Computer Vision - Automating Generic Detection of Key Metrics

With the printed part metrics and the surrounding powder metrics reported independently, prioritization of design change performance evaluation was supportable.

OpenCV - Automated Segmentation, Filtering, & Data Sampling

With any discipline of work given enough time it becomes clear which steps are most critical. It makes sense to invest time making more efficient the time-consuming or arduous steps when working in a resource or time constrained environment.

One component of the thermal analysis work to report metrics used in feedback to process improvement required sampling the thermal trends for separated printed parts and surrounding powder during a print job. Because the print job contains a high mix of geometric content variability, either a human or an automated system needs to determine regions for data sampling. In the early phases of development the human was myself, but due to the increasing volume of print jobs and their printable content, I needed to find an automation solution. I did this using the OpenCV bindings available using Python. This allowed engineers to run varied experiments and collect the thermal data without my continued intervention.