Metrology is the science of measurement. In industry, it ensures that the measurements used to manufacture, inspect and accept a product are reliable: suitable instruments, calibrated, traceable to references, with a known uncertainty. It is the basis of quality control, which decides whether a part conforms on the strength of these measurements.
Measuring means comparing
For LNE, the French national metrology and testing laboratory, every measurement compares an unknown quantity with a known reference, using a reliable instrument. A measurement result has three elements:
- a numerical value;
- a unit, from the International System of Units (SI);
- an uncertainty, which quantifies the doubt about the measured value. Without uncertainty, two results cannot be compared.
The three branches of metrology
- Scientific metrology: national laboratories realise, maintain and transfer measurement references. In France, LNE leads the national metrology network.
- Legal metrology: a state mission that ensures the reliability of measurements used for trade or regulation and prevents fraud. It covers, for example, retail scales, fuel pumps, water meters and speed cameras.
- Industrial metrology: the control of measurements within the company, for inspection and manufacturing. It is linked to quality management.
LNE points to two industrial stakes: without reliable measurements, a customer cannot be sure that subcontracted parts meet its requirements; and a plant that measures its production closely reduces the cost of scrap and rework.
Calibration, traceability and uncertainty
- Calibration compares an instrument with a reference standard. It tells you about the instrument's condition at the time it is carried out and about its past drift, but does not predict its future behaviour: hence regular calibration, at a suitable interval.
- Metrological traceability: a succession of calibrations, with ever smaller uncertainties, links the shop-floor instrument to the national reference, then to the SI. It is an unbroken chain. Between the plant and the national laboratory are accredited laboratories, each in a defined field.
- Measurement uncertainty is calculated through a budget that takes every source of error into account: reference, measuring range, test conditions. The method is described in the Guide to the expression of uncertainty in measurement (GUM), published under the auspices of the International Bureau of Weights and Measures (BIPM). LNE provides free software for evaluating uncertainties.
For LNE, the uncertainty required is first and foremost the one that meets the need: there is no point measuring a roof frame to the nanometre.
What the ISO 9001 standard requires
Clause 7.1.5 of ISO 9001:2015 covers monitoring and measuring resources. The company must:
- provide the resources that give valid and reliable results when a measurement is used to verify product conformity;
- ensure they are suitable for the type of measurement and maintained;
- retain documented information as evidence;
- identify each instrument subject to traceability, to know the status of its calibration.
The Collège français de métrologie, the French metrology association, explains that these resources are all the components of the measurement process, the "5Ms": manpower, means, milieu (environment), method and material. To demonstrate reliability, repeatability and reproducibility tests are often carried out. The capability of a measurement process is frequently calculated as the ratio between the part tolerance and the expanded measurement uncertainty.
Measuring equipment in the plant
- The coordinate measuring machine (CMM) records the coordinates of points on a part and compares them with the nominal dimensions. Its probes touch the part point by point, scan it continuously or measure it without contact, optically. Software calculates deviations in form, position, parallelism or coaxiality against the tolerances.
- Optical scanners (laser, structured light), trackers and photogrammetry measure large or complex parts without contact.
- Machine vision inspects the appearance, presence or position of parts on the line: our machine vision guide covers this part.
Metrology and quality control: measure well to produce well
Quality control decides whether a product conforms. Metrology ensures that this decision rests on accurate measurements. The Collège français de métrologie puts forward a simple idea, "measure well to produce well": a measurement used only to declare the conformity of the finished product adds little; it becomes valuable when it is used to control the process and detect an anomaly before the part is finished. Measurement data then feed into production software: our MES software guide presents these tools.
Quality control with AI
AI is mainly used for visual inspection. Software connected to cameras learns to recognise conforming parts, since defects are rare, then flags those that deviate. The French start-up Visionairy, for example, rejects only the defective parts rather than the whole batch, and cross-references data from several lines to trace the origin of a defect (Techniques de l'Ingénieur, 21/05/2024).
What manufacturers said at Tech for Industry Show
At Tech for Industry Show 2026, at Paris Expo Porte de Versailles:
- L'Oréal presented in-line quality control using vision and AI, able to spot defects on transparent glass jars, previously hard to detect, and to characterise the defect in the flow to correct it faster. Read the article; the keynote is available as a replay: L'Oréal operations: turning complexity into a sustainable advantage.
- Renault said it has deployed more than 6,400 vision-based quality checks in its plants, built on the data of its digital twins. Read the article
- Forvia has been using machine learning for more than five years for quality inspection by image and by sound. Read the article
An example in France
Merck opened in May 2026 in Saint-Ismier, near Grenoble, a 4,500 m² site dedicated to semiconductor metrology and inspection, an investment of 20 million euros. Its tools measure the depth, thickness or height of structures to adjust processes in real time, and its inspection systems detect chip defects (Le Journal des Entreprises, 12/05/2026).
The market
Global Market Insights estimates the global industrial metrology market at 15.97 billion dollars in 2025, including 3.61 billion for coordinate measuring machines and 4.15 billion for optical scanners. Other firms give different amounts, depending on the scope.
The 2027 exhibitor equipping quality control
To meet at Tech for Industry Show 2027:
- Azumuta: a platform that digitises shop-floor work, including work instructions, quality control and audits.
How to organise metrology in a plant
1. List and identify the instruments whose results are used to accept a product or control a process.
2. Define the need: the tolerance of each characteristic and the acceptable uncertainty.
3. Calibrate regularly, with traceability to national references, through an accredited laboratory where necessary.
4. Assess the capability of measurement processes, through repeatability and reproducibility tests.
5. Train everyone who measures, not only the metrology department.
6. Measure upstream of final inspection, to correct the process before the part is finished.
Frequently asked questions
What is industrial metrology?
The control of measurements within the company: suitable instruments, calibrated and traceable to references, with a known uncertainty, to manufacture and inspect.
What are the three types of metrology?
Scientific metrology (references), legal metrology (measurements for trade or regulation) and industrial metrology (the company's measurements).
What is the difference between metrology and quality control?
Quality control decides whether a product conforms; metrology ensures that the measurements behind this decision are reliable.
Which standard governs metrology in companies?
ISO 9001:2015, clause 7.1.5, for monitoring and measuring resources and their traceability.
How is AI used in quality control?
Mainly for visual inspection: software connected to cameras learns to recognise conforming parts and flags those that deviate.
