By Gabriele Mangiafico, Business Development Manager, EMEA, Axis Communications
The regulatory framework net is tightening, customers care about traceability more than ever, and timelines are constantly being squeezed. The role of quality control has moved beyond catching defects at final test: it has, rightly, become part of every step of production. It is the necessary proof that each process has met specification, been assembled correctly, and that the product can be traced back to its originating materials.
Test data is vital, but it does not suffice alone. There is a context gap in testing, and if manufacturers are to keep pace with growing demand, it must be filled. Compliance records do not supply evidence that production actually occurred as designed. Network video, integrated into the production environment, changes what can be documented from the ‘what’ to the ‘why’, complementing and adding process context to existing automated inspection systems.
Building greater process visibility
In EMS environments, automated inspection stations are vital, as are the process steps surrounding them. Feeder loading, manual assembly, hand soldering, et al; these are the moments where variation, handling error, or deviation from work instruction can create downstream failures.
Cameras positioned at assembly stations, solder points, placement equipment and test fixtures provide a visual record of the path of a board’s manufacture. Synchronised with time stamps from your MES or inspection and test systems, this video then becomes part of the quality record.
Field failures are not uncommon. A board could pass final testing, but still a failure might appear in service. Investigating such failures – finding the ‘why’ – requires a clear understanding of how that board was assembled. Test logs could show the board passed electrical and thermal tests at production. But a video record from assembly supports the process of verifying correct process, whether the board has reached the right station, the operator has followed the right work instruction, and so on. Combined with test and inspection data, cameras provide visual context that dramatically shortens root cause analysis.
Video allows forensic review of the exact procedures used compared to their work instruction. This level of traceability is increasingly demanded by OEMs in high-reliability segments, helping strengthen audit readiness and support investigations. It also supports faster problem resolution, replacing multiple test cycles or an expensive trip to the line by an engineer with a detailed remote review of the assembly sequence.
Integration with production systems
Effective visual process monitoring does not mean dropping in a camera and calling it done. Integration with your MES, QMS or test equipment is what makes its forensic applications work; the video must be searchable by board serial number, work order, date range or test result. A technician investigating a failure should be able to query the MES, locate the board, and immediately access the corresponding video from assembly and test.
Integration also enables closed-loop quality assurance. When a board fails final electrical test, the system can flag the corresponding assembly station video for immediate review. This kind of on-line fault finding is the route to more efficient and effective process management, but without integration these kinds of investigations are usually manual and time-consuming.
Setting up video integration requires planning. Your camera system must be able to work with standard industrial networking protocols like Modbus TCP, OPC UA, open standard APIs and event-driven integration. It cannot be a closed-camera island. The MES must be able to tie together video footage and testing data based on board identifiers. Time synchronisation across systems must be tightly controlled so that video, MES events and test records align tightly.
While possibly challenging, these are engineering problems with established solutions, and camera systems running on open platforms enable the creation of the kind of custom integrations that make this work.
Process verification without manual inspection
The camera reduces, but does not replace, the need for manual visual inspection. The human eye is and remains important in many scenarios. But the inspection process itself can be documented via video. Combine an inspector’s assessment alongside related video, and this creates an auditable inspection record. Later, if a defect escapes and reaches the field, you can review both the video and the original inspection decision.
Video also supports training and consistency. By using recordings of good assemblies and failed assemblies, inspectors and assemblers can train against real-world examples rather than abstractions. A new technician can see how an experienced assembler performs a particular procedure. A quality engineer can use video examples to conduct training on defect identification. Video ties the process together.
Data retention and compliance
Not everything can be retained, since video creates large data volumes. Retention policy should be driven by customer, regulatory and product-liability requirements. In regulated sectors, retention periods can be long and must be defined carefully. The key is not to over-retain by default, but to retain the right evidence, at the right quality, for the right duration.
Storage requirements are manageable with modern on-camera compression, using high quality, low size codecs like AV1 where possible, and with tiered storage. Cloud-based video management systems can scale, and while doing so may be costly, any spend is likely justified if video prevents a field failure or allows rapid problem resolution.
Regulatory considerations also include access control and privacy. Video of the production floor must be secured. Access to video records for a particular board should be restricted to relevant personnel. Where video captures workers, privacy and labour-law considerations must be addressed from the outset. In Europe especially, deployments should be purpose-limited, proportionate, transparent to employees, and governed by clear access and retention policies. Cybersecurity therefore needs to be a primary concern, in software and on connected camera hardware.
Getting started
Implementation should start small, beginning with the process step that creates the most field failures or customer complaints. Add video at one assembly station or one test fixture. Integrate it with your MES. Document a few boards end-to-end and verify the process works. Once the initial infrastructure is in place, the case for expansion is proven.
Not that there is much of an argument: for EMS providers managing complex supply chains and OEM requirements, video-enabled traceability is a vitally important and practical way to strengthen root-cause analysis. The organizations that benefit most will be the ones that implement it selectively, integrate it properly, and govern it as part of a wider quality architecture rather than as a standalone camera project.
About the Author:
Gabriele Mangiafico is part of Axis’ EMEA Business Development team, driving growth through digital transformation and innovative solutions in the industrial and manufacturing sectors. With a background that includes technical consultancy in NLP and computer vision, as well as business development roles for leading industrial companies, he has led complex projects and strategic roadmaps across the industry and is passionate about shaping the future of smart manufacturing.
Learn more about how Axis solutions enable situational awareness and actionable insights:
https://www.axis.com/en-gb/solutions/industrial
https://www.axis.com/customer-story/axis-industrial-vehicle-production
