separate properties.

6.1 Deviation maps

A deviation map assigns a signed distance from measured points to a nominal surface. The sign convention must be stated—for example, positive may indicate material outside the nominal surface. Colour bands are useful for visual diagnosis, but pass/fail decisions should use numerical tolerances and a defined evaluation rule. Excessive smoothing, hole filling or mesh editing can remove real defects and must be controlled and documented.

7. Inspection Planning and Data Interpretation

An inspection plan links product requirements to measurement methods. A practical plan contains:

Plan element	Required decision
Characteristic	What dimension, form, location or defect is measured?
Specification	What nominal value, tolerance and datum structure apply?
Method	Which sensor, probe, optics and software are used?
Sampling	Which parts, features and points are measured?
Environment	What temperature, vibration and lighting controls are required?
Traceability	Which calibration artefacts and standards support the result?
Decision rule	How are uncertainty and borderline results treated?
Record	What data, alignment, program version and operator information are retained?

7.1 Selecting a method

Select a CMM when datum-related three-dimensional geometry and high accuracy are dominant. Select machine vision for rapid two-dimensional inspection and pattern-based decisions. Select structured light or laser scanning for free-form surfaces, reverse engineering and high-density deviation analysis. Select interferometry for very small displacement, flatness or precision calibration. Select an optical comparator for profile comparison where silhouette information is sufficient.

Important selection criteria include tolerance-to-uncertainty ratio, surface accessibility, part size, inspection rate, automation requirements, material reflectivity, allowable contact force and data volume.

7.2 Sampling and repeatability

Measurement points must represent the feature. A circle cannot be reliably characterized from two points; a plane requires spatially distributed points; a free-form surface requires a sampling strategy related to curvature and tolerance wavelength. Repeated measurements under unchanged conditions estimate repeatability. Measurements by different operators, setups or days contribute reproducibility information.

A useful gauge capability index is the ratio of measurement variation to tolerance, but no universal acceptance threshold should be assumed without the applicable quality system or customer requirement. The study design, number of trials and statistical method must be stated.

7.3 Interpreting results

A result near a tolerance limit should not be interpreted without uncertainty. If the measured value is x and the upper specification limit is USL, the apparent margin is USL−x, but the decision risk depends on the uncertainty interval and the selected decision rule. A conservative guard band may prevent accepting conforming-appearing product when measurement uncertainty overlaps the limit, although it can also increase false rejection. The rule must be agreed before inspection.

8. Worked Example: Interferometric Displacement

An interferometer in reflection records a shift of 125 complete fringes using a laser with wavelength 632.8 nm. The surface displacement is:

The result assumes that all counted fringes correspond to the same direction of surface movement, the wavelength is specified in the measurement medium, and no fringe-counting ambiguity exists. If the instrument instead measured a one-way optical path directly, the factor of 1/2 would not apply; the optical configuration must therefore be established before calculation.

9. Applied Case: Inspection of an Additively Manufactured Housing

Background

This illustrative case concerns an aluminium additively manufactured housing containing two precision bearing seats, a sealing face and several free-form cooling channels. The drawing specifies datum-related bore locations, sealing-face flatness and a CAD-defined external profile.

Problem

A conventional CMM can measure the bores and sealing face accurately, but access to the internal channels is limited. A full optical scan can capture the external profile rapidly, yet the dark, partially reflective surface and internal occlusion produce incomplete data.

Analysis

The inspection team divides the requirements by function. The bores and sealing face control assembly performance and require traceable datum-related measurements. The external form requires dense comparison with CAD. Internal channels require either sectioning of a qualification part, computed tomography or an accessible optical strategy.

Approach and solution

The part is temperature-conditioned and mounted on a fixture that reproduces the primary datum without overconstraint. A calibrated CMM touch probe establishes the primary datum plane, locates the secondary bore pattern and measures bore size and position. A structured-light scanner captures the external surface. Targets and controlled repositioning permit multiple views to be registered. The scans are merged, filtered using a documented rule and aligned to the established datum frame rather than by unconstrained best fit. A deviation map is generated, while numerical acceptance uses specified tolerances and sampling rules.

A gauge study is performed for the bore measurement, and the scanner is verified with a calibrated artefact before use. Reflective regions are treated with an approved removable coating only after confirming that coating thickness will not affect the result. The report records environmental conditions, calibration status, software version, alignment method, point spacing and uncertainty information.

Outcome and learning points

The hybrid plan provides reliable functional measurements and efficient surface coverage. It also reveals that a visually attractive colour map cannot substitute for datum-controlled measurement. The case demonstrates that inspection planning is a systems problem: sensor selection, fixturing, coordinate strategy, data processing and decision rules must be designed together.

10. Limitations, Governance and Contemporary Developments

Optical and scanning measurements may be affected by privacy or cybersecurity concerns when production geometry is transferred to cloud platforms. Measurement programs and processed point clouds are intellectual property and should be access-controlled. Automated defect classification using machine learning requires representative training data, validation against expert decisions and monitoring for changes in material, lighting or product design. An AI system should support—not silently replace—the approved measurement method and engineering decision authority.

Relevant quality practice includes calibration traceability to national or international standards, documented environmental control, measurement-system analysis, configuration control of inspection programs and retention of raw data. Standards such as ISO 17025, ISO 10360, ISO 14253 and GPS-related specifications may be applicable depending on the laboratory, CMM and product requirement; the current edition and contractual requirements must always be checked.

Emerging developments include high-speed inline 3D inspection, multisensor CMMs, optical coherence methods, digital twins, uncertainty-aware CAD comparison and closed-loop manufacturing. These developments improve speed and data availability, but they increase the need for validated algorithms, robust data governance and competent interpretation.

Best-Practice Checklist

Chapter Summary

Optical, coordinate and non-contact metrology extend measurement beyond conventional contact gauges. Interferometry converts phase or fringe movement into highly sensitive displacement information. Optical comparators and machine vision provide rapid profile and feature inspection when illumination, calibration and edge definition are controlled. Laser triangulation and structured light reconstruct surface geometry, while CMMs establish traceable three-dimensional coordinates through probes, datums, alignments and fitted features.

No measurement method is universally superior. Accuracy, uncertainty, surface condition, access, speed, data density and functional datum requirements determine the appropriate technology. A defensible inspection plan defines the measurand, method, sampling, environment, calibration, data processing and decision rule. Dense point clouds and attractive colour maps are useful only when their coordinate frame, registration, filtering and uncertainty are understood. Effective metrology therefore combines physical principles, geometric reasoning, statistical interpretation, software control and professional judgement.

Key Terms

Optical path difference: Difference in distance travelled by interfering optical waves.

Machine vision: Automated image-based sensing and interpretation for inspection.

Telecentric optics: Optics designed to reduce magnification change with object distance.

Laser triangulation: Distance measurement using a projected laser and geometric imaging.

Coordinate measuring machine: Machine that determines locations of points in a coordinate frame.

Structured light: Projected coded illumination used for three-dimensional reconstruction.

Best-fit alignment: Alignment minimizing an overall discrepancy between measured and nominal geometry.

Traceability: Unbroken documented relationship to recognized measurement standards.

Expanded uncertainty: Combined standard uncertainty multiplied by a stated coverage factor.

Decision rule: Prescribed method for accepting or rejecting a result while considering uncertainty.

Review Questions

Short Answer Questions

Long Answer Questions

Analytical / Application Questions

Multiple Choice Questions with Answers

Question 1

In a reflection interferometer, one complete fringe change corresponds to a surface displacement of:

Question 2

Which instrument is most appropriate for rapid two-dimensional inspection of a stamped profile?

Question 3

Question 4

Question 5

If 125 fringes are counted with λ = 632.8 nm in reflection, the displacement is:

Question 6

Which factor is especially important in machine vision because it determines feature contrast?

Question 7

Question 8

The relationship d = √[(x₂−x₁)² + (y₂−y₁)² + (z₂−z₁)²] calculates:

Question 9

Which method is generally most suitable for dense measurement of an accessible free-form exterior?

Chapter 8 Optical, Coordinate and Non-Contact Metrology

Optical, Coordinate and Non-Contact Metrology

Introduction and Context