A defective part often only becomes costly once it reaches the next stage of production.
A component may be defective, incomplete, or simply misaligned. If it enters an automated assembly station unnoticed, this can result in rework, scrap, production interruptions, or additional manual inspections.
As production speeds increase and the variety of product variants grows, this challenge becomes more significant. At the same time, there is a need to reduce manual inspections and make feeding systems more flexible.
This brings one point in the production process into sharp focus: the output of the feeding system.
Here, the parts have already been separated and are moving toward the next production step. It is precisely at this point that industrial image processing can inspect, make decisions, and, if necessary, immediately trigger a sorting action—before a minor deviation turns into a major production problem.
More Automation—Without Increased Mechanical Complexity
Manufacturers continue to invest in smart and flexible production systems.
According to Deloitte’s 2026 Manufacturing Industry Outlook, 80% of manufacturing executives surveyed plan to invest at least 20% of their improvement budget in smart manufacturing initiatives. These include, among other things, automation hardware, sensor technology, and data analysis. The key question, however, is not how much additional technology can be deployed.
Rather:
How can a process be automated more flexibly without simultaneously making it mechanically more complex?
This question plays a particularly important role when it comes to spiral conveyors. Mechanical guides are traditionally often used to ensure the correct orientation of components. They must be tailored to the specific component and may require adjustments when new geometries or variants are introduced.
Camera-based sorting offers a different approach here.
Instead of relying solely on mechanical means to ensure that every part arrives in the correct orientation, the vision system can detect the actual position and then decide how to handle the component. In suitable applications, this allows for the reduction or complete replacement of complex, component-specific mechanical guides.
From the Exit of the Spiral Conveyor to the Quality Control Point
A feeding system separates components from the bulk material and feeds them into the subsequent production process. When a part leaves the spiral conveyor, it is thus in a suitable position for optical inspection.
Before the component continues on, an inspection system can determine, for example:
- Is it the correct part?
- Is it correctly oriented?
- Are all required features present?
- Does the contour match the specification?
- Are there any visible defects?
- Are the defined dimensional criteria met?
This transforms a simple transfer point into an automated quality control point prior to assembly.
The EyeSorter Checkbox from Eye Vision Technology combines inspection, orientation checking, counting, and automated sorting into a compact system for feeding and assembly processes.
GOOD, ROTATE, or BAD: Not every deviation is a defect
Industrial sorting is often reduced to two outcomes: Good or Bad.
In practice, however, there is an important third case.
A component may meet all quality requirements but still arrive misoriented.
If it is rejected for this reason, unnecessary scrap is generated. If, on the other hand, it is passed on without inspection, problems may arise at an orientation-dependent assembly or handling station.
The EyeSorter Checkbox therefore distinguishes between three result classes:
- GOOD – The part meets the defined criteria and can proceed.
- ROTATE – The part itself is acceptable but is oriented incorrectly. It can be diverted or reoriented accordingly.
- BAD – The part does not meet the defined quality criteria and is rejected.
Separate parameters, real-time counters, and digital outputs are available for these classifications.
This ensures that a usable part does not automatically become scrap simply because it was fed in the wrong orientation.
What can an automated parts inspection system detect?
Which features are inspected depends on the specific component and the application.
Typical inspection criteria include:
- Orientation and position
- Presence or absence of features
- Contour and shape
- Dimensional characteristics
- Visible damage
- Foreign or incorrect parts
- Piece counting and production quantities
EyeSorter systems can be used for various types of parts—for example, screws and fasteners, nuts and metal parts, plastic components, caps, turned parts, or other automatically fed components.
The size of the component alone is not the decisive factor. Transport, geometry, required features, and desired inspection speed determine which camera concept is best suited.
An EyeSorter family for various inspection tasks
Not every sorting task requires the same image capture.
That is why different EyeSorter configurations are available.
The EyeSorter Checkbox Standard uses an integrated line scan camera and achieves a line rate of up to 14,000 lines per second.
For more extensive applications, the EyeSorter Checkbox Advanced is available with an integrated computing and control unit for external line scan camera configurations.
The EyeSorter Checkbox Flex uses high-resolution area-scan cameras and is therefore also suitable for more complex inspection features or applications that require a complete image of the component.
Depending on the application, the system can also be expanded with additional cameras or sensors—for example, for an additional incident light inspection or to detect other optical features.
This allows the image capture concept to be adapted to the actual inspection task without changing the basic operating and sorting concept.
Quick Inspection—and Reacting at the Right Moment
A high frame rate alone is not sufficient for reliable sorting.
After a component has been inspected, the system must know where it is located when the sorting decision is to be implemented.
EyeSorter therefore combines image processing with encoder information and digital I/Os.
This allows the result to be used immediately, for example, to:
- send a GOOD part onward,
- reorient a ROTATE part,
- eject a BAD part,
- control pneumatic or mechanical sorting mechanisms,
- or pass the result on to the higher-level automation system.
The inspection result does not remain on a screen.
It directly drives a process decision.
AI-Assisted Teach-In Instead of Complex Programming
Flexibility becomes particularly important when a system processes different components or variants.
A purely mechanical solution may need to be adjusted in the event of a geometric change. Even a traditional vision system can require additional engineering effort if a new inspection program must be created for each variant.
The EyeSorter Checkbox therefore employs a teach-in approach.
The component is taught via the intuitive user interface. AI-assisted evaluation helps capture relevant features with just a few user actions and use them for recognition.
If necessary, the inspection parameters can then be further adjusted.
Part types that have already been set up can be reselected. This allows for faster product changeovers and reduces the need for specialized image processing knowledge in daily operations.
The basic principle remains the same across the various EyeSorter configurations:
Teach-in. Inspect. Sort.
Less Mechanics, More Flexibility for Variants
Part orientation, in particular, is often a challenging task in traditional feeding technology.
Mechanical guides must be designed to match the part’s geometry, set up, and, if necessary, adjusted when changes occur.
Camera-based orientation detection can make this process more flexible.
EyeSorter detects whether a part is correctly oriented. If the part is of acceptable quality but incorrectly oriented, it receives the classification “ROTATE.”
This means that correct orientation does not always have to be fully enforced via a part-specific mechanical guide.
Especially when dealing with changing parts and variants, this can reduce the effort required for retooling and setup and make the spiral conveyor more flexible to use.
From the sorting station to SCADA and digital production
A modern inspection application does not end with a digital output. The results obtained can also be used for higher-level automation and data systems.
EyeVision supports industrial communication standards such as PROFINET, OPC UA, Modbus, Ethernet/IP, and EtherCAT, thereby enabling communication with PLC, machine, and SCADA environments.
This allows, for example:
- Inspection results to be transmitted to the machine control system,
- Good/bad/rework counters to be recorded centrally,
- Process data to be visualized in a SCADA system,
- Quality data to be documented,
- or information to be provided for higher-level digital production models.
This means the sorting station can not only make immediate production decisions but also provide data for networked manufacturing concepts. Such data can, for example, serve as a source of information for a digital twin of the production process. The EyeSorter is therefore not the digital twin itself—but it does provide relevant real-world process and quality data that can be integrated into corresponding systems.
Integration Instead of Rebuilding the Entire Feeding System
Manufacturers rarely want to redesign an entire feeding system just to make quality control or orientation more flexible.
This is precisely where a vision-based approach offers a significant advantage.
The EyeSorter Checkbox can be integrated into existing production environments via digital I/Os, Ethernet, and encoder interfaces. EyeVision also provides interfaces to industrial communication and SCADA systems.
Instead of setting up an additional offline inspection station or developing a complex mechanical orientation system for each new variant, inspection, position control, and sorting can be integrated directly into the existing material flow.
Why inspect before assembly and not after?
An end-of-line inspection remains important.
However, it answers a different question:
Was the finished product manufactured correctly?
Inspection at the exit of the spiral conveyor addresses an earlier question:
Should this component even proceed to the next process?
An unsuitable part that never reaches the assembly line in the first place can:
- prevent a defective assembly from being created,
- prevent a station dependent on orientation from being
- blocked,
- prevent avoidable rework,
and prevent the need for later troubleshooting.
Automated part sorting is therefore not just quality control.
It is also process protection.
Stopping problems before they reach assembly
Smart manufacturing does not necessarily mean redesigning an entire production line.
Sometimes the greatest benefit comes from making a better decision at the right moment.
At the exit of the spiral conveyor, EyeSorter can inspect a component, determine its orientation, evaluate defined quality characteristics, and trigger the appropriate sorting action—before the part reaches the next production step.
- GOOD: continue.
- ROTATE: reorient.
- BAD: Reject.
And because different tasks require different image capture methods, EyeSorter Checkbox Standard, Advanced, and Flex offer various configurations based on the same fundamental, intuitive operating concept.
The goal always remains the same:
The right part must reach the next process in the correct condition and orientation.
Which EyeSorter configuration is right for your application?
The appropriate solution depends on component geometry, feeding system, inspection criteria, camera design, and production speed.
From fast line inspection to more complex area image analysis to an additional camera for supplementary visual inspections, the system can be tailored to a variety of sorting tasks.
Talk to Eye Vision Technology about your application—and see how much manual labor can be replaced by intelligent, vision-based sorting.