Vision for Tracking

Globalization continues to shape the economies of various regions. As more Western manufacturers move East in search of low-cost labor and other resources, Asia's regulatory structures are also evolving to enforce greater accountability in the supply chain. This applies not only for goods that are produced for export to first-world countries, but also increasingly for product that are meant for domestic consumption.

For instance, in India, the government has announced that all pharmaceutical export shipments must have a track & trace system, with it being mandatory for exporters of pharmaceutical products to attach bar codes on their export products. This requirement is to facilitate the tracking and tracing of their products in accordance to GS1 global standards.

It is necessary for GS1 Global Trade Item Numbers (GTIN) to be printed on primary packaging (e.g. blister strips, vials and bottles) from July 1 next year. While data matrix code is the preferred carrier, EAN/ UPC, GS1-128 and GS1 Databar are also allowed. For secondary packaging, 2D or 1D bar codes are required with effect from January 1, 2012.

Meanwhile, China is taking active steps to combat its domestic counterfeit drug problem. The country's State Food and Drug Administration (SFDA) announced a supplementary notice in the first quarter of 2011 to promote the electronic supervision of essential drugs. All imported bid-winning essential drugs are required to be entered into the network and be coded accordingly. The management of printing/pasting electronic supervision codes have to be standardized on some packages.

Mandatory serialization on sellable pharmaceutical product units is required by December 2011, and all products that are manufactured in China must have GS1 bar code labels. Imported products may have Health Industry Business Communications Council (HIBCC) bar codes for now, but will soon have to be mandated to the GS1 standard.

South Korea also has a significant drug counterfeit problem. In 2010, it was revealed that US$5.58 billion worth of counterfeit goods had been seized in the country since 2006, and these included a substantial quantity of fake medicines, according to The Partnership for Safe Medicines.Org. Hence, the country is enforcing the requirement for all drugs to have a GS1 barcode (GTIN- 13, GTIN14 or GS1-128). Specified drugs must have GS1-128 codes with lot number and expiry date from 2012 and special drugs from 2013.

Food scares
As well as the pharma sector, the food and beverage (F&B) industry in Asia is also undergoing changes in its regulatory space. In 2010, it was reported that Chinese authorities closed down a dairy producer and arrested three of the latter's executives after tests revealed that the company was producing melamine-contaminated milk powder. This is the same chemical that created one of China's worst scandals in food safety in 2008.

The arrests indicate that China's regulatory authorities are monitoring the dairy market more aggressively after six children died and more than 300,000 became ill after consuming melamine-tainted milk in the 2008 scandal. China only allows a little over a milligram of melamine to be present in every pound of food product. Beyond this limit, consumers especially children, run the risk of health problems such as kidney stones.

On June 2, 2011, Taiwan's cabinet announced that five types of food would require safety certificates in order for them to obtain clearance for export. These are: sports drinks, juices, teas, syrups and jams, and tablets and powders.

The move was in response to a widespread food contamination scandal where toxic plasticizers and clouding agents were discovered in food products in an attempt enhance the latter's visual appeal. It is possible that the affected products has been in circulation within the domestic market for over three decades, and had already had a detrimental health effect.

According to its Premier Wu Den-Yih, the Taiwanese government has increased its efforts to investigate the sources of the chemicals, destroy tainted products, revamp the necessary laws and establish a food traceability system.

Information integrity
The above are just some examples in Asia where there is an increasing requirement for manufacturers to raise their standards. This applies to quality assurance (QA) in production, as well as supply chain traceability.

In the F&B and pharmaceuticals industries, lot codes, manufacturing and expiry dates serve as a means of tracking legitimate products in the supply chain, while helping to prevent counterfeit drugs from entering the distribution channels. This information also informs the customer about the shelf life of products and helps to prevent the accidental consumption of products that have already passed their expiration dates.

This means that manufacturing plants which are still incorporating human vision for QA compliance processes, need to start considering adopting automated vision technology. Why? Because, given the limitations of human vision, human operators that are deployed along a production line can only check for the presence of the required information on packaging.

But in the face of high production speeds, subjectivity in operator judgement and fatigue are factors that make it nearly impossible to identify production errors at a 100 percent success rate. Furthermore, the operators are not able to easily identify incorrectly printed dates nor read bar codes.

Automated vision systems can play a major role in this space by ensuring that packaging information such as lot codes, expiry dates are correct and are readable during production and throughout the supply chain, and that the product packaging complies with consumer safety standards and is not damaged in any way.

Seeing with vision
With smart camera vision systems, manufacturers can raise the quality of their product and packaging at nearly every stage of the production line. Applications include checking each container's orientation, verifying code quality, and ensuring label accuracy at printing and marking stations.

ID readers are used to decode bar codes, and verifiers are applied to assess the quality of bar codes on labels or marked on containers. Item-level traceability is also achievable by reading 2D bar codes carrying unique numbering schemes.

These technologies help manufacturers to confirm that the label and product match. They also help and track lots, batches, and individual items through the supply chain to reduce the cost of recalls due to product defects or contamination.

Another benefit is that using machine vision in guidance, inspection, gauging and identification eliminates the need to physically handle the products. This can improve product safety by minimizing the risk of contamination due to human contact. It also reduces the chance that workers will suffer from repetitive strain injuries, and any liability that may go with it.

Advanced systems are also capable of addressing “blind spots”, meaning that the reading of bar codes can still be performed regardless of the angle in which the container (e.g. tins and cans) presents itself to the cameras (i.e. can be in any 360-degree orientation).

In other instances, the labels on packaging such as jars and bottles could be missing or be misaligned. There could also be situations where the wrong label has been applied to the packaging, e.g. chilli sauce label being applied to a tomato ketchup bottle and vice versa. Left unchecked and undetected, such production errors could result in embarrassment and a loss of reputation for the manufacturer.

In the case of drug packaging, such errors could have farreaching implications on consumer health. Consumption of an incorrect drug or dosage can result in massive product recalls and expensive litigation.

Another dimension
Most machine vision systems are integrated into production lines in the form of fixed-mount sensors. They are used for identifying parts that are handled and moved automatically by conveyor, indexer or robot. In operation, this type of reader is mounted in a fixed position where the mark can repeatedly be placed in front of the reader in either continuous or indexed motion. They can often be configured with either an integrated light source or an external light source, as required by the application.

To provide maximum flexibility, advanced ID code readers are also available as handheld devices. These are preferred in environments where part handling is not automated or where parts vary greatly in size. This method can be applied across various industries (especially in the logistics sector) where a product often needs to be checked in several locations before it reaches its final destination.

While both 1D and 2D readers are used, the growing trend is for implementing readers that have 2D capabilities. The main reason is that the 2D readers are future proofed, meaning that they have the capability to read both 1D and 2D codes, whereas 1D laser scanners can only read 1D barcodes.

In addition, 1D laser scanners suffer from the need to have 80 percent contrast between the foreground and the background whereas 2D code readers can read down to levels of 20 percent contrast (and sometimes less), as well as being unaffected by code rotation.

Many manufacturers are faced with the prospect of maximising available space on the production line while adopting technology that can perform the job effectively. Understanding these needs, vendors have introduced advanced readers that represent a breakthrough in combining high-level code performance (up to 45 reads per second), ease of use, and an extremely small size. Lighting, camera, processor, and communications are all integrated into an industrial housing, making the fixed-mount readers suitable for demanding applications.

Marked for ID
ID code readers have an important role to play in industries such as automotives and electronics manufacturing – for track & trace applications.

Automotives manufacturers aim to achieve defect free products as dealer repair charge backs can be costly. This is especially true for the safety aspects of critical parts where there are warranty costs involved as well as recall liabilities. Enforcing part identification while the product is still being made helps to enable flexible manufacturing and traceability for product genealogy.

In a similar fashion, vision systems are also applied in the electronics industry for inspecting integrated circuit (IC) marks consisting of alphanumeric characters. Machine vision can also be used to read 2D data matrix codes that are directly marked on wafers, IC components, printed circuit boards or other parts and assemblies for the purpose of identification and traceability.

In addition to recognizing and verifying characters and reading bar codes for identification, machine vision can also examine patterns to identify items based on color, shape, size or assembly. For example, color tools detect and measure color features in a wide range of applications and can be used to find gold pads, verify resistor markings or verify LED color quality. Color search and histogram tools are easy to train and provide a powerful alternative to grayscale.

Furthermore, with the aid of advanced software, degraded, reflective or low-contrast marks from die backsides, substrates, encapsulated packages and PCBs can be reliably read, making full wafer to die to package traceability a reality.

Direct part marking and corresponding machine vision technology have been evolving to meet consumer demands and industry regulations. Once the code is marked on the particular part or product, it is of little use unless it can be accurately read. This is where machine vision takes up the reins and ensures that full product traceability is achieved.

In a typical manufacturing scenario, the marked part is passed in front of a vision camera/sensor. Specialized image pre-processing and identification algorithms capture and process an image of the data matrix code. By adopting this technology, code reading performance is unaffected by low contrast or poorly formed codes – which can result from marking issues or general wear and tear of the product. Besides reading data that is stored within the code, the sensors can also give feedback on production processes that pertain to the quality of the specific marking. Perfecting the quality of the codes helps to eliminate waste and leads to improved manufacturing efficiency and a reduction of operating costs. Direct part mark verification (DPMV) allows manufacturers to improve the marking process, increase read rates, and cut down the cost of rejecting parts that are a result of unreadable codes. It also helps to ensure that parts do not lose their lifetime identity. Certain companies adopt DPMV with the aim of attaining process control, while others want to achieve contract compliance.

From a process control point of view, ensuring the quality of the mark that is made on the part is necessary for part tracking. This enables manufacturers to identify the root cause of poorly formed codes by using universal standards, and allows them to configure marking equipment for better results. By adopting these procedures, poor quality codes can be prevented from entering the production line and ultimately the supply chain.

Looking at it from a contract compliance point of view, DPMV can help suppliers to ensure that marked codes are in line with specific requirements in the contract. Suppliers of parts can improve their own ability to read codes, while helping improve data capture, part lifecycle management, and logistics activities.

Didier Lacroix is Senior VP, International Sales & Services, Cognex (www.cognex.com).

---

Validating Vision

In line with advancements in technology, certain automated vision systems have incorporated special programs that allow manufacturers to test and verify that their systems are working correctly. This feature allows users to have greater confidence in the validity of their vision inspection results. Users can confirm program changes by quickly checking against the “baseline” configuration.

The test application/program provides ease-of-use to system integrators and original equipment manufacturers (OEMs). It allows them to demonstrate that the vision system meets acceptance criteria when the system is installed at the end user's facility.

Machine operators are able to minimize poor performance by ensuring that the vision hardware is performing properly. By running the test program, quality managers are able to identify “borderline” inspections and refine pass/fail criteria.

Consistent vision system performance is dependent on maintaining consistent image quality. Vision programs may contain many parameters during set up and the test program assigns range limits to these parameters. When the test application executes, confirmation that the current setup parameters are within the expected limits will be given.

To verify that the system is performing image inspection correctly, the program runs a validation test against an image database. The library of images comprises a collection of “good” and “bad” parts. The reference images of bad parts can also be graded based on specific defects that are found on each part. When the test application is run, each image is evaluated by the vision program. If the system is running properly, the results from the vision program should match the “expected” results.