Zivid 2+
Zivid 2
Software
Get Started
Industries
Get Started
Support
Downloads
Customers
Zivid 2+
Zivid 2
Software
Get Started
Industries
Get Started
Support
Downloads
Customers
.svg "Zivid")
.png "piece picking demo zivid automatica 2023")
The high-speed and demanding world of e-commerce has been at the vanguard of robotic automation deployment for many years. Modern fulfillment centers are alive with automated transport of thousands or, in some cases, millions of SKUs (stock-keeping units). Whilst automation thrives in these centers, one stubborn problem has remained: reliable piece-picking of all items, regardless of opacity, transparency, shape, or size. As part of the Automatica trade show, held in Munich, Germany, Zivid has brought together the best experts in robotics, picking software and machine vision to create a smart, fast, and reliable piece-picking demo that solves this challenge.
Zivid, The Gripper Company, and Fizyr have come together to create a solution and see what a truly specialized piece-picking cell can achieve. The Zivid 2+ M130 provides high-resolution 2D images with complete 3D points clouds on some of the toughest of objects, even transparent ones. With this data, the Fizyr software can find and identify the complete content of the bin, including items like shampoo bottles and clear plastic containers. To take advantage of this advanced imaging and detection system, The Gripper Company has implemented their MAXXgrip gripper system, designed to access items in deep bins and be able to carry and place them in a gentle manner.
The piece-picking demo, shown for the first time at Automatica, proved to be able to handle a wide variety of objects without damaging the goods. The parts were moved back and forth between the two bins continuously, being picked and placed with ease. This picking cell proved the synergy that occurs when you integrate these three high-performance components to make a specialized picking solution.
With this demo cell, we have brought to the market a new 3D camera that is the first in the industry that can "see" transparent items. This is quite disruptive technology."
Henrik Olsen-Schumann, CTO and co-founder of Zivid
The E-commerce companies of the world keep their eyes turned toward automation. The giants of the online shopping market need to continually update their order fulfillment methodology if they want to stay at the top of the market. One of the areas that have remained largely unexplored is having a piece-picking cell that can handle virtually anything in the inventory. Most of the piece-picking cells now are expected to handle the easy and simple items, relying heavily on human intervention on items that are thought to be difficult to see, or difficult to pick. These cells are also seen as being at a high risk of damaging items, so delicate objects also would be funneled over to human operators.
For the Automatica trade show, held every two years in Munich, Germany, our goal was to create a piece-picking cell that addresses the key corner cases that are seen in the order fulfillment industry. This includes polybags for clothes, transparent boxes for food storage, anything that has a clear or transparent container, extremely shiny items, and more, all thrown together in the same bin. We went at this with the mindset of "if we can solve the hardest bits, we know we have the rest".
We also knew that the vision system is just one piece in a large intricate puzzle in creating this piece-picking cell, so we teamed up with other companies that are pushing the edge of what is possible in their fields.
Get a sneak peek at the piece-picking cell in action:
The biggest challenge that creating a piece-picking cell brings is being able to capture the wide variety of objects in the bin. Usually, to cover a wide range of objects, you need multiple capture settings or a very long time to make sure you cover the full range of objects. This includes everything from dark, extremely absorptive objects, to translucent shiny materials. The camera must be prepared to capture and image anything. These items need to be captured with high accuracy, not just for their pickability but for their placement, too. Eventually, this will allow for the packing of the items for shipment and a better storage solution.
The demands on the accuracy of the system in piece picking are high, and not just on the camera, but on the detection system and picking system as well. The camera is the first piece of the puzzle in what it takes to excel at handling items. The next step is using that data to the best of our ability: we needed a system that could detect and handle a wide range of items without a provided CAD model.
The final step in making sure we can get any item we see is how we execute the pick pose. There are many gripper designs out there, but finding one that can act almost universally is a tall task. It needs to pick accurately and gently and be able to handle oddly shaped objects. If it picks books, how will it handle the fact the pages can separate? What about shoe boxes? Will the gripper be able to pick soft items wrapped in plastic without damage?
For this application, we have chosen to use our new 3D camera: Zivid 2+ M130. The Zivid 2+ delivers true-to-reality point clouds with a spatial resolution of 0.35 mm and dimension trueness >99.7%. Thanks to the Omni Engine, the Zivid 2+3D camera also offers innovative 3D technology to consistently unveil transparent and reflective objects with temporal structured light. This is truly a revolution, as no 3D camera has been able to image transparent objects before. Today, the Zivid 2+ is the only 3D camera capable of such performance.
In the 2+ family, the M130 is specifically designed to accelerate automation in the world of e-commerce order fulfillment by enabling performant piece-picking across millions of inventory SKUs. With an optimized field-of-view of 79x65 at 130 cm and a working distance between 80 and 200 cm, the M130 ensures precise and accurate imaging with a generous working volume and spacious robot clearance. Even when utilizing the full working range of the Zivid 2+ M130 at 2 meters, you can still expect the best. But our cameras are just one component in the system our customers need to succeed, and we wanted to see what we could do if we collaborated with our cohorts in the automation market.
Fizyr has been working with Zivid cameras for years, using them across multiple customer projects as their vision provider. Zivid and Fizyr have worked to build a symbiotic relationship that can provide high-quality picking results from the precise and clean point clouds the Zivid cameras provide. With this software, finding what we need to see is a breeze. One more piece to the puzzle needed to be solved: How will we pick what we can now find and see?
This is where The Gripper Company comes in. The Gripper Company focuses on designing grippers that go above and beyond the standard pinching or suction procedure, and on making grippers that are truly specialized. In this case, they have designed a gripper that excels at versatility. The MAXXgripp gripper system is not only a suction cup gripper, but also a claw. It is designed to pick effectively and carry gently, regardless of the shape.
With the MAXXgripp attached to an industrial robot (Fanuc m-10id/12), two Zivid 2+ M130 cameras fixed to the walls, and the Fizyr software implemented into the system, the demo cell was finally complete. Have a look at this interview of Zivid, Fizyr, and The Gripper Company talking about the different components:
The piece-picking cell turned into a success, being a focal point of the stand that Zivid held at the Automatica 2023 trade show in Munich, and attracting many curious visitors. With the combination of hardware and software from all three companies, the system could successfully pick objects that were normally considered unpickable for most of the piece-picking world: clear boxes, shampoo and soap bottles, and boxes with other items rattling around inside them... These key items were normally seen as human handling only but this demo opens new opportunities for the logistics and e-commerce industries.
