AIA: Machine Vision Helps Bring New Medical Device to Market
Actimed Laboratories Inc. (Burlington NJ) manufactures cholesterol home-screening kits. The kits allow individuals to check their cholesterol levels without the need for laboratory analysis of blood samples.
In developing these at-home kits, Actimed ventured into new technological territory. But after nearly 5 years of research and development, the company received Food and Drug Administration clearance to market the product. A user simply pricks a finger with a tiny lancet and places a drop of blood into a small sample well at one end of the device. The blood is automatically absorbed into a series of four tiny pads where it reacts with biochemical compounds. When the reaction process is complete, the device produces a cholesterol reading in a capillary channel running along its length.
Engineering a high-speed production process for the device's numerous lightweight components required exacting tolerances and control over several assembly operations. Actimed uses custom thermoforming equipment to produce the thin plastic device body in webs of six units across. Then, the reagent pads are placed within the small well of each device.
To meet the stringent quality requirements, the company must also verify zero defects through automated inspections of each device. From the start, it was evident that this would require powerful pattern recognition capabilities available only with machine vision technology.
To help it engineer the most effective machine vision system, the company turned to Xyntek Inc. (Yardley, PA), a provider of turnkey machine vision systems for pharmaceutical and medical device manufacturing. Ultimately, the technical demands and programming requirements of the application led to the Checkpoint 800 machine vision system manufactured by Cognex Corp. (Natick, MA).
From start to finish, Xyntek completed the application development phase of the Actimed project in just over 6 weeks.
In a battery of preliminary tests, Actimed had validated the accuracy and reliability of the cholesterol-screening device. One of the critical factors in its performance proved to be the placement of the tiny absorbent reagent pads in the well. The pads and the taped strip that keeps them in place must be located precisely for the device to work correctly. If a pad is hiked up on the side of the well or pulled away from the back, the device cannot take an accurate reading.
Using the image recognition and analysis capabilities of the Checkpoint 800, engineers from Xyntek and Actimed designed a turnkey inspection system to measure these tolerances in the X, Y and Z axes.
The vision system comprises processing hardware, vision tool software, cameras and a monitor. The automated inspection system also employs structured visible lasers to project reference lines to triangulate the key Z dimensions, and high-frequency lighting for general illumination.
The inspection cycle takes approximately 90 milliseconds per device and about a half-second for an entire row of six.
The unique nature of the Actimed device imposes some especially tough demands on the machine vision system. Using the built-in capabilities of the system, Xyntek developed application code to handle them in a number of ways.
First, there's the inspection process. Because it is thin and flexible, the web has a degree of latitude as it moves along the index conveyor into the inspection point. Therefore, it is necessary to determine exact orientation before inspecting it.
Using a search tool, the application code finds the two fiducial marks stamped onto the front of each device. It then performs a normalized correlation to match what it sees against a model that was trained during development. From that, it can determine the rotation and placement of the web.
The system proceeds with the actual inspection, using its edge detection tool to find the red-laser lines projected onto the device. By correlating the various X-Y-Z coordinates provided by this measurement, it determines the placement of the absorbent pads in the wells.
Before each production run, the operator calibrates the system to make sure that it can take the measurements accurately. During calibration procedures, the Checkpoint 800 examines each of the six inspection frames through two references: a fixtured reference frame for the camera and laser support structure, and a "floating reference" frame for the web sheet. There is also an error-checking procedure programmed into the calibration process. If the vision system sees that the alignment or skew of a laser is not correct, it will alert the operator.
Before the start of a run, a synthetic search tool performs a binary black-and-white search to find fiducial marks on a machined calibration plate. It then obtains a measurement of the exact pixel distance between the fiducial marks. Using that measurement, the inspection is calibrated to the correct scale factors by converting pixel units to mil units. It also calibrates for camera skew and rotation.
For more information on machine vision, call 877-264-6391 or visit www.cognex.com.