How to quantify coating defects by using advanced image recognition methods

 
Have you ever asked yourself how to turn subjective opinions on coating defect evaluation into structured data? The optical evaluation, especially of large test series, by well-trained lab personnel has several disadvantages: The evaluation is often time-consuming and highly subjective. A quantifiable, reproducible result is hardly possible. A precise and objective analysis of coated test samples is possible by using computer vision. These methods allow lab workers to make a comprehensible and reproducible decision for the objectively best formulation at any time. The technology is very simple to use: the user takes a photo of the coated sample. The software evaluates micro and macro foam as well as craters in seconds. The results are stored in the browser-based app and can be documented as a test series. In this way, even a larger number of formulations can be evaluated in a short time. In the presentation, the backgrounds of image recognition on coating samples will be explained, and the software will be shown live using laboratory samples.