Product Quality Assurance Using Computer Vision

The Problem

Many manufacturers still rely on manual inspection for quality assurance, especially detection of flaws on the assembly line. Visual inspection performed by humans, however, have limitations as the results vary depending on several factors depending on attention spans, energy levels, human senses and more. Such inspections can therefore lead to costly errors. Can AI be used to improve the product quality assurance process?

Solution

Computer vision models deployed on Kranium can be used to improve quality assurance. The models can be trained to identify anomalies from images or real-time video feeds from cameras. The models can review multiple features like curves, colors, corners and shapes, and then evaluate the product against a set of standards to determine if the products pass or fail the inspection.

The computer vision models on Kranium offer high levels of flexibility and can be easily trained to adapt to any product. Such models can drastically reduce the likelihood of flaws and frees us worker time for use elsewhere.

Furthermore, computer vision models can monitor real-time video surveillance feeds to ensure workers are wearing their safety gear and are not engaged in unsafe practices.