Smart camera vision systems: the new approach to track and trace

Increasing global regulations on healthcare products have many pharmaceutical manufacturers preparing to implement serialised packaging in support of full traceability and authentication requirements. However, the traditional approach of deploying turnkey serialisation solutions that integrate PC-based vision can be expensive to install, validate and maintain.

A new leaner, more cost-effective approach to track and trace demands new technology. That’s why pharmaceutical manufacturers are increasingly opting for networked smart camera vision systems over PC-based inspection systems.

Now is the time to start serialisation

An increasing number of European countries are, or will be, adopting sophisticated tracking and ePedigree infrastructures throughout their supply chains. In the near term, Turkey and France will require item-level serialisation starting in 2011. Other countries with regulatory deadlines include Brazil, the United States (California) and South Korea. Countries having entered regulatory negotiations to set a deadline include Spain, Germany, Italy and Bosnia.

For manufacturers that have yet to begin learning about serialisation, now is the time to initiate a program or resume a stalled pilot project. Why? Because eventually these requirements will impact all pharmaceutical plants around the world.

As the pharmaceutical industry has consolidated, plants have become increasingly specialised, producing fewer products for an increasing number of markets. Therefore many producers manufacture one medication in a few or even a single location and then distribute that product globally.

Even though not all regions have enacted ePedigree legislation, it’s very important to know that manufacturers exporting into regulated markets must adapt their packaging to conform to specific regulations even at facilities located in countries that have not yet imposed legislative deadlines.

Of course, with compliance comes improved patient safety, product integrity and supply chain security. However, beyond compliance, drug manufacturers have discovered real value in being able to stop counterfeiting, prevent parallel trade through unauthorised channels and achieve greater visibility into how products are made, distributed and used.

Implementing traceability for product safety and quality control not only helps firms isolate the source and extent of safety or quality control problems to minimise production and distribution of unsafe or poor quality products, but it also reduces the potential for bad publicity, liability and recalls. And, in the event of a recall, the more granular the tracing system, the faster a producer can identify and resolve product safety or quality problems.

With the tremendous amount of work that will be required, it’s not too early for pharmaceutical producers to actively engage with equipment and software suppliers to map out their traceability requirements. Now is the time to learn about data carriers, coders and markers, labellers and printers, ID readers and vision systems, and to start planning to implement the software infrastructure required to support data sharing with trading partners.

New requirements drive new technology

Whether implementing traceability at the batch level, or implementing serialised packaging to support full traceability for ePedigree, producers must deploy a broad range of technology and software platforms, spanning all levels, processes and systems. At the highest levels, ERP systems typically interface between the supply chain and plant-level systems such as MES, line management and serialisation databases.

Here, PCs are an essential component of a traceability system, where they are a great tool for user access control, database management and enterprise-level software applications. PCs are also generally required at the next level down where the line management and HMI software resides.

However, at the machine level, which includes packaging equipment, material handling, machine controls, ID readers and vision systems, Windows-based PCs are often unnecessarily costly to install and complex to validate when compared to smart camera vision systems. And thanks to the latest generation of smart camera vision systems, deploying PCs in machine-level applications such as bar-code reading, text verification, mark quality assessment, label inspection and general machine vision applications is no longer necessary.

As a result, more and more manufacturers - as well as their equipment and materials suppliers - point to smart camera vision systems as key technology helping them to lower costs, improve supply chain security, and achieve product and package compliance in an increasingly regulatory and competitive environment.

PC-based vision systems versus smart camera vision systems

Generally, today’s vision systems are divided into two groups: PC-based vision systems and smart camera vision systems. Key differentiators between these two types of vision systems include architecture, cost and development environment. In this article, smart cameras are defined as modular DSP-based vision systems that are self-contained and don’t require the use of a PC, VME, PCI or similar architecture to run vision tools.

Centralised versus distributed processing

The primary architectural difference between PC-based vision systems and smart camera vision systems is one of centralised versus distributed processing. PC-based vision systems generally multiplex industrial cameras from a single processor in order to distribute vision at multiple points on the production line.

This centralised approach typically increases software complexity and integration costs to the point where the resulting system is not as easily scaled as a distributed smart camera architecture. PC-based vision systems also require more physical space on the machine, and can make centralising the HMI a challenge when multiple stations are deployed.

In contrast, smart camera vision systems combine low-cost distributed processing with high-speed networking to provide infinite scalability. Smart camera vision systems generally have one or two processors per camera, and because they can be easily linked together and managed as a system over a network, the overall costs and complexities of implementing distributed vision are dramatically reduced.

Today’s most advanced smart camera vision systems also offer compatibility with low-cost, touch-screen displays that can be easily deployed wherever factory floor operators need more control over their vision applications and better insight into what is happening on the production floor.

Cost of ownership

In terms of cost, PC-based vision systems typically have a much higher cost of ownership because they often require more IT department oversight and management to deal with service pack updates and other items like virus protection software. Because smart cameras function independently of the Windows operating system, they don’t require pharmaceutical manufacturers to repeatedly quarantine, test and deploy patches.

PC performance increases with each boost in processor speed, which makes new PC-based vision systems well suited to the most complex or mathematically intensive applications. However, because PC technology changes so rapidly, it’s not as easily replicated as standard off-the-shelf smart camera vision systems. In as little as one year after installation, for example, it’s much more difficult to source and configure a new PC with identical specifications than it would be to replicate a smart camera.

Similarly, replacing failed PCs or expanding the line to include more PCs frequently results in many different PC models on the line. The end result is that some PC-based machine vision systems may offer the most sophisticated vision tools, and provide the fastest performance, because they rely on the latest CPU architectures, while others may not because they rely on older PC technology.

In contrast, smart camera vision system technology is inherently much more stable over time. As a result, it’s much easier to find a commercial off-the-shelf replacement unit for many years after the initial installation. It’s also less costly to stock spares and to maintain consistent vision system performance across multiple production lines and inspection points.

Programmable versus configurable development environment

The development environment allows users to build (set up and program) vision applications to meet specific needs. While many PC-based systems have a programmable environment, most smart camera vision systems generally provide a configurable environment that’s easier to use, integrate and maintain.

Programmable PC systems are generally more costly and time consuming to integrate because they require more vision expertise and knowledge of low-level programming languages such as C++ or Visual Basic. Consequently, producers without in-house machine vision expertise in these low-level programming languages either incur the extra costs of contracting a specialist every time production requirements change or they pay for costly annual service and support arrangements with equipment suppliers.

This can be extremely expensive in today’s regulatory environment because the path to compliance remains unclear with current data formats and marking standards varying from country to country and from region to region. In this uncertain environment, producers implementing programmable PC-based systems have limited ability to quickly and efficiently adapt. Every time production changes require code rewrites and costly line revalidations, programmable systems put profit at risk.

In contrast, the most advanced smart camera vision systems generally require no programming and provide more user-friendly interfaces that make it easy for pharmaceutical manufacturers to cost-effectively bring more of the machine level integration for their packaging lines in house. Because they require no programming, these smart camera vision systems are much easier to adapt for compliance with emerging global standards as they evolve.

Consequently, producers implementing smart camera vision systems will have greater flexibility to quickly respond to new production requirements, and configurable software reduces the need for line revalidations because code rewrites are not required to accommodate production environment changes.

Turnkey versus best-of-breed serialisation solutions

In a sprint to compliance with product serialisation and authentication regulations, over the last few years many major pharmaceutical manufacturers have defined traceability requirements and initiated package serialisation pilot programs. Some have standardised on integrated solution providers that offer turnkey serialisation systems. Others are strategically leveraging multiple partners to design a best-of-breed serialisation system at each level of the enterprise.

Many manufacturers that have standardised on a single turnkey vendor in an effort to reduce integration time are pausing to reassess best tools and practices. After rethinking their initial strategies, many have found that relying on one vendor to supply all equipment and software needed for serialisation - from the machine and line levels up through the MES and enterprise levels - provides limited or no choice in component selection and technology choice.

Serialisation success depends on not only selecting the right ERP, MES and packaging line control software, but also choosing the right hardware for marking, coding and labelling, and especially for identification. Because companies with integrated serialisation solutions must offer a variety of products and services beyond vision hardware and software, they generally have limited machine vision offerings.

In contrast, companies strictly focused on identification and machine vision offer a broad range of smart camera vision systems in a variety of form factors, resolutions and processing speeds. These companies deliver the most accurate and reliable vision and ID tools for text verification, label inspection, bar-code reading and other general vision applications. In addition, companies with core competencies in machine vision and industrial ID generally provide both smart camera and PC-based vision system architectures for a price- and performance-optimised solution at each application level.

Conclusion

Low cost, ease of deployment and ease of maintenance remain the key attributes of smart camera vision systems, and over the last several years they have become more powerful. While PCs are required at the higher enterprise and production line levels, at the machine level smart camera vision systems can more cost-effectively address ID code reading, text verification, mark quality assessment, label inspection and general inspection applications.

Compared to PC-based systems, smart camera vision systems are easier to configure and validate. As a result, they are less costly to integrate and maintain. A distributed architecture allows manufacturers to more easily reduce scrap, rework and inventory problems by networking vision systems throughout production to catch defects at the source and potentially prevent errors altogether.

Smart camera vision systems intrinsically provide a more stable platform than Windows-based alternatives, without any service packs and patches to install. And they readily handle all future changes in regulations and standards without costly code rewrites.

With built-in ethernet, smart camera vision systems provide higher level computing systems access to plant floor data. Ethernet also links enterprise-level networks with production control and device networks, allows intelligent control devices to share information required for tasks such as automating production line changeovers and offers high-speed access to data generated by a broad range of plant floor devices for statistical process control.

By John Lewis, Market Development Manager, Cognex Corporation