4D Vision in the Real World: Where 3D Fails and 4D Succeeds

One key to achieving industry-wide adoption of 3D vision systems for robotic guidance is the ability to deal with a large variety of materials and objects. However, traditional 3D cameras, such as structured light cameras, fail to properly capture many objects that are made with certain materials. Failure is due either to the object's color and reflectivity or its geometry.

Khatoonabadi introduces 4D vision systems and presents three common types of materials that 3D vision systems struggle with. He shows how a 4D vision system can handle the challenges presented by these materials.

Translucent material. Objects made of this type of material offer variable geometry, depending on angle of view. Thus, they are particularly challenging to capture with traditional 2D or 3D vision systems, and difficult to locate using CAD-based methods.

Reflective material. Objects made of this type of material saturate the camera sensor at certain angles due to high reflection. 3D cameras by nature aggravate the issue because of to the light they emit from their projector.

Thin-sheet metal. When viewed from the thin side, objects made of this type of material present a very small area for 3D cameras to register.

About the presenter
Armin Khatoonabadi is co-founder and CEO of Apera AI, a company working to develop human-like vision for industrial robots. He has a background in identifying new applications for technologies and extending them into new industries. He started his career in the late 1990s as an entrepreneur and established his first company in the field of internet and VoIP (Voice over Internet Protocol) services. Since then, Khatoonabadi has continued as an entrepreneur and investor in high-tech startups.