How Does a Vision Inspection Machine Catch Defects in Filled Medicine Bottles
You can picture a run of 10 ml glass vials that moves down a filling line at a steady pace. Most vials look the same. But one vial sits 2 mm lower than it should. Another vial has a tiny chip on the rim. That chip is barely visible under the plant lights. An operator who scans bottles all shift will likely miss both problems. A vision inspection machine does not miss them. It looks at every bottle, every time. It makes a decision in a fraction of a second. The steps below show how it does that.
Bottles Are Positioned Before the Camera Ever Sees Them
A camera can only measure what it sees consistently. Star wheels or conveyor guides hold each container steady as it passes the inspection station, so the camera sees every bottle from the same distance and angle. Even a small wobble can make a correct fill look low or turn a reflection into something that looks like a crack.
Consistent distance reduces false rejects
Matching angles make results repeatable from bottle to bottle
When the machine knows exactly where each bottle sits, every later step works from the same starting point.
Lighting and Cameras Capture Several Views
Once the bottle is steady, high-resolution industrial cameras take images under carefully tuned lighting. Fill level is usually measured with backlighting, since light passing through the glass and liquid creates a sharp boundary at the surface. Chips and scratches show up better under angled or ring lighting, which highlights edges and shadows.
Most systems capture several views per bottle, covering the body, shoulder, neck, and base. A defect hidden from one angle can still be caught from another. Lighting and camera settings are locked during setup, so every bottle on every shift is judged under the same conditions.
Software Measures Fill Level Against Your Target Range
After the system captures the images, the software turns them into measurements. For fill level, it finds the liquid line in the image. Then it converts its position into a height. The software compares that height against the target range you entered for that bottle. The exact tolerance is set according to your product and dose needs.
Software Looks for Cracks, Chips and Seal Defects
Fill level is only half the job. The software also checks the glass itself for problems that could make the bottle unsafe or not sterile. It scans the body for hairline cracks. It checks the neck and mouth for chips. For ampoules, it checks the sealed tip for tiny fractures. An operator at normal speed may not see these fractures.
Each defect type has its own detection rule. The system compares the image against a reference of a good bottle. It marks any area that differs beyond the set threshold. Resolution reaches ≥40 μm for visible foreign particles and light inspection; on the ampoule combi machine the leak-testing zone reaches ≥0.2–0.3 μm. The rules are written into the recipe. So the machine applies them the same way to every container. The machine has no fatigue and no drift between shifts.
Every Bottle Gets a Pass or Fail Decision
After the system analyzes all the images, the machine makes one decision per bottle: pass or reject. A good bottle moves on to capping and labeling. A flagged bottle is pushed off the conveyor by a small air jet or mechanical gate. The rest of the line keeps running without interruption.
Different Containers Need Different Inspection Settings
Vials, ampoules, and oral liquid bottles do not look alike. So they do not get inspected the same way. Vials are usually checked for stopper seating and fill height through the glass. Ampoules need close attention at the sealed tip. A hairline crack there can be easy to miss. Oral liquid bottles often come in plastic or glass and in a wider range of shapes. So the system needs flexible recipes that can be adjusted for each product.
This is also where equipment upstream matters. If your line includes a vial filling machine, an aseptic filling machine, or an oral liquid filling machine, the inspection station usually sits after filling and capping. Catching a short-dosed or damaged bottle at that point keeps it out of labeling and packing. Fixing the problem later gets much more expensive.
Rejected Units Tell You Where the Process Is Drifting
A reject count on its own does not say much. The real value comes from patterns. If a filling nozzle starts drifting, the fill-level rejects will slowly climb. If a feeding system is scratching bottles, chip-related rejects will appear more often. Good systems store images, timestamps, batch numbers, and reject reasons. So the quality team can see these trends early.
Questions to Ask Before You Choose a System
Not every vision inspection machine suits every line. You should start with your container types. Vials, ampoules, and oral liquid bottles each need different camera setups and reject rules. Then you should look at your required speed, the defect types you most need to catch, and how much room you have on the line.
You should ask suppliers how the system handles a change in bottle size. You should ask how long a recipe switch takes. You should ask whether the software can export inspection data in a format your quality team already uses. Finally, you should request a test run with your own bottles and your real defect samples. Seeing the machine catch the problems you actually face is the most reliable way to judge it.
A vision inspection machine works by doing three things consistently. It holds each bottle still. It looks at the bottle from several angles. It compares what it sees against a clear set of rules. That combination catches underfilled vials, cracked ampoules, and damaged glass before they reach patients. If you want to see how the system would handle your containers, you should visit the vision inspection machine page and ask for a sample test.