Vision Detection and System Integration

Customized vision solutions and their integration into your process or workflow. With our great experience in vision technology and the integration into real life environments we are able to provide you with the inspection solution for your needs.

We have designed, built and implemented computerized vision detection systems for many applications in a large variety of fields. This includes applications in production departments, laboratories, and quality departments. Our solutions start with simple inspection cells operated manually or by a third party handling system and they go as far as fully automatic inspection machines including automatic loading, handling, sorting, unloading, and physically testing; supervised by several computerized vision detection systems controlling the sample testing step by step and interfering accordingly.

Appearance

By applying appearancebased methods automatic product identification and classification can be performed. Non-expected changes in product or batch appearance will be detected. Furthermore, with deep learning based approaches normal changes in products or batch appearance and environmental variability can be suppressed; therefore, only the defect product with non-expected behavior will be detected. These deep learning based software frameworks are algorithms in machine learning which allow to vision detect otherwise impossible inspection and classification challenges. This makes vision detection suitable for a whole new area of applications in production, packaging, laboratory, and quality departments.

Detection of missing or wrong objects in delivery trays

 

 

 

 

Detection of RFID within injection molded  products

 

 

 

 

Inspection of glue line on product against   missing glue

 

Measuring

Advanced vision detection systems allow not only to detect features on samples or products but also allow to measure them. This includes measuring of: Liquid heights in containers, overall dimensions of parts or features, distances between features, circularity of holes, straightness of edges, angles between features, and many other measurements. By use of several camera systems also features and distances between different images from different cameras on large products can be measured without getting the product moved. Furthermore, automatic reading of numbers, letters, and barcodes by the vision detection system will allow to check for correct labeling of products or samples.

 

 

 

Detection of
objects in samples
and measuring of
fill levels

 

 

Measuring
of diameters,
circularity, and
positions

 

 

Counting of
objects and
patterns

 

 

 

Recognition of
numbers, letters,
and barcodes

 

 

 

 

 

 

 

 

Aestethic inspection

Aesthetic inspection includes the detection of defects on surfaces regarding surface structure as well as printing or engraving on surfaces.
Surfaces of products or samples come in many different types made from different materials and produced in different processes. This results in wanted and unwanted surface structures. Wanted surface structures such as grinded surfaces or brushed surfaces for decoration are surface variabilities which have to be suppressed during the vision detection process by use of deep learning based approaches or similar methods. After suppressing such wanted surface structures the unwanted surface defects such as scratches, holes, or stains can be detected.
Printed or engraved surfaces often include letters, numbers, or symbols. Those letters, numbers or symbols might still be readable but form quality aspect they do not reflect the high quality of a product. Therefore, by use of deep learning based approaches such deviations from wanted printings can be detected and appropriate measures can be taken.

 

 

 

Detection of scratches and other anomalies on a surface.

 

 

 

Detection of stains or changes in colour.