What needed solving
Six parallel positions needed a consistent presence check. The initial rim-based definition also rejected too many cones, so the customer and Qualitas revised the damage criteria to focus on loose pieces inside the cone.
How Qualitas solved it
One camera and controlled lighting serve each position. EagleEye Edge applies the selected recipe, sends results to the PLC and displays inspection decisions. The final damage scope covers internal cone fragments greater than 10 × 10 mm; rim-only damage and cracks are not rejection conditions under that scope.
Watch the cone-inspection application
Six cone positions checked together
One camera looks into each cone position. Triggered lighting gives the software a consistent view of the wafer and sleeve, while the six positions are checked in the same inspection cycle. The camera arrangement follows the parallel lanes of the line.
Defining a useful reject condition
The initial rim-based check rejected too many cones. The customer-approved production scope focused instead on loose cone pieces visible inside the cone, with a specified minimum size greater than 10 × 10 mm.
Rim damage without loose pieces inside the cone remained acceptable under the revised scope. Crack detection and smaller fragments were excluded. This boundary matters: the inspection must apply the agreed production criteria, rather than reject every visible imperfection.
From line trigger to operator decision
- The line trigger starts image capture for the six camera positions and their lights.
- Images pass to the computer running EagleEye Edge, where the inspection checks are applied.
- The software sends the inspection result to the PLC and displays pass/fail information. A reject indicator identifies the affected position.
- Inspection images and accepted/rejected counts can be retained for review.
Operators select the recipe for the product and collar combination before inspection. The operating procedure lists 14 recipes covering 50 ml and 100 ml products with different cone colours and collar types.
Reviewable decisions and reports
Saved images let the team review the examples behind inspection decisions. The project’s grading workflow supports manual review and labelling during model development.
The inspection report can be filtered by date range, product variant and shift, then exported as a CSV file. Operators can review accepted and rejected counts alongside the image history.
Delivered capability and go-live
| Delivered capability | What it means on the line |
|---|---|
| 6 camera positions | A dedicated image and reject indication for each of the six parallel positions. |
| 14 recipe selections | Configured product and collar combinations listed in the operating procedure. |
| Pass/fail and saved images | Inspection decisions can be reviewed alongside the captured images. |
| Live from March 2024 | The system went live in March 2024, with software inputs, outputs and reports validated. |
The delivered system brings presence and damage checks into one line-connected inspection workflow. It identifies incomplete or damaged cones and retains information that operators can use to investigate repeat failures.
Planning cone inspection on your line
Share lane count, cone and sleeve variants, collar types, the exact conditions to reject, and the line-trigger and reject interfaces. We can assess camera coverage, lighting and a validation set that distinguishes unacceptable damage from acceptable variation.
Share lane count, cone and sleeve variants, collar types, acceptable and unacceptable examples, and the line-trigger and reject interfaces.
Request project details
