Each publication stems from industry studies, field research and analysis of global trends integrated with knowledge and expertise gained from transformation projects, with the aim of promoting business culture.<\/em>
Published 05\/14\/2026<\/h4>\n\n\n\n
Summary<\/strong><\/h3>\n\n\n\nThis article guides manufacturing companies through the implementation of computer vision for quality control, with the goal of moving closer to a zero-defect model. It begins by examining the current landscape\u2014where product variety is increasing but inspection often remains manual and based on sampling\u2014to explain how computer vision enables fully automated, more standardized, continuous, and data-driven inspection.<\/p>\n\n\n\n
The two main approaches (classical Machine Vision and Deep Learning) are analyzed, along with the value of a hybrid strategy, and the fundamental components for building a reliable industrial dataset: scene, lighting, labeling, and model lifecycle.<\/p>\n\n\n\n
The article then moves on to the reference standards (ISO 9001, EMVA 1288, OPC UA, ISO\/IEC 42001) and updated ISTAT and Eurostat 2025 data, which show still selective adoption and therefore a concrete competitive opportunity.<\/p>\n\n\n\n
The operational heart is a 6-step roadmap, from CTQ selection to industrialization and MLOps, accompanied by the 5 most common mistakes to avoid. The conclusion reiterates that Zero Defect is not an algorithm, but an integrated system of measurement, decision-making, and continuous improvement.<\/p>\n\n\n\n
ISTAT data, ISO standards, and a 6-step roadmap to transform visual inspection into a measurable competitive advantage<\/strong><\/em><\/h3>\n\n\n\nComputer Vision allows manufacturing companies to transition from sample and manual inspections to a model of Automated visual inspection on the 100%<\/strong>, measurably and repeatably reducing defects and scrap. In this guide, you will find updated data, reference standards, and a Concrete roadmap for industrial rigor in its implementation.<\/strong><\/p>\n\n\n\n\u201cZero defects\u201d quality is not just an ambitious goal; it is the result of a system that measure<\/strong>, decide<\/strong> e improve<\/strong> in a structured way. Technology really works when it is designed as Process system<\/strong>, supported by reliable data and anchored in AI quality, measurement, integration, and governance standards.<\/p>\n\n\n\n\ud83d\udca1 Do you want to understand where your company stands on the path to the Smart Factory?<\/strong> Discover the Benchmarking Study Operations 2025 by Bonfiglioli Consulting<\/a><\/em><\/p>\n\n\n\n<\/p>\n\n\n\n
Why computer vision is transforming quality control in factories<\/strong><\/h2>\n\n\n\nIn manufacturing, quality control often faces a paradox: while product mix and variability are increasing, inspection capabilities remain limited to manual sampling and checks. Machine vision resolves this imbalance by making the process more standardized\u2014same scene, same criteria, less operator subjectivity\u2014more continuous, with inspections up to 100% in-line where necessary, and more data-driven, because every image generates a traceable result and a corrective or preventive action.<\/p>\n\n\n\n
Robustness grows further when different modalities are combined: the 2D<\/strong> by contrast, texture, and color, the 3D<\/strong> for defects in shape and surface. An often underestimated element is the need to take into account Continuous maintenance<\/strong> of data and models: it's not a failure, but a physiological characteristic of any industrial AI system.<\/p>\n\n\n\n<\/p>\n\n\n\n
Machine Vision vs. Deep Learning: When to Use the Hybrid Approach<\/strong><\/h2>\n\n\n\nWhen discussing a vision for quality, it's useful to distinguish between two main approaches. The Classic Machine Vision<\/strong> relies on predefined rules and thresholds: it is deterministic, fast, and easy to audit, but requires stable conditions and well-defined defects. Computer Vision with Deep Learning<\/strong> It is, however, more suitable when finishes, textures, reflections, batches, and process conditions change, or when defects are numerous, rare, or varied.<\/p>\n\n\n\nIn practice, the most solid projects are often hybrids<\/strong>rules for simple constraints and deterministic checks, AI for classification and anomaly detection where variability makes fixed threshold logic fragile. The choice of approach is not ideological, but depends on the type of defect, the stability of the scene, and the required level of traceability.<\/p>\n\n\n\n<\/p>\n\n\n\n
How to build a reliable industrial dataset for computer vision<\/strong><\/h2>\n\n\n\nAlmost everything is decided here. An automated \u201cfactory\u201d visual inspection system requires treating the image as process data<\/strong>, not as a simple photo.<\/p>\n\n\n\nAcquisition and Scene: Optics, Lighting, Mechanics<\/strong><\/h3>\n\n\n\nEven before training, we need to make Repeatable scene<\/strong>: position, distance, exposure times, vibrations, reflections, physical shielding. Lighting is often the factor that makes or breaks the quality of the data\u2014and the most common mistake is to put off standardizing it until later.<\/p>\n\n\n\n2D vs 3D Inspection: How to Choose Based on Defect Type<\/strong><\/h3>\n\n\n\n![\"Informative](\"data:image\/svg+xml;base64,PHN2ZyB3aWR0aD0iMSIgaGVpZ2h0PSIxIiB4bWxucz0iaHR0cDovL3d3dy53My5vcmcvMjAwMC9zdmciPjwvc3ZnPg==\")
<\/figure>\n\n\n\n<\/p>\n\n\n\n
<\/p>\n\n\n\n
Ground truth and labeling: discipline, not bureaucracy<\/strong><\/h3>\n\n\n\nIf AI is used, labels must be consistent: defined defect taxonomy, clear rules for borderline cases, \u201cgolden set\u201d samples for regression testing. The model must be updated periodically to keep up with new cases and process variability: this is normal, it's part of the life cycle.<\/p>\n\n\n\n
Life cycle and skills<\/strong><\/h3>\n\n\n\nMachine vision systems are not \u201cset & forget.\u201d They require support for development and maintenance, and often, new internal roles and responsibilities. Devices (cameras, edge computing) are subject to rapid obsolescence: it is advisable to design from the outset with replaceability and standardization in mind.<\/p>\n\n\n\n
<\/p>\n\n\n\n
ISO 9001, EMVA 1288, and OPC UA: Standards for Industrial Vision<\/strong><\/h2>\n\n\n\nTo bring Computer Vision to industrial-grade quality with rigor, it is useful to rely on recognized standards at multiple levels.<\/p>\n\n\n\n
System quality standards<\/strong><\/h3>\n\n\n\nISO 9001<\/a><\/strong> it is the most widespread reference for setting up and improving a quality management system: processes, evidence, continuous improvement.<\/p>\n\n\n\nStandards for sensors and performance<\/strong><\/h3>\n\n\n\nEMVA 1288<\/a><\/strong> it is the standard for measuring and comparably presenting the specifications of sensors and cameras for machine vision: useful for technical selection, specifications, and system validation.<\/p>\n\n\n\nInteroperability Standard<\/strong><\/h3>\n\n\n\nGenICam<\/strong> <\/a>offers a plug & play interface for managing rooms and devices with a common interface, reducing complexity and vendor lock-in. OPC UA for Machine Vision<\/a><\/strong> allows instead for the integration of inspection systems with production control and IT systems, enabling vertical and horizontal integration of quality data along the entire supply chain.<\/p>\n\n\n\nAI Governance and Risk Management<\/strong><\/h3>\n\n\n\nISO\/IEC 42001<\/a><\/strong> provides the standard for establishing an AI management system with roles, controls, and continuous improvement. ISO\/IEC 23894<\/a><\/strong> offers guidance on managing specific AI-related risks, while the NIST AI RMF 1.0<\/a><\/strong> It is the international reference framework for managing the risks and reliability of AI systems throughout their entire lifecycle.<\/p>\n\n\n\nMetrology and Prevention<\/strong><\/h3>\n\n\n\nWhen vision enters measurement and decision logic, it's useful to reason about accuracy and repeatability as with any measurement system. On the prevention side, tools like FMEA\/FMECA<\/strong> they support prioritizing checks on CTQs (Critical To Quality).<\/p>\n\n\n\n<\/p>\n\n\n\n
How many companies use computer vision? ISTAT and Eurostat data 2025<\/strong><\/h2>\n\n\n\nThe current numbers send a clear message: in manufacturing, quality is already one of the most mature AI use cases, but the adoption of \u201call-enterprise\u201d based artificial vision is still selective. The gap between those who have already industrialized these solutions and those who are still in the PoC phase represents a concrete competitive opportunity.<\/p>\n\n\n\n
In global manufacturing, the Google Cloud report \u201cAI acceleration among manufacturers\u201d indicates that 39%<\/strong> companies are already using AI for quality inspection and 35%<\/strong> per product\/production line quality checks. At the European level, Eurostat<\/a> report that in 2025 the 19,95%<\/strong> At least 10% of EU companies with at least 10 employees use at least one AI technology, while technologies related to computer vision (image recognition\/processing) are adopted by between 3,78% and 7,22%<\/strong> of businesses. In Italy, according to ISTAT 2025<\/a>, the 16,4%<\/strong> of companies use at least one AI technology, but \u201cimage recognition\/processing\u201d technology is adopted by just 2,9%<\/strong> of businesses \u2014 a figure that highlights significant untapped potential.<\/p>\n\n\n\nThe Benchmarking Study Operations 2025 by Bonfiglioli Consulting<\/a><\/strong> invites you to read these rates in terms of overall maturity: a PoC is not enough; a Lean & Digital operating model supported by data, integration, and skills is needed to transform computer vision from an isolated initiative into a system capability.<\/p>\n\n\n\n<\/p>\n\n\n\n
Computer Vision Roadmap: From Proof of Concept to Industry Standard<\/strong><\/h2>\n\n\n\nAn effective roadmap follows this order and does not skip intermediate steps:<\/p>\n\n\n\n
\n- Select 1\u20132 CTQs<\/strong> where scraps, returns, or non-compliance are truly relevant<\/li>\n\n\n\n
- Standardize the scene<\/strong> \u2014 lighting, positioning, trigger, mechanical safeguards<\/li>\n\n\n\n
- Build datasets and labeling rules<\/strong> - defect taxonomy, documented edge cases, golden set for tests<\/li>\n\n\n\n
- Choose the approach<\/strong> (rules \/ Deep Learning \/ hybrid) and set clear metrics: False Acceptance Rate<\/em> vs. false rejection rate<\/em><\/li>\n\n\n\n
- Industrialize<\/strong> \u2014 integration with PLC, MES, and QMS, traceability of results, model version management<\/li>\n\n\n\n
- MLOps and continuous improvement<\/strong> \u2014 drift monitoring, controlled updates, AI governance<\/li>\n<\/ol>\n\n\n\n
<\/p>\n\n\n\n
\ud83d\udca1 Bonfiglioli Consulting supports manufacturing companies at every stage of this journey, from initial diagnosis to industrialization. Contact us for an estimate<\/a><\/em><\/p>\n\n\n\n<\/p>\n\n\n\n
The 5 most common errors in implementing computer vision<\/strong><\/h2>\n\n\n\n\n- \u201cWe'll sort out the lighting later.\u201d<\/strong> \u2014 it is the most frequent and most expensive mistake. Scene standardization comes first.<\/li>\n\n\n\n
- Dataset too \u201cclean\u201d<\/strong> \u2014 variable batches, reflections, dirt, wear, vibrations enter production. The dataset must reflect reality, not the ideal case.<\/li>\n\n\n\n
- Misaligned Quality Criteria<\/strong> \u2014 without a shared defect taxonomy, the model \u201clearns\u201d the opinion of whoever performed the labeling<\/li>\n\n\n\n
- No maintenance plan<\/strong> \u2014 Models and devices require periodic updates; obsolescence must be prevented from the initial design stage<\/li>\n\n\n\n
- Uncalibrated expectations<\/strong> Actual earnings may differ from those estimated if integration into the process is not managed methodically.<\/li>\n<\/ul>\n\n\n\n
<\/p>\n\n\n\n
Conclusion: Zero Defect is a system, not an algorithm<\/strong><\/h2>\n\n\n\nArtificial vision can become a pillar of Zero Defect manufacturing only when it is treated as a three-level integrated system: a measurement system<\/strong> with calibrated sensors, technical standards, and metrological validation; a decision-making system<\/strong> with clear rules, defined responsibilities, and integration with the QMS; a adaptive system<\/strong> with continuous monitoring, controlled updates, and structured AI governance.<\/p>\n\n\n\nTo move from \u201cquality control\u201d to \u201cquality by design,\u201d the starting point is not the model, but the Data\u2013standard\u2013process system<\/strong><\/p>\n\n\n\nFrequently Asked Questions<\/strong><\/h2>\n\n\n\n<\/p>\n\n\n\n
What is Computer Vision for quality control in factories?
<\/strong><\/h3>\n\n\n\nComputer Vision is a technology based on industrial cameras and artificial intelligence algorithms that allows for the automatic inspection of products during production, detecting defects, dimensional anomalies, or non-conformities in real-time. Unlike human visual inspection, it guarantees repeatability, speed, and traceability of results.<\/p>\n\n\n\n
Machine vision offers several advantages over manual quality control. It can inspect products faster, more consistently, and with higher accuracy. Machine vision systems can detect defects that are invisible to the human eye, such as microscopic imperfections or subtle color variations. They can also operate continuously without fatigue, reducing the risk of human error. Additionally, machine vision can collect and analyze data, providing valuable insights for process improvement and traceability.
<\/strong><\/h3>\n\n\n\nComputer Vision eliminates the variability of human judgment, reduces waste and complaints, cuts down on quality-related costs, and enables full traceability of every inspection. According to Google Cloud data, 39% of manufacturers already use it for quality inspection. Compared to manual inspection, it ensures 100% coverage of manufactured parts, 24 hours a day.<\/p>\n\n\n\n
How much does it cost to implement a computer vision system for quality control?<\/strong>
<\/h3>\n\n\n\nThe cost varies based on the complexity of the inspection scene, the number of cameras, the type of defects to detect, and the level of integration with PLCs, MES, and QMS. Simpler projects start from tens of thousands of euros; solutions based on Deep Learning and multi-station require more significant investments, but with ROI often less than 12\u201318 months due to reduced scrap and rework costs.<\/p>\n\n\n\n
What technical standards must be met to implement machine vision in production?
<\/strong><\/h3>\n\n\n\nThe main regulatory and technical references are: EMVA 1288<\/strong> for the characterization of industrial cameras, GenICam<\/strong> for the standard interface with the chambers, OPC UA for Machine Vision<\/strong> for IT\/OT integration, and quality system standards ISO 9001<\/strong>, ISO\/IEC 42001<\/strong> (AI management) e ISO\/IEC 23894<\/strong> (AI risk management). Adherence to these standards ensures metrological validation and system governance.<\/p>\n\n\n\nHow can Zero Defects be achieved with Computer Vision?
<\/strong><\/h3>\n\n\n\nZero Defect manufacturing is not achieved with a single algorithm, but by building an integrated three-level system: a measurement system<\/strong> (calibrated sensors, technical standards, validation), a decision-making system<\/strong> (clear rules, QMS integration, defined responsibilities) and a adaptive system<\/strong> (drift monitoring, controlled AI model updates, structured governance). The starting point is always data\u2013standards\u2013process system, not the technology itself.<\/p>\n\n\n\nSources and references<\/strong><\/h2>\n\n\n\n\n- Eurostat, Use of artificial intelligence in enterprises<\/em> \u2014 EU 2025: 19,951 TP3T AI adoption; 3,781 TP3T\u20137,221 TP3T for image recognition\/processing<\/a><\/li>\n\n\n\n
- ISTAT, Businesses and ICT \u2013 Year 2025<\/em> \u2014 Italy: 16.41 TP3T AI adoption; 2.91 TP3T image recognition\/processing<\/a><\/li>\n\n\n\n
- Google Cloud, AI acceleration among manufacturers<\/em> \u2014 39% quality inspection; 35% quality checks<\/li>\n\n\n\n
- ISO 9001 \u2014 Quality management system<\/a><\/li>\n\n\n\n
- ISO\/IEC 42001 \u2014 AI management systems<\/a><\/li>\n\n\n\n
- ISO\/IEC 23894 \u2014 Guidance on AI risk management<\/a><\/li>\n\n\n\n
- NIST AI RMF 1.0<\/a><\/li>\n\n\n\n
- EMVA 1288 - Specifications and measurements for machine vision cameras<\/a><\/li>\n\n\n\n
- GenICam \u2014 Generic Interface for Industrial Cameras<\/a><\/li>\n\n\n\n
- OPC UA for Machine Vision \u2014 Vision Integration with IT\/OT<\/a><\/li>\n\n\n\n
- Bonfiglioli Consulting, Benchmarking Study Operations 2025 \u2013 What's next in Operations?<\/em><\/a><\/li>\n<\/ul>\n\n\n\n
<\/p>","protected":false},"excerpt":{"rendered":null,"protected":false},"author":9,"featured_media":73935,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[117,115],"tags":[],"class_list":["post-73915","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-digital-transformation","category-operational-excellence"],"acf":[],"yoast_head":"\n
Come usare la Computer Vision per il Controllo Qualit\u00e0 in fabbrica<\/title>\n<meta name=\"description\" content=\"Scopri come la Computer Vision trasforma il controllo qualit\u00e0 verso lo Zero Defect: dati ISTAT, standard ISO e roadmap.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.bonfiglioliconsulting.com\/en\/computer-vision-quality-control-zero-defect\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Come usare la Computer Vision per il Controllo Qualit\u00e0 in fabbrica\" \/>\n<meta property=\"og:description\" content=\"Scopri come la Computer Vision trasforma il controllo qualit\u00e0 verso lo Zero Defect: dati ISTAT, standard ISO e roadmap.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.bonfiglioliconsulting.com\/en\/computer-vision-quality-control-zero-defect\/\" \/>\n<meta property=\"og:site_name\" content=\"Bonfiglioli Consulting\" \/>\n<meta property=\"article:published_time\" content=\"2026-05-14T07:30:34+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2026-05-14T20:32:38+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/www.bonfiglioliconsulting.com\/wp-content\/uploads\/2026\/05\/COMPUTER-VISION-1024x585.png\" \/>\n\t<meta property=\"og:image:width\" content=\"1024\" \/>\n\t<meta property=\"og:image:height\" content=\"585\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/png\" \/>\n<meta name=\"author\" content=\"Silvia Bolognini\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Silvia Bolognini\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"12 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/computer-vision-controllo-qualita-zero-defect\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/computer-vision-controllo-qualita-zero-defect\\\/\"},\"author\":{\"name\":\"Silvia Bolognini\",\"@id\":\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/it\\\/#\\\/schema\\\/person\\\/9914763a8be447279cf8103d52b6a6b9\"},\"headline\":\"Come implementare la Computer Vision per il Controllo Qualit\u00e0: guida pratica verso lo Zero Defect\",\"datePublished\":\"2026-05-14T07:30:34+00:00\",\"dateModified\":\"2026-05-14T20:32:38+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/computer-vision-controllo-qualita-zero-defect\\\/\"},\"wordCount\":2069,\"image\":{\"@id\":\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/computer-vision-controllo-qualita-zero-defect\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/wp-content\\\/uploads\\\/2026\\\/05\\\/COMPUTER-VISION-scaled.png\",\"articleSection\":[\"Digital Transformation\",\"Operational Excellence\"],\"inLanguage\":\"en-US\"},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/computer-vision-controllo-qualita-zero-defect\\\/\",\"url\":\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/computer-vision-controllo-qualita-zero-defect\\\/\",\"name\":\"Come usare la Computer Vision per il Controllo Qualit\u00e0 in fabbrica\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/it\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/computer-vision-controllo-qualita-zero-defect\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/computer-vision-controllo-qualita-zero-defect\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/wp-content\\\/uploads\\\/2026\\\/05\\\/COMPUTER-VISION-scaled.png\",\"datePublished\":\"2026-05-14T07:30:34+00:00\",\"dateModified\":\"2026-05-14T20:32:38+00:00\",\"author\":{\"@id\":\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/it\\\/#\\\/schema\\\/person\\\/9914763a8be447279cf8103d52b6a6b9\"},\"description\":\"Scopri come la Computer Vision trasforma il controllo qualit\u00e0 verso lo Zero Defect: dati ISTAT, standard ISO e roadmap.\",\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/computer-vision-controllo-qualita-zero-defect\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/computer-vision-controllo-qualita-zero-defect\\\/#primaryimage\",\"url\":\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/wp-content\\\/uploads\\\/2026\\\/05\\\/COMPUTER-VISION-scaled.png\",\"contentUrl\":\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/wp-content\\\/uploads\\\/2026\\\/05\\\/COMPUTER-VISION-scaled.png\",\"width\":2560,\"height\":1463,\"caption\":\"Una fabbrica futuristica con nastri trasportatori, illuminata da luci blu, presenta un'interfaccia grafica digitale che si sovrappone alla scena, mostrando l'automazione, l'analisi dei dati, la visione computerizzata del controllo qualit\u00e0 e la tecnologia avanzata nella produzione.\"},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/it\\\/#website\",\"url\":\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/it\\\/\",\"name\":\"Bonfiglioli Consulting\",\"description\":\"\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/it\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/it\\\/#\\\/schema\\\/person\\\/9914763a8be447279cf8103d52b6a6b9\",\"name\":\"Silvia Bolognini\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/88b4aff51854ff4456e63e579fae780f59b95a16d8f04b9cd8d992fb5e257357?s=96&d=mm&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/88b4aff51854ff4456e63e579fae780f59b95a16d8f04b9cd8d992fb5e257357?s=96&d=mm&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/88b4aff51854ff4456e63e579fae780f59b95a16d8f04b9cd8d992fb5e257357?s=96&d=mm&r=g\",\"caption\":\"Silvia Bolognini\"},\"url\":\"https:\\\/\\\/www.bonfiglioliconsulting.com\\\/en\\\/author\\\/sbolognini\\\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"How to Use Computer Vision for Quality Control in Factories","description":"Discover how Computer Vision is transforming Quality Control towards Zero Defects: ISTAT data, ISO standards, and a roadmap.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.bonfiglioliconsulting.com\/en\/computer-vision-quality-control-zero-defect\/","og_locale":"en_US","og_type":"article","og_title":"Come usare la Computer Vision per il Controllo Qualit\u00e0 in fabbrica","og_description":"Scopri come la Computer Vision trasforma il controllo qualit\u00e0 verso lo Zero Defect: dati ISTAT, standard ISO e roadmap.","og_url":"https:\/\/www.bonfiglioliconsulting.com\/en\/computer-vision-quality-control-zero-defect\/","og_site_name":"Bonfiglioli Consulting","article_published_time":"2026-05-14T07:30:34+00:00","article_modified_time":"2026-05-14T20:32:38+00:00","og_image":[{"width":1024,"height":585,"url":"https:\/\/www.bonfiglioliconsulting.com\/wp-content\/uploads\/2026\/05\/COMPUTER-VISION-1024x585.png","type":"image\/png"}],"author":"Silvia Bolognini","twitter_card":"summary_large_image","twitter_misc":{"Written by":"Silvia Bolognini","Est. reading time":"12 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/www.bonfiglioliconsulting.com\/computer-vision-controllo-qualita-zero-defect\/#article","isPartOf":{"@id":"https:\/\/www.bonfiglioliconsulting.com\/computer-vision-controllo-qualita-zero-defect\/"},"author":{"name":"Silvia Bolognini","@id":"https:\/\/www.bonfiglioliconsulting.com\/it\/#\/schema\/person\/9914763a8be447279cf8103d52b6a6b9"},"headline":"Come implementare la Computer Vision per il Controllo Qualit\u00e0: guida pratica verso lo Zero Defect","datePublished":"2026-05-14T07:30:34+00:00","dateModified":"2026-05-14T20:32:38+00:00","mainEntityOfPage":{"@id":"https:\/\/www.bonfiglioliconsulting.com\/computer-vision-controllo-qualita-zero-defect\/"},"wordCount":2069,"image":{"@id":"https:\/\/www.bonfiglioliconsulting.com\/computer-vision-controllo-qualita-zero-defect\/#primaryimage"},"thumbnailUrl":"https:\/\/www.bonfiglioliconsulting.com\/wp-content\/uploads\/2026\/05\/COMPUTER-VISION-scaled.png","articleSection":["Digital Transformation","Operational Excellence"],"inLanguage":"en-US"},{"@type":"WebPage","@id":"https:\/\/www.bonfiglioliconsulting.com\/computer-vision-controllo-qualita-zero-defect\/","url":"https:\/\/www.bonfiglioliconsulting.com\/computer-vision-controllo-qualita-zero-defect\/","name":"How to Use Computer Vision for Quality Control in Factories","isPartOf":{"@id":"https:\/\/www.bonfiglioliconsulting.com\/it\/#website"},"primaryImageOfPage":{"@id":"https:\/\/www.bonfiglioliconsulting.com\/computer-vision-controllo-qualita-zero-defect\/#primaryimage"},"image":{"@id":"https:\/\/www.bonfiglioliconsulting.com\/computer-vision-controllo-qualita-zero-defect\/#primaryimage"},"thumbnailUrl":"https:\/\/www.bonfiglioliconsulting.com\/wp-content\/uploads\/2026\/05\/COMPUTER-VISION-scaled.png","datePublished":"2026-05-14T07:30:34+00:00","dateModified":"2026-05-14T20:32:38+00:00","author":{"@id":"https:\/\/www.bonfiglioliconsulting.com\/it\/#\/schema\/person\/9914763a8be447279cf8103d52b6a6b9"},"description":"Discover how Computer Vision is transforming Quality Control towards Zero Defects: ISTAT data, ISO standards, and a roadmap.","inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.bonfiglioliconsulting.com\/computer-vision-controllo-qualita-zero-defect\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/www.bonfiglioliconsulting.com\/computer-vision-controllo-qualita-zero-defect\/#primaryimage","url":"https:\/\/www.bonfiglioliconsulting.com\/wp-content\/uploads\/2026\/05\/COMPUTER-VISION-scaled.png","contentUrl":"https:\/\/www.bonfiglioliconsulting.com\/wp-content\/uploads\/2026\/05\/COMPUTER-VISION-scaled.png","width":2560,"height":1463,"caption":"Una fabbrica futuristica con nastri trasportatori, illuminata da luci blu, presenta un'interfaccia grafica digitale che si sovrappone alla scena, mostrando l'automazione, l'analisi dei dati, la visione computerizzata del controllo qualit\u00e0 e la tecnologia avanzata nella produzione."},{"@type":"WebSite","@id":"https:\/\/www.bonfiglioliconsulting.com\/it\/#website","url":"https:\/\/www.bonfiglioliconsulting.com\/it\/","name":"Bonfiglioli Consulting","description":"","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.bonfiglioliconsulting.com\/it\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Person","@id":"https:\/\/www.bonfiglioliconsulting.com\/it\/#\/schema\/person\/9914763a8be447279cf8103d52b6a6b9","name":"Silvia Bolognini","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/secure.gravatar.com\/avatar\/88b4aff51854ff4456e63e579fae780f59b95a16d8f04b9cd8d992fb5e257357?s=96&d=mm&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/88b4aff51854ff4456e63e579fae780f59b95a16d8f04b9cd8d992fb5e257357?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/88b4aff51854ff4456e63e579fae780f59b95a16d8f04b9cd8d992fb5e257357?s=96&d=mm&r=g","caption":"Silvia Bolognini"},"url":"https:\/\/www.bonfiglioliconsulting.com\/en\/author\/sbolognini\/"}]}},"atf_populated":false,"_links":{"self":[{"href":"https:\/\/www.bonfiglioliconsulting.com\/en\/wp-json\/wp\/v2\/posts\/73915","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.bonfiglioliconsulting.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.bonfiglioliconsulting.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.bonfiglioliconsulting.com\/en\/wp-json\/wp\/v2\/users\/9"}],"replies":[{"embeddable":true,"href":"https:\/\/www.bonfiglioliconsulting.com\/en\/wp-json\/wp\/v2\/comments?post=73915"}],"version-history":[{"count":14,"href":"https:\/\/www.bonfiglioliconsulting.com\/en\/wp-json\/wp\/v2\/posts\/73915\/revisions"}],"predecessor-version":[{"id":73963,"href":"https:\/\/www.bonfiglioliconsulting.com\/en\/wp-json\/wp\/v2\/posts\/73915\/revisions\/73963"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.bonfiglioliconsulting.com\/en\/wp-json\/wp\/v2\/media\/73935"}],"wp:attachment":[{"href":"https:\/\/www.bonfiglioliconsulting.com\/en\/wp-json\/wp\/v2\/media?parent=73915"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.bonfiglioliconsulting.com\/en\/wp-json\/wp\/v2\/categories?post=73915"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.bonfiglioliconsulting.com\/en\/wp-json\/wp\/v2\/tags?post=73915"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
Computer Vision allows manufacturing companies to transition from sample and manual inspections to a model of Automated visual inspection on the 100%<\/strong>, measurably and repeatably reducing defects and scrap. In this guide, you will find updated data, reference standards, and a Concrete roadmap for industrial rigor in its implementation.<\/strong><\/p>\n\n\n\n \u201cZero defects\u201d quality is not just an ambitious goal; it is the result of a system that measure<\/strong>, decide<\/strong> e improve<\/strong> in a structured way. Technology really works when it is designed as Process system<\/strong>, supported by reliable data and anchored in AI quality, measurement, integration, and governance standards.<\/p>\n\n\n\n \ud83d\udca1 Do you want to understand where your company stands on the path to the Smart Factory?<\/strong> Discover the Benchmarking Study Operations 2025 by Bonfiglioli Consulting<\/a><\/em><\/p>\n\n\n\n <\/p>\n\n\n\n In manufacturing, quality control often faces a paradox: while product mix and variability are increasing, inspection capabilities remain limited to manual sampling and checks. Machine vision resolves this imbalance by making the process more standardized\u2014same scene, same criteria, less operator subjectivity\u2014more continuous, with inspections up to 100% in-line where necessary, and more data-driven, because every image generates a traceable result and a corrective or preventive action.<\/p>\n\n\n\n Robustness grows further when different modalities are combined: the 2D<\/strong> by contrast, texture, and color, the 3D<\/strong> for defects in shape and surface. An often underestimated element is the need to take into account Continuous maintenance<\/strong> of data and models: it's not a failure, but a physiological characteristic of any industrial AI system.<\/p>\n\n\n\n <\/p>\n\n\n\n When discussing a vision for quality, it's useful to distinguish between two main approaches. The Classic Machine Vision<\/strong> relies on predefined rules and thresholds: it is deterministic, fast, and easy to audit, but requires stable conditions and well-defined defects. Computer Vision with Deep Learning<\/strong> It is, however, more suitable when finishes, textures, reflections, batches, and process conditions change, or when defects are numerous, rare, or varied.<\/p>\n\n\n\n In practice, the most solid projects are often hybrids<\/strong>rules for simple constraints and deterministic checks, AI for classification and anomaly detection where variability makes fixed threshold logic fragile. The choice of approach is not ideological, but depends on the type of defect, the stability of the scene, and the required level of traceability.<\/p>\n\n\n\n <\/p>\n\n\n\n Almost everything is decided here. An automated \u201cfactory\u201d visual inspection system requires treating the image as process data<\/strong>, not as a simple photo.<\/p>\n\n\n\n Even before training, we need to make Repeatable scene<\/strong>: position, distance, exposure times, vibrations, reflections, physical shielding. Lighting is often the factor that makes or breaks the quality of the data\u2014and the most common mistake is to put off standardizing it until later.<\/p>\n\n\n\n <\/p>\n\n\n\n <\/p>\n\n\n\n If AI is used, labels must be consistent: defined defect taxonomy, clear rules for borderline cases, \u201cgolden set\u201d samples for regression testing. The model must be updated periodically to keep up with new cases and process variability: this is normal, it's part of the life cycle.<\/p>\n\n\n\n Machine vision systems are not \u201cset & forget.\u201d They require support for development and maintenance, and often, new internal roles and responsibilities. Devices (cameras, edge computing) are subject to rapid obsolescence: it is advisable to design from the outset with replaceability and standardization in mind.<\/p>\n\n\n\n <\/p>\n\n\n\n To bring Computer Vision to industrial-grade quality with rigor, it is useful to rely on recognized standards at multiple levels.<\/p>\n\n\n\n ISO 9001<\/a><\/strong> it is the most widespread reference for setting up and improving a quality management system: processes, evidence, continuous improvement.<\/p>\n\n\n\n EMVA 1288<\/a><\/strong> it is the standard for measuring and comparably presenting the specifications of sensors and cameras for machine vision: useful for technical selection, specifications, and system validation.<\/p>\n\n\n\n GenICam<\/strong> <\/a>offers a plug & play interface for managing rooms and devices with a common interface, reducing complexity and vendor lock-in. OPC UA for Machine Vision<\/a><\/strong> allows instead for the integration of inspection systems with production control and IT systems, enabling vertical and horizontal integration of quality data along the entire supply chain.<\/p>\n\n\n\n ISO\/IEC 42001<\/a><\/strong> provides the standard for establishing an AI management system with roles, controls, and continuous improvement. ISO\/IEC 23894<\/a><\/strong> offers guidance on managing specific AI-related risks, while the NIST AI RMF 1.0<\/a><\/strong> It is the international reference framework for managing the risks and reliability of AI systems throughout their entire lifecycle.<\/p>\n\n\n\n When vision enters measurement and decision logic, it's useful to reason about accuracy and repeatability as with any measurement system. On the prevention side, tools like FMEA\/FMECA<\/strong> they support prioritizing checks on CTQs (Critical To Quality).<\/p>\n\n\n\n <\/p>\n\n\n\n The current numbers send a clear message: in manufacturing, quality is already one of the most mature AI use cases, but the adoption of \u201call-enterprise\u201d based artificial vision is still selective. The gap between those who have already industrialized these solutions and those who are still in the PoC phase represents a concrete competitive opportunity.<\/p>\n\n\n\n In global manufacturing, the Google Cloud report \u201cAI acceleration among manufacturers\u201d indicates that 39%<\/strong> companies are already using AI for quality inspection and 35%<\/strong> per product\/production line quality checks. At the European level, Eurostat<\/a> report that in 2025 the 19,95%<\/strong> At least 10% of EU companies with at least 10 employees use at least one AI technology, while technologies related to computer vision (image recognition\/processing) are adopted by between 3,78% and 7,22%<\/strong> of businesses. In Italy, according to ISTAT 2025<\/a>, the 16,4%<\/strong> of companies use at least one AI technology, but \u201cimage recognition\/processing\u201d technology is adopted by just 2,9%<\/strong> of businesses \u2014 a figure that highlights significant untapped potential.<\/p>\n\n\n\n The Benchmarking Study Operations 2025 by Bonfiglioli Consulting<\/a><\/strong> invites you to read these rates in terms of overall maturity: a PoC is not enough; a Lean & Digital operating model supported by data, integration, and skills is needed to transform computer vision from an isolated initiative into a system capability.<\/p>\n\n\n\n <\/p>\n\n\n\n An effective roadmap follows this order and does not skip intermediate steps:<\/p>\n\n\n\n <\/p>\n\n\n\n \ud83d\udca1 Bonfiglioli Consulting supports manufacturing companies at every stage of this journey, from initial diagnosis to industrialization. Contact us for an estimate<\/a><\/em><\/p>\n\n\n\n <\/p>\n\n\n\n <\/p>\n\n\n\n Artificial vision can become a pillar of Zero Defect manufacturing only when it is treated as a three-level integrated system: a measurement system<\/strong> with calibrated sensors, technical standards, and metrological validation; a decision-making system<\/strong> with clear rules, defined responsibilities, and integration with the QMS; a adaptive system<\/strong> with continuous monitoring, controlled updates, and structured AI governance.<\/p>\n\n\n\n To move from \u201cquality control\u201d to \u201cquality by design,\u201d the starting point is not the model, but the Data\u2013standard\u2013process system<\/strong><\/p>\n\n\n\n <\/p>\n\n\n\n Computer Vision is a technology based on industrial cameras and artificial intelligence algorithms that allows for the automatic inspection of products during production, detecting defects, dimensional anomalies, or non-conformities in real-time. Unlike human visual inspection, it guarantees repeatability, speed, and traceability of results.<\/p>\n\n\n\n Computer Vision eliminates the variability of human judgment, reduces waste and complaints, cuts down on quality-related costs, and enables full traceability of every inspection. According to Google Cloud data, 39% of manufacturers already use it for quality inspection. Compared to manual inspection, it ensures 100% coverage of manufactured parts, 24 hours a day.<\/p>\n\n\n\n The cost varies based on the complexity of the inspection scene, the number of cameras, the type of defects to detect, and the level of integration with PLCs, MES, and QMS. Simpler projects start from tens of thousands of euros; solutions based on Deep Learning and multi-station require more significant investments, but with ROI often less than 12\u201318 months due to reduced scrap and rework costs.<\/p>\n\n\n\n The main regulatory and technical references are: EMVA 1288<\/strong> for the characterization of industrial cameras, GenICam<\/strong> for the standard interface with the chambers, OPC UA for Machine Vision<\/strong> for IT\/OT integration, and quality system standards ISO 9001<\/strong>, ISO\/IEC 42001<\/strong> (AI management) e ISO\/IEC 23894<\/strong> (AI risk management). Adherence to these standards ensures metrological validation and system governance.<\/p>\n\n\n\n Zero Defect manufacturing is not achieved with a single algorithm, but by building an integrated three-level system: a measurement system<\/strong> (calibrated sensors, technical standards, validation), a decision-making system<\/strong> (clear rules, QMS integration, defined responsibilities) and a adaptive system<\/strong> (drift monitoring, controlled AI model updates, structured governance). The starting point is always data\u2013standards\u2013process system, not the technology itself.<\/p>\n\n\n\n <\/p>","protected":false},"excerpt":{"rendered":null,"protected":false},"author":9,"featured_media":73935,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[117,115],"tags":[],"class_list":["post-73915","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-digital-transformation","category-operational-excellence"],"acf":[],"yoast_head":"\nWhy computer vision is transforming quality control in factories<\/strong><\/h2>\n\n\n\n
Machine Vision vs. Deep Learning: When to Use the Hybrid Approach<\/strong><\/h2>\n\n\n\n
How to build a reliable industrial dataset for computer vision<\/strong><\/h2>\n\n\n\n
Acquisition and Scene: Optics, Lighting, Mechanics<\/strong><\/h3>\n\n\n\n
2D vs 3D Inspection: How to Choose Based on Defect Type<\/strong><\/h3>\n\n\n\n
![\"Informative](\"https:\/\/www.bonfiglioliconsulting.com\/wp-content\/uploads\/2026\/05\/2D-e-3D-scegliere-in-funzione-del-difetto-1024x451.png\")
<\/figure>\n\n\n\nGround truth and labeling: discipline, not bureaucracy<\/strong><\/h3>\n\n\n\n
Life cycle and skills<\/strong><\/h3>\n\n\n\n
ISO 9001, EMVA 1288, and OPC UA: Standards for Industrial Vision<\/strong><\/h2>\n\n\n\n
System quality standards<\/strong><\/h3>\n\n\n\n
Standards for sensors and performance<\/strong><\/h3>\n\n\n\n
Interoperability Standard<\/strong><\/h3>\n\n\n\n
AI Governance and Risk Management<\/strong><\/h3>\n\n\n\n
Metrology and Prevention<\/strong><\/h3>\n\n\n\n
How many companies use computer vision? ISTAT and Eurostat data 2025<\/strong><\/h2>\n\n\n\n
Computer Vision Roadmap: From Proof of Concept to Industry Standard<\/strong><\/h2>\n\n\n\n
\n
The 5 most common errors in implementing computer vision<\/strong><\/h2>\n\n\n\n
\n
Conclusion: Zero Defect is a system, not an algorithm<\/strong><\/h2>\n\n\n\n
Frequently Asked Questions<\/strong><\/h2>\n\n\n\n
What is Computer Vision for quality control in factories?
<\/strong><\/h3>\n\n\n\nMachine vision offers several advantages over manual quality control. It can inspect products faster, more consistently, and with higher accuracy. Machine vision systems can detect defects that are invisible to the human eye, such as microscopic imperfections or subtle color variations. They can also operate continuously without fatigue, reducing the risk of human error. Additionally, machine vision can collect and analyze data, providing valuable insights for process improvement and traceability.
<\/strong><\/h3>\n\n\n\nHow much does it cost to implement a computer vision system for quality control?<\/strong>
<\/h3>\n\n\n\nWhat technical standards must be met to implement machine vision in production?
<\/strong><\/h3>\n\n\n\nHow can Zero Defects be achieved with Computer Vision?
<\/strong><\/h3>\n\n\n\nSources and references<\/strong><\/h2>\n\n\n\n
\n