Skip to content

Inspect turned parts

Inspection of turned parts with image processing and deep learning

From belt tensioners to the pins that hold watch straps. Turned parts can be found everywhere in industry. Each of these parts must be inspected and quality documented after production, and turned parts are often manufactured with high precision and extremely tight tolerances. Machine vision systems must therefore offer very high resolution and measurement accuracy in order to detect minute deviations.

EyeVision 3D software can help automate production processes by inspecting turned parts quickly, accurately and repeatably, monitoring both dimensional accuracy and surface quality.

Common defects that arise are

  • Dimensional deviations: Exceeding or falling below tolerances in terms of diameter, length, holes, angles and other geometric properties.
  • Surface defects: Scratches, dents, cracks or material defects.
  • Shape deviations: Deviations from the desired shape or symmetry.
    Missing or incomplete features: For example, an opening that has not been fully drilled or a slot that has not been made.

In addition to the software, EVT also offers an all-in-one solution for turned part inspection:

The turned parts inspection system based on the EyeSVC (Smart Vision Cabinet) enables precise inspection and testing while the turned parts are still in motion. A trigger input ensures that the turned parts are detected either in flight or as they glide past and the inspection result is available a few milliseconds later to control the actuator.

A comprehensive statistics module makes it possible to display the production quantity and the measured development values on the integrated monitor at any time. As the entire system is housed in an IP65 cabinet, it is not at risk from the abrasion of the turned parts. In addition, the ready-to-use system can be configured quickly and without programming, so that turned parts of all kinds can be optically measured with just a few clicks. The EyeSVC can be retrofitted anywhere and has numerous interfaces.

With the Deep Learning add-on module, many errors can be detected that are not based on measurement technology but require classification. For example, whether there is a chip in a cut.

Various cameras are available for the EyeSVC to ensure optimum image capture. Inspection from several sides is also easy to implement. The standard version of the system can support up to 4 cameras. If this is not sufficient, you can switch to the next larger system.

 

Try out the EyeVision Software for free and convince yourself!