The CONT industrial partner supplies automotive components for different car producers. To ensure the highest quality of products it uses automatic and manual final tests. The automatic test presumes comparison of images taken by one or more cameras, placed in the check stations along the assembly line, with images of expected product. If the product fails during the automatic test it is sent to the operator that performs manual test, while comparing the product image with reference image. In order, to reduce the number of quality incidents (e.g. related to false positives) during automatic test, ZDMP platform offers a set of services utilizing Artificial Intelligence to learn defects types and acceptance limits, improve testing programs though analysis of false positives resulting in improvements of base models for optical check.
In the case of the negative automatic test, operator performs the manual check of the product comparing it with the reference images. The operator decisions with corresponding images are collected and stored to learn or extract the defects types and acceptance limits.
In the case of the negative automatic test, operator performs the manual check of the product comparing it with the reference images. The operator decisions with corresponding images are collected and stored to learn or extract the defects types and acceptance limits.
The ZDMP platform provides an optimisation services for the quality check performed within the CONT assembly line, resulting in the reduction of false-positives during the automatic test, as well as creating the models based on operator decisions used for generation of acceptance patterns.
This use-case has the goal to optimize the quality check process within the CONT assembly line. Data that are not utilized by the previous quality check process are used by ZDMP platform services to complement the process.
ZDMP platform assists the internal quality check system. The quality check stations along the assembly line provide the images for optical quality control. These data along with quality test results are accumulated and analysed by ZDMP platform, and the feedback is generated.
The test check stations along the assembly line equipped with the cameras serving the goal of optical quality control. Data in the form of images taken within these check stations is a valuable resource that is used not only to check the quality of product, but also to improve the efficiency of quality testing programs. The images taken allow detecting, for instance, defects related to the shape of the product.