Point Cloud and How to Digitally Transform Your Manufacturing Capacity

In the past three decades, the advancement of technologies has changed human lives at the speed of light. As the business landscape undergoes tectonic shifts, manufacturers are faced with the daunting task of finding automation solutions that allow for production agility while achieving the best quality on time. The answer is to team up humans with robots. With that said, the manufacturing sector is facing an estimated 2.69 million baby boomers expected to retire by 2028, digitizing critical manufacturing knowledge became especially pressing.

Over the years, manufacturers have started to incorporate Point Cloud and 3D machine vision to codify employees’ expertise and fortify their knowledge database. While robots can take over rote tasks, organizations need to start retraining their counterpart to facilitate more interactions between workers and automation systems. This will create a symbiotic smart factory that can quickly pivot in any economic environment.

What Makes Manufacturers Choose Point Cloud?

A point cloud is a data set that defines the shape or structure of an object in space. The base element of a point cloud is the single point. A point represents a specific location in space and is comprised of three coordinate values. The Cartesian XYZ coordinate system is most common, but cylindrical, spherical, or any other coordinate system that can define a location in 3D space is also applicable.

A point cloud is made up of one or more points representing the 3D characteristics of the objects being viewed — both big and small. From kilometers to nanometers, the units are arbitrary and application-dependent. The data’s representation is as simple as a list, where each row identifies the three coordinates for each point.

Point cloud data can be augmented with additional data attached to each point. Characteristics such as reflectance, RGB color, and surface normal direction are commonly used to help downstream processes identify features, detect objects and perform a 3D inspection. Because of its robust digital representation of objects that were often hard to identify by a 2D vision system, industrial manufacturers now rely on 3D machine vision to recreate human judgment. In summary, there are three major advantages to automate using point cloud–

  • Captures high-accuracy data regardless of the object size,
  • Easy-to-understand 3D digital representation of target objects, and
  • Lossless data that can be fed to other downstream processes.

How to capture point cloud data?

Point cloud data can be captured using many different technologies: stereo vision, laser triangulation, coded light, structure light, and many more. Amongst these technologies, the robustness, and simplicity of integrating a laser-triangulation scanner make it the leading method of capturing 3D data in industrial environments. Watch this video to find out how each method works to advance industrial automation.

Tracking and Tracing Steel Billets

Steelmaking requires stringent quality monitoring as many major structures rely on steel for a durable and strong foundation. Tracking and tracing a semi-product through the manufacturing process ensures mills deliver top-quality steel billets to their customers.

Traditionally, this was done by mill workers manually affixing information tags using a nail gun or welding them to the hot steel billets. During this stage, the working temperature can reach up to 430°C/800°F. Not only was it dangerous for workers, but it was prone to human error.

The solution was to implement a robotic arm equipped with a 3D machine vision camera. The 3D scanner captures the point cloud data of the rough flame cut billet surface. The data is then analyzed to determine if a tag is already present. If not, the software looks at the 3D data to calculate an optimal tagging spot.

The dangerous and error-prone task of attaching tracking information to hot steel billets was eliminated, with the operators upskilled to provide the system’s onsite supervision. It not only automates a repetitive task, but the system has helped improve workplace safety, streamlined production, and significantly eliminated information errors.

Helping Palletizers See Clearly

Major produce warehouses pack and dispatch fresh fruits and vegetables every day. It is a process that requires care and efficiency to prevent goods from being damaged and spoiled.

Before dispatch, an OCR scanner reads the label on the produce box and directs it down to the right conveyor belt for pelletizing. However, the robotic palletizer would frequently pick up boxes with open flaps, deformed boxes or place them in the wrong orientation, resulting in workflow disruption and sometimes even damaging the products.

The solution is to install a 3D machine vision camera that captures every shipping box’s point cloud data. The data is analyzed to identify packages with damage, open flaps, or bulging contents. The system then diverts these boxes down to another conveyor belt so operators can repackage them. Since the point cloud data is inherently spatial, it provides the data necessary to precisely position and orient the palletizer’s end-effector, ensuring it grips and stacks the box in the correct direction.

Upskilling to Solve the Talent Shortage

History has shown that with every industrial revolution, more jobs were created as a result. Many skilled workers whose tasks were automated went on to become an integral part of the company that further propelled a more human-centric working environment.

 Teaming Coordinator

From the previous case study, the produce warehouse upskilled some of their workers to become teaming coordinators. Teaming coordinators are responsible for identifying opportunities where vision-guided robots can improve productivity and throughput. They are also tasked to train human team members to collaborate efficiently with robots. Their knowledge and experience in the production process are invaluable in linking these two assets together.

Safety Supervisor

A safety supervisor is becoming increasingly crucial in smart factories, especially post-pandemic. For instance, by training a former steel billet tagger with workplace health and safety regulations, he or she can utilize their experience to identify a potentially hazardous or unsafe site that needs to be digitally transformed. Tapping into their knowledge can also help system engineers design safety measures between humans and machines.

Quality Analyst

While machine vision has proven its capabilities in automating inspection tasks, human inputs are the key to developing algorithms that work with point cloud data. A quality analyst with solid operational competency is the right candidate for this position. They should receive training in data analysis to create new inspection instructions that improve the facility’s quality rate. Their keen insights on how new QC technologies can be incorporated across the facility determines how successful they are at their job.

When Scale No Longer matters

By now, most manufacturers across industries are familiar with the benefits of embracing digital transformation. The PWC industrial manufacturing trends 2020 also shows that companies that invest in technology and talents with Industry 4.0 training have a stronger position in adapting to the ever-changing consumer demands. As previously outlined in this article, a successful manufacturing DX strategy relies not only on technology but also on transforming your employees’ experience and knowledge into actionable intelligence.

Understanding how to incorporate 3D machine vision into your business case is the first step to fortify and expand your digital manufacturing capabilities. To learn more about Hermary’s technology, visit our Video or Learning page. Alternatively, you can contact us with any questions.