Collaborative robots need flexible 3D vision

Zivid is at the forefront of a new generation of 3D machine and robot vision companies for Industry 4.0.

Our products, Zivid One 3D cameras, are an essential step forward for the machine vision industry, with a unique combination of features in a robust, compact and cost-effective package. Of all the industry-leading features packed into the award-winning industrial design product, perhaps the most sought-after attribute of Zivid's cameras is flexibility.

Let me start by defining what flexibility: 

A truly flexible 3D camera is one that has the capability to adapt to a broad set of applications, distances, and environments.

Companies address flexibility from many angles. Some label their whole portfolio as flexible. However, this does not live up to the spirit of flexibility as I have defined it. The ease of use across a broad portfolio is lost, and the burden of climbing learning curves is moved to the end-user as they swap between products within the portfolio. These unintended costs outweigh the benefits of a portfolio that can address a broad set of applications.

Zivid 3D cameras are designed deliberately with a scalable hardware and software platform which allows flexibility across a family of cameras as well as within a particular camera itself. This flexibility shows up in multiple forms.

First, by providing a 3D camera platform capable of producing high-quality point clouds on a wide range of distances, Zivid One 3D cameras offer developers complete working distance flexibility to operate within their physical environment. From 30cm to 3meters users can capture HD point clouds of their target objects.

Second, flexibility applies to the user's hardware and software environment as well. Every developer has personal preferences or practical reasons for their PC's technical specs and operating systems. Zivid One cameras work independently of a user's operating system and are compatible with a wide variety of programming languages via an easy to use API. Developers can pick Python, C, C++, or another programming language suited for their tasks. To avoid creating an application from scratch, developers can also opt to directly grab the data format they need, including

• XYZ
• RGB
• XYZ+RGB
• Pure depth data

The camera's output can then work with off the shelf imaging tools from industry leaders such as MVTec and Matrox.

Zivid's vision hardware is designed to be modular. With the use of an external GPU, the flexibility offered from a processing performance standpoint is inherent in the camera's design. Without the need to change existing setups or touch the 3D color camera itself, the cost for customers to improve processing performance is as easy as a GPU replacement throughout the lifetime of their application

To us, providing flexibility for our customer's applications is derived by our camera's fast processing and acquisition times. The Zivid One acquires and processes images at 13 Hz (or a 3D image in less than 100ms) while providing a high quality and dense point cloud. With typical cycle times of 1 second, there is sufficient headroom for the end application.

Zivid's detail-oriented approach to industrial design is critical in offering another form of flexibility; the camera is built for use in manufacturing environments. With the advent and high growth of the collaborative robot market, it is important that Zivid One 3D cameras deliver the gold standard in quality of data in a small and lightweight form factor that is suitable for production line and robot automation. With IP65 specification, small size and weight footprint, Zivd One 3D cameras can be integrated with automation systems without compromising the workload capability of modern cobots. This type of flexibility enables integration with automation systems without compromising load capability of modern collaborative robots.

At the core, Zivid offers real-time capture of high-quality 3D data with ultra-high accuracy and resolution. These features enable our customers to take images of a broad set of objects and surfaces, and allows for faster, more flexible and reliable robotic operations.