top of page
Defect detection
in mass production using
images
Until now, monitoring the quality of manufactured products has been a time-consuming process with inconsistent accuracy. With recent advances in Deep Learning, this process can be easily automated. At the end of this article, you will find a link that shows you how to create your own defect detection model using zerocode.ai.
The Challenge
Conventional product quality inspection is slow and inefficient, which can lead to production bottlenecks. With human inspection, accuracy depends largely on each individual inspector, and traditional automated systems are both expensive and difficult to implement.
The opportunity for deep learning
A Deep Learning approach is going to increase efficiency by enabling the integration of quality control into a fully automated production line and a more accurate analysis of the quality of each individual part. Because these deep learning models are automated and fast, they can be used in more production stages to avoid defective parts going through additional processes in the production process.
How is the AI model implemented?
Along the production line, cameras transmit images to the model. Modern production lines can then react to the model's prediction and either remove the part from production or release it for the next step.
Data requirements
The model is trained using images of manufactured parts that are classified as either defective or not defective.
bottom of page