The Question Plant Heads Ask, and the One They Should Ask
"Can we add vision inspection without replacing the line?" is the question that starts almost every retrofit conversation. The honest answer is almost always yes — a vision inspection station is a self-contained unit that sits at a point on the existing conveyor, not a replacement for the filler, capper, or labeller.
The question that actually determines project cost and timeline is different: where does the defective unit go once the camera has found it? That single question — the reject mechanism — is where most retrofit projects spend the majority of their engineering effort, and where most vendor quotes understate the real scope of work.
What "Retrofit" Actually Involves
A retrofit vision inspection system has four functional pieces, and only one of them is the part most people picture when they hear "machine vision":
- The inspection station — camera, lens, illumination, and enclosure mounted over or beside the existing conveyor at a chosen inspection point
- The controller and software — processes the image, makes the pass/fail decision, and logs the result
- The signal handshake — communicates the inspection trigger and result to the existing line PLC, typically via digital I/O or a fieldbus protocol the line already uses
- The reject mechanism — physically removes the defective unit from the line without disturbing adjacent product, in the time available before the next unit arrives
Vendors selling primarily on camera specification tend to under-scope the third and fourth items — because they are mechanical and controls engineering, not vision engineering, and they vary enormously from one existing line to the next.
The Real Constraints — Not the Ones Most Plant Heads Expect
| Constraint | Why It Matters | Typical Impact |
|---|---|---|
| Available physical space | Camera, lighting, and enclosure need a clear mounting point without blocking operator access or existing maintenance points | Mechanical design |
| Line speed and indexing | Continuous-motion lines need trigger sensors and strobe illumination; indexed lines are generally easier to inspect but slower overall | Inspection timing |
| Reject point availability | Most existing lines were not built with a reject mechanism at all — this usually has to be engineered from scratch | Highest cost driver |
| PLC protocol and spare I/O | The vision controller must talk to the existing PLC using a protocol and spare I/O points it actually supports | Controls integration |
| Ambient light and vibration | Existing lines were not designed around consistent lighting for a camera — stray light and conveyor vibration both affect image quality | Illumination & mounting design |
| SKU variation on the same line | Multiple container shapes, label positions, or product colours running on one line require recipe management, not just camera hardware | Software configuration |
The most expensive line item on most retrofit quotes: a properly engineered reject mechanism — air blast nozzle, pneumatic pusher, or diverter gate — sized correctly for the container weight, line speed, and available reaction time. Lines with no existing reject point often need a section of conveyor modified to accommodate one.
The Retrofit Process, Step by Step
-
Line survey and speed audit. Measure actual line speed, container spacing, available inspection points, and confirm physical clearance for a camera and lighting enclosure without disturbing operator access.
-
Identify the inspection point and the reject point. These are rarely the same location — the reject point needs enough downstream distance from the camera to allow processing and actuation time before the unit passes it.
-
Mechanical mounting design. A bracket or frame engineered to the specific conveyor geometry, designed not to interfere with cleaning, maintenance access, or changeover procedures.
-
Illumination selection for the existing environment. Lighting has to overpower ambient light from the factory floor and any reflective surfaces nearby — this is evaluated on-site, not from a datasheet.
-
PLC and controls integration. Confirm the existing PLC's available I/O and communication protocol, then wire the trigger signal (part-present sensor) and the reject output back into line control logic.
-
Reject mechanism installation and timing validation. Install and tune the reject actuator so it fires within the reaction window for the fastest SKU the line runs — the tightest timing case, not the average one.
-
Trial run across every SKU the line produces. Run production samples for each container type, label, and colour variant the line handles, confirming detection accuracy and reject timing hold across all of them — not just the SKU used for the demo.
-
Validation and operator handover. For regulated manufacturers, this includes IQ/OQ/PQ documentation. For all installations, it includes operator training on recipe selection, fault indication, and the manual override procedure.
Optomech's applications engineers conduct an on-site line survey before quoting — so the reject mechanism, PLC integration, and mounting design are scoped accurately from the start, not discovered mid-installation.
What Most People Get Wrong
The most common misconception is treating a vision retrofit as primarily a software or camera-selection decision. In practice, the camera and software are usually the least variable part of the project — a well-established inspection algorithm for cap presence, label placement, or fill level does not change much from one installation to the next. What changes, every time, is the mechanical and controls integration around it.
The second common mistake is underestimating the reject mechanism as an afterthought — "we'll just add an air jet." A reject mechanism that fires too early, too late, or with too much force can damage adjacent good product or fail to remove the defective unit cleanly. Sizing and timing this correctly requires the same line speed and product weight data used to design the inspection point itself.
The third mistake is assuming a single inspection recipe will handle every SKU the line runs. A vision system tuned for one bottle shape and label position will either miss defects or generate false rejects the moment the line changes over to a different SKU without a corresponding recipe change. Recipe management — and a clear changeover procedure for selecting the right one — has to be part of the original scope, not an add-on request after commissioning.
Retrofit vs. New Line — What Actually Differs
A new line is designed around the inspection station from the start. A retrofit has to fit the inspection station, lighting, and reject mechanism into a layout that was never planned around them.
With mechanical design and pre-fabrication completed offline, physical installation of the inspection station and reject mechanism is typically a planned shutdown window of hours to one shift — not a multi-day line rebuild.
Whether retrofit or new-build, a pharma or FMCG line still needs the same trial runs across every SKU and the same IQ/OQ/PQ documentation before go-live. Retrofit does not shortcut validation.
Practical Takeaway
Before requesting a retrofit quote, walk your line with these questions in hand: where would the inspection point sit, where would the reject point sit, and does a reject mechanism already exist anywhere on the line that could be repurposed?
Answering these upfront — even roughly — gives any vendor a far more accurate basis for a quote, and avoids the most common source of retrofit project overruns: discovering mid-installation that the reject point needs a section of conveyor modified to accommodate it.
Optomech's vision engineers scope every retrofit project with an on-site line survey before quoting — covering mechanical mounting, reject mechanism design, and PLC integration for the specific line in question. Our installed systems run across pharma, FMCG, and nutraceutical packaging lines in India and export markets, most of them added to lines that were already in production.