Ai Build implements AI to detect and correct 3D printing errors in real time

We reported only yesterday on Inkbit’s new 3D printer – it is powered by machine vision and learning systems to correct printing errors in real time and to predict the warping behaviour of materials. Similarly, the London-based startup, Ai Build, just unveiled its new real-time inspection technology for additive manufacturing using computer vision and machine learning.

An obstacle to large-scale 3D printing are production defects, therefore Ai Build’s technology will prove beneficial as it will automatically detect the most common printing issues, such as material shrinkage and warping in real time. As we know, additive manufacturing technologies were mostly used for rapid prototyping or low-volume production until recently. However, this is changing as predicted by multiple reports on the state of AM in the last few months. We are now entering an age where large-scale 3D printing could become the norm for certain industries or products – new technologies are therefore necessary and artificial intelligence is likely to play an important role.

Example of production defects such as overextrusion or underextrusion | Photo Credits: AiBuild

Ai Build is the startup working towards providing a completely autonomous large-scale 3D printing technology to produce their designs at their own facilities. They also help other businesses to set up their manufacturing environment, give initial training and provide technical support. So, what are the benefits of this new technology? As mentioned above it is a real-time detection of 3D printing defects with artificial intelligence technology. It should increase productivity and level automation in additive manufacturing facilities and at the same time reduce production costs and material waste. Their AiMaker, the high-precision robotic end-effector that attaches to industrial robotic arms is now embedded with a camera and GPU module that detects manufacturing defects by comparing its real-time camera view to previously recorded images of 3D printed parts.