Vision sensors can prove particularly useful for inspecting and tracking pharmaceutical products to improve product integrity and satisfy industry quality standards.

To meet ever-increasing industry regulations, pharmaceutical companies need to employ strict inspection and verification tools during product manufacturing and packaging processes. While simple sensors can be used for presence detection on the automated production line, to achieve accurate results in more sophisticated pharmaceutical applications—such as those requiring analysis of more than one product feature at a time—vision sensors can deliver a more comprehensive and cost-effective solution.

Determining whether a vision sensor is required for the application at hand involves several decision criteria. First, if the application requires sensing a larger area—such as a blister pack, a label, or a date code, rather than a single point, vision generally provides a more complete solution. Vision sensors are additionally helpful when more than one feature must be inspected on a single part. For example, while a photoelectric sensor may be able to determine vial fill level, a vision sensor can confirm that both the level and vial stopper alignment are correct. Some vision sensors also prove advantageous on high-speed production lines and for identifying color, translucency or other traits of the feature of interest.

(Click image for larger version.)
Along with these benefits, some vision sensors also have the ability to read, verify, and/or grade bar codes. Industry requirements have been established for pharmaceutical manufacturers to identify and track manufactured products, and bar codes offer a cost-effective method to implementing this traceability. However, unless the bar code is legible in the production process and throughout the supply chain, this tool is useless. Some vision sensors can both read bar code information—tracking products and identifying rejects on the production line—as well as grade bar codes to determine whether or not the code will be able to be identified further down the production line.

The following scenarios are a few examples of how the appropriate vision system can be used to efficiently and effectively solve common pharmaceutical applications.

Inspect for complete blister packs

Vision sensors serve multiple purposes when used for verifying blister pack contents and packaging integrity. Vision inspection tools allow users to set parameters for the inspection and specify which features of interest within the blister pack must be identified. This teaches the sensor to inspect the packs for tablet presence, verify that no tablets are broken, and alert operators of any blisters that are empty or contain foreign, unwanted debris.

To perform this task, the vision sensor is taught to identify a known "good" reference part—a correctly filled blister pack—from a "bad" part, rejecting all defective packaging and product. Once the inspection region and features of interest have been properly illuminated to create the highest contrast possible, the sensor acquires the image by a 2-dimensional CMOS or CCD imager chip. Next, the vision sensor compares this acquired image to the stored "good" part and notes any discrepancies between the two, thereby allowing users to quickly identify rejects and divert them from the production line.

(Click image for larger version.)
Color vision sensors take this technology to the next level with the ability to identify a specific color and simply differentiate between three or more colors. A special optical filter is placed over the imager chip that allows the sensor to distinguish between two colors that would appear identical to a grayscale sensor. These sensors may be used in applications where a tablet of a specific color must be identified, or they can be employed to inspect a group of tablets for correct color—detecting slight hue variations that could mean the difference between a good and bad product.

Accurate, efficient vial inspection

Another common use for color vision sensors is in vial packaging, in cases where the vials' cap color must be verified. Once all the vials are placed into a tray based upon the vials' color-coded lids, a color vision sensor "sees" the color pattern of each, allowing the sensor to identify each vial's cap color and verify that they are correctly placed.

To verify vial fill level as well as stopper seal and alignment, a high-speed vision sensor can deliver results rapidly, right on the production line. The sensor captures the image of each conveyed vial, then compares it to the image of a good part—a correctly filled and capped vial—that is taught to the sensor. This comparison allows the vision sensor to identify improperly filled or capped vials and reject them.

(Click image for larger version.)
Vials can also be inspected for overall quality, ensuring that they contain no chips or cracks before and after the automated liquid filling station. This inspection is performed using the vision sensor's Geometric Count tool that allows users to search the image obtained for the pattern of a good vial. Geometric Count is one of many common tools in a vision sensor's toolset. In this case, the vial's top rim is uniformly illuminated using an on-axis light and the pattern consists of edge outlines of the inside and outside diameters of the vial opening. If this edge is chipped, the Geometric Count tool recognizes that the edge is not continuous, and the vial is rejected.

Read, verify, and grade bar codes for product traceability

The vision-based Bar Code Reader (BCR) tool is particularly useful for identifying and tracking pharmaceutical products. Since bar codes are comprised of light and dark patterns, a vision sensor with the BCR tool can scan a bar code pattern as an image, then compare it to the taught good bar codes stored in the sensor. In this way, the sensor can determine if the bar code read is legitimate and, if so, "translate" the symbol into encoded information—generally a string of numbers or letters.

(Click image for larger version.)
In the pharmaceutical industry, vision sensors are used to read bar codes to identify and track batches of product on the production line. If a product's bar code pattern does not match up with what the sensor has been taught is acceptable—either because the product is incorrect or the bar code is unable to be read—the product is rejected. This ability also proves useful for grading bar codes, in order to ensure the symbols can be read by other devices throughout the production process and downstream in the supply chain. The vision sensor is programmed to assign a letter grade—A through F—to each passing symbol based upon a number of parameters specified in the IEC standard. For instance, if the device marking the bar codes runs low on ink, thereby producing symbols with low legibility, the vision sensor can report the low bar code quality grade and stop the process.

Simplify inserts and outserts verification

Along with confirming product integrity and quality control, vision sensors can verify that a product insert is included with each unit. The sensor is mounted above a pharmaceutical packaging line and checks each passing package for the information leaflet, using the Geometric Find vision tool to look for the insert within a defined search area. This tool allows the sensor to identify the insert despite its orientation, and any packages that do not contain the leaflet are diverted from the production line.

(Click image for larger version.)

A similar method is used to ensure that an outsert is included with each package. The vision sensor can be taught to let only bottles with an outsert attached to the cap pass the inspection. As each bottle passes on the production line, the vision sensor captures and analyzes an image of the bottle cap to verify the outsert's presence. This vision sensor capability, combined with the aforementioned tools for image acquisition and analysis, allows these sensors to deliver a comprehensive inspection and verification solution for a broad range of pharmaceutical industry applications.