Vision-guided robotic cells · Life Science

No two parts are alike.
Each one is inspected.

If your components vary in position and shape, our robotic cell recognizes each part individually, processes it precisely in the right place, and verifies the result seamlessly — in one pass.

1 µm
Hardware accuracy
100 %
Checked before and after
OPC UA
AGV/FTS-ready
IVD · DD · SU
Drug Delivery / Single-use
The challenge

Three challenges that only a cohesive architecture can solve.

Varied components, varying features, and 100% verification requirements—many vendors address two of these separately. Only a closed-loop architecture consisting of before-vision, processing, and after-vision can handle all three in a single pass.

If that applies to you
  • Their components are arranged differently in the tray each time.
  • The area to be worked on is never in exactly the same position.
  • Each individual part must be verifiably tested.
  • Production volumes are increasing — manual inspection can no longer keep up.
  • The audit requires complete documentation for each part.
Challenge · 01

The situation varies

Gripper transfers, injection molding tolerances, and sensitive materials prevent hard fixturing. Rigid machining stations fail due to this variation.

Our approach:
Image processing captures each part individually and transmits the exact position to the robot.

Challenge · 02

The trait migrates

Due to injection molding and bonding tolerances, islands, septa, ports, and membranes are not reproducibly positioned in the same relative location. Robotics based solely on a program will fail.

Our approach:
We relocate the feature on each part and correct the robot path in real time.

Challenge · 03

The sample size is insufficient

Regulatory requirements and risk profiles demand that each processing step be independently verified and documented with an OK/NOK result. End-of-line inspection is too late—the rejects have already been generated.

Our
After Vision approach compares against tolerance in the same cell, automatically sorts IO/NIO and passes the result per piece to the MES.

Series production at QIAGEN

In series production at QIAGEN.

For QIAGEN, we operate test cells in the ongoing series production of diagnostic components: feeding, recognizing, processing, testing, sorting — fully automatically and with documented proof for each individual part.

15 µm
Camera resolution
up to 2 × 96 positions per plate
Components/Layer
KLT manual + AGV via lifting door
Supply
Second SCARA can be retrofitted
Scalable
What the cell can do — and what it cannot

Does this fit your role?

What our cells are built for — and what they are not. This clarity belongs at the beginning for us, not in the ongoing project.

Within the scope

  • Vision-guided handling of scattered components
  • Real-time feature localization and path correction
  • Machining steps with SCARA and 6-axis robots
  • 100% inspection before and after the step
  • Sorted by test result
  • Traceable documentation for each component
  • Connection to line, MES, AGV/FTS via OPC UA

Not in scope

  • Mass production at high speeds beyond robot cycle times
  • Heavy handling beyond the SCARA load capacity
  • Purely manual workplaces without automation
  • Chemical or wet chemical process steps
  • Pure software/AI projects without equipment
  • Standard pick-and-place without variation or inspection requirements
Application areas

Where our cells are in use today.

Proven · in series

Microfluidics, lab-on-a-chip and IVD consumables.

For components with inlets, outlets, air equalization openings, reagent ports or other defined functional points.

Implemented
  • Opening or perforating ports
  • Processing of reagent or fluidic sites
  • Camera-guided position detection
  • Verification against component reference (offset OK/NOK)
  • IO/NIO sorting

Directly related to our QIAGEN reference: Injection-molded consumables with microstructures — from quantities where manual handling and end-of-line inspection reach their limits.

component
microfluidic chip
function
Open reagent port
Camera resolution
15 µm
Test
Before, afterwards
Possible application

Pens, autoinjectors and cartridge components.

For scattered injection-molded parts that must be individually identified in their position, processed or joined, and fully documented.

Applicable to
  • Position detection of each component before gripping
  • Editing or adding defined function points
  • Verification of safety-relevant dimensions
  • Verification against component reference
  • Traceable documentation for each component

The same combination of image processing and SCARA processing as in microfluidics — applied to drug delivery components with high traceability requirements.

component
Cartridge / Pen part
function
Joining & Testing
Camera resolution
after component
Test
Before, afterwards
Possible application

Disposable components for bioprocess engineering.

For connectors, bag ports and hose sets whose functional dimensions must be verified and traceable before installation.

Applicable to
  • Testing of functional and sealing dimensions
  • Dispersion-tolerant handling of sensitive parts
  • Camera-guided position detection
  • Verification against component reference
  • Sorted by findings

Single-use components come in a wide variety of variants — the vision-guided cell adapts to the respective geometry instead of relying on a fixed fixture.

component
Connector / Port
function
Measurement verification
Camera resolution
after component
Test
Before, afterwards
MV.BRIX — modular platform

Configure instead of developing from scratch

Our MV.BRIX platform consists of proven hardware and software modules — automation, vision, software, inspection. We configure your customized system from these building blocks. Reduced risk. Shorter delivery times. Lower total cost of ownership.

  • Scalable · maintainable · future-proof
  • 100% perfectly tailored to your requirements
  • Time to first measurement 4-8 months
Why MABRI.VISION?

Six questions for a quick assessment.

A simple "yes" is enough to initiate an initial technical conversation. Select the option that applies to you.

hit 0 / 6

The more points that apply, the more worthwhile a feasibility discussion becomes. However, a single applicable statement is sufficient grounds for such a discussion.

Discuss component process
  • Does a defined area on the component need to be machined with micrometer precision?
  • Is the component or location not always in the same place?
  • Does the step need to be checked 100% to ensure it was correct?
  • Are the components supplied from KLT, tray or via AGV?
  • Is the application in diagnostics, drug delivery, or bioprocessing?
  • Are you planning a new line or the scaling of an existing one?
What happens inside the cell

Six steps. One cell.

From delivery to documentation, the part remains in a single cell — without changing stations, without being moved again.

Step 1

Supply

Components enter the cell as bulk material, in trays or via AGV/FTS — without complex singulation hardware.

Step 2

Location recognition

The image processing determines the position and orientation of each component individually — the gripping position is calculated for each part.

Step 3

editing

SCARA or 6-axis robots perform the work step — the path is corrected to the actual feature position.

Step 4

Test

The result is checked 100% before and after the step — dimensions, completeness and position.

Step 5

Sorting

Good and bad parts are separated and removed based on the test result — without manual sorting.

Step 6

documentation

Each step is documented and traceable for each component and transferred to your line or MES.

REFERENCES

AR / Smart Glasses
CA · US

EUV Semiconductors
NL

Tier-1 Automotive
Munich · DE

MedTech Implants
Boston, MA

Diagnostic chips
Hilden · DE

Drug Delivery
Massachusetts · US

Inquire now

Request a feasibility study - free of charge.

30 minutes online, starting with a concrete component description. Together we'll determine whether a vision-driven cell is suitable for your application — objectively, without sales pressure.