- Why ZIVID
The Zivid 2+ 3D cameras are now available and have a lot to offer in terms of cutting-edge innovation. But probably the feature that really captures people’s attention is its unique ability with a range of transparent objects. The source of these astonishing capabilities is our brand-new vision pipeline technology called Omni Engine.
With the introduction of the Omni Engine in Zivid 2+, there is now finally an industrial 3D camera that has all the capabilities required to be a perfect fit for the challenges posed by multi-SKU piece picking in the world of e-commerce fulfillment.
Here are some examples of the results you can expect on transparent objects with Zivid Omni Engine:
This new machine vision innovation is truly a system-level improvement. Its implementation is a combination of new hardware and firmware, optical system, dynamic projector, software support, and networking improvements. While on the outside Zivid 2+ looks like the Zivid 2, inside, it is a very different beast. It is a 3D machine vision product that can see a whole new range of items, not previously possible to capture, with absolute clarity and completeness.
Transparent objects were considered impossible at this stage. This was stopping perhaps the biggest automation innovation possibility out there, piece picking in E-commerce. Zivid simply couldn’t accept that, so we went and did it. We can proudly say that for the first time anywhere, Zivid has unlocked transparency in 3D machine vision.
- Henrik Schumann-Olsen, Co-Founder and CTO
Watch Henrik's keynote "Unlocking 3D imaging of transparent objects":
Zivid has a broad range of customers across typical automation segments such as manufacturing, inspection, and logistics, but now we are expanding into the world of piece-picking in e-commerce. A typical order fulfillment center has extensive use of automation across its operations. But piece-picking automation has proven to be sporadic and largely unsuccessful, and there are a couple of reasons for this:
The transparency challenge is such: Transparent items: plastics, polybags, shrink-wrap, and bubble wrap have long been considered impossible to reliably capture point clouds on. Even in machine vision research, there has not been a formal method or solution found to reliably capture 3D data on such materials. There are myriad and very complex reasons why, but in short, transparent materials scatter and reflect light in such complex ways it becomes very difficult to identify what is good data, to build the point cloud from, and then what spurious data should be discarded.
The transparency challenge had been considered by many in machine vision as extremely difficult to overcome (it was), in some quarters it was considered impossible to fix anytime soon (we proved it wasn’t). We found this to be enormously irritating, an entire market was desperate for a solution. Henrik Schumann-Olsen, Zivid co-founder and CTO stated that reliable 3D capture of transparent objects was considered impossible to fix in the near term by the machine vision community. The inability to deliver on transparency was seriously holding back large-scale automation in an industry that was one of the most eager to automate, namely e-commerce.
The plan began to be assembled in 2020 when we began engaging very closely with the major players and companies with a strong focus on e-commerce innovation. We soon understood how hungry these companies were to solve the piece-picking transparency challenge in a repeatable and reliable way.
In parallel, we were designing the Zivid 2+ 3D camera platform which would incorporate all the learnings and solutions to challenges that we would discover. Then, with Zivid 2+ ready to launch, we knew we had a 3D camera platform that was going to really shock people with its unique capabilities. What really excited us was that we knew we had a genuine world first on our hands with an industrial 3D camera that worked reliably on transparent items.
Here is a visualization of a series of transparent objects captured first without and then with Zivid Omni Engine:
The Zivid 2+ and the Omni Engine are a genuinely unique, and significant leap forward for 3D machine vision. Encompassing hardware, software, optics, and more. It can best be described as a completely new vision pipeline incorporating a broad range of innovations. We demonstrated the capabilities of Zivid 2+ and the Omni Engine at this year’s Automatica show, and you can read about the results here or you can just watch this short interview:
With the Omni Engine and the supporting SDK 2.10, Zivid 2+ is already being deployed at large fulfillment centers where it is proving to be a game-changer. The Zivid 2+ 3D camera family has not just unlocked transparency but has opened the door to a massive market; piece-picking at scale in e-commerce order fulfillment.