Skip to main content

While the teams in every factory strive to produce defect-free parts, in some cases, subtle surface defects may be acceptable provided the integrity of the product is not compromised. But that’s never the case for parts with Class A surfaces. From B-pillars to bumpers to visible surfaces on home appliances and electronics, OEMs and end customers expect them to be perfect. 

Ranging from highly reflective surfaces to parts with matte finishes, detecting quality issues on Class A surfaces has been a challenge.  Many defects are almost impossible to see with the human eye. Traditional machine vision solutions have helped catch some defects but have been difficult to deploy. And, even if these systems can catch the defects, they don’t provide any insights into what caused them. 

“Detecting defects in parts with Class A surfaces is one thing, preventing them is another,” explains Cristian Filipczak, Application Engineer. “For those looking to move beyond sample testing and basic detection to leveraging machine vision to prevent common, difficult-to-detect surface issues, our solution can help.”


For machine vision to deliver detection of multiple common defects on injection molded, Class A surfaces, special image capture and illumination is required. We’ve created two configurations that create the ideal lighting conditions for capturing images that provide a consistent view of the surface. Our software then merges and enhances the images to provide a virtual view where various “invisible” defects become obvious.

Along with generating detailed images, our solution captures the machine sensor or process data that you need to see (i.e. injection pressure, cycle time, fill time, hold time, plastic cooling, melt and tool temperatures, etc.) and merges it with the image data to create a digital record of each part.

The combined image and process data on each part are used to generate machine learning models that deliver real-time alerts to the shop floor team if any issues are detected. While flagging issues in real-time lets you take immediate action on defective parts, it’s the analytics and insights generated by our software that set you on the path to defect prevention.


Monitoring the surface of every part and exposing defects in real-time allows you and your team to focus on any parts flagged by our system.  Using our online software, you can analyze flagged parts and connect the dots between process variation and defects.  And, if you find it difficult to find the time to conduct the analysis, our software does it for you.

“While our active monitoring detects a variety of common defects on high gloss or matte surfaces ranging from short-shot and flow lines to ghosting, you need more context around the cause of these defects,” says Filipczak. “Think of our solution as a virtual data analyst. It correlates critical data, identifies trends, and delivers insights that allow you to get proactive about avoiding defects and to feel more certain in the quality of every part.” 


These real-time insights can be scaled across all your molding machines making the same or similar parts offering you access to standardized digital records across all deployed lines in our online software.

And if the molded parts are moving on to another process such as welding or painting, data from those applications can be linked to the digital part records. 

“For parts and components where multiple processes are involved, quality and process issues can arise at a variety of different points,” explains Filipczak. “With additional data and broader analytics and insights, you can decrease the chances that a defect will be missed. You can start making the concept of zero defect manufacturing a reality.”  

Follow us on LinkedIn to keep up on the latest news about our machine vision solutions for plastics.

Check out our case studies for more information on our injection molding, welding and paint solutions.