Cisco-Eagle, the company that built the automated system for applying the stamps, was only able to achieve about 95% accuracy with the first barcode reader that it tried. This meant that the reader failed to read a barcode about every 20 seconds. Next Cisco-Eagle tried the Cognex DataMan® 500 barcode reader and discovered it rarely was unable to read a barcode. "Cognex customer support is remarkable," said Sam Chen, Senior Robotics and Software Engineer for Cisco-Eagle. "They worked with us continuously to make this application a success."
When Joe Jansen purchased the Goo Goo liquor store in January 1964, he decided to call it Goody Goody to save money by not having to buy a new sign. At this time, the store was in a 15 foot wide building with an inventory valued at $2,400. The company reached $1,000,000 in sales in 1969. Goody Goody opened a wholesale division in 1966, sold it to focus on its retail business in 1974, and started a new wholesale division in 1995. Today, Goody Goody Liquors is the number one volume retail store chain and its wholesale division services more than 600 hotels and restaurants in the Dallas area.
When Goody Goody Liquors wanted to automate the application of tax stamps, management began talking to Cisco-Eagle because of their experience in, as the company puts it: "getting the right amount of the right material to the right place at the right time in the right sequence at the right position in the right condition for the right cost in a safe manner." Cisco-Eagle was founded with the incorporation of Cisco Material Handling in 1953, followed by the founding of Eagle Material Handling in Tulsa, Oklahoma, in 1985. The companies merged in 1995. Cisco-Eagle provides solutions for the movement, storage, retrieval, control, and protection of materials and products throughout their manufacture, distribution, consumption and disposal.
Challenging machine vision application
Cisco-Eagle worked with Goody Goody Liquor to develop a system that automatically applies the tax stamps with 1-D barcodes and reads the barcode for verification and tracking. The liquor bottles are processed at high speeds on conveyor belts. An applicator automatically applies the stamp but the system must be able to handle the case where the applicator fails. Machine vision is required to identify bottles that do not have a stamp on them so they can be removed from the line. The image-based reader also captures the tax stamp number so it can be entered into a database along with the bottle number in case of an audit.
"We have tried various machine vision systems in the past for this application but have experienced reliability problems," Chen said. "One of the challenges is that the labels pass by the barcode reader very quickly. The bottles themselves are different with some being round and others being rectangular, which affects the way the barcodes are presented. The line processes many different liquor bottles, many of which provide a background that can be difficult to read against, e.g., very dark, very bright, or striped packaging and opaque or transparent liquid content. Finally, the line is used in a warehouse where the lighting is shifting constantly as lights are turned on and off and the degree of natural lighting changes with the time of day and weather."
Selection of barcode reader
Cisco-Eagle tried a number of different barcode readers. One barcode reader was only able to read the barcode against the widely varying bottle backgrounds about 95% of the time. This resulted in a failure to read about every 20 seconds which negated the main benefits of automation. Cisco-Eagle engineers tried another barcode reader but had difficulty integrating it with the Rockwell ControlLogix programmable logic controller (PLC) controlling the bottling line. They received very little direct help from the manufacturer of the barcode reader to resolve this critical issue.
"We decided to go with the Cognex DataMan 500 barcode reader because it has a larger depth of field, higher resolution, and decoding technology that made it possible to reliably read the barcodes in spite of the challenges of this application," Chen said. DataMan 500 is the first reader for the logistics industry to use liquid lens autofocus technology that maximizes the depth of field for greater reliability in applications where package position varies. DataMan 500 offers 1024 X 768 image resolution. DataMan 500 also offers advanced decoding that makes it possible to easily read badly printed codes, damaged codes, distorted codes, codes on a noisy background, extreme perspective, codes with low height, blurred codes, scratched codes, etc.
"The other barcode reader manufacturers gave us their manuals and technical support numbers and we had to figure out how to apply and integrate the new products largely on our own," Chen said. "On the other hand, Cognex taught us step by step how to use the product and helped us integrate their barcode reader into our system. When we had problems they came and worked with us until the application was successful." The Cognex barcode reader is configured by connecting it to a personal computer. A graphic user interface is used to set parameters such as aperture, gain, focus, communications parameters, etc.
Integration of barcode reader and applicator system
The DataMan 500 communicates using Rockwell's EtherNet/IP protocol and has a driver for Rockwell PLCs that simplifies the integration task. Once the driver is installed in the PLC, the camera shows up as a component. The PLC can then be easily programmed to issue any commands to the camera such as to capture an image with minimal programming. In this application, issuing the command to the barcode reader to capture an image, provide feedback on whether or not a barcode was detected, and send the barcode number was accomplished with only three lines of code.
Just before the bottles are put on the conveyor, an operator uses a hand scanner to read the barcode on the label automatically entering the number into a database. The operator then places the bottle on the line. A photo sensor detects the bottle as it begins to move on the line. The line can move at varying speeds and the PLC tracks the position of each bottle by tallying the encoder pulses it has detected from when the bottle passed the first photo sensor. As the bottle moves down the line, the tax stamp is applied to the bottle. Another photo sensor is mounted on the line just before the barcode reader and provides a signal to the PLC every time a bottle passes by on the conveyor.
The PLC sends a signal to the camera to begin capturing images. The barcode reader captures images until it finds and reads a barcode. Based on the time and travel that has elapsed since the bottle was loaded on the line, the PLC knows which particular bottle was read by the barcode reader and stores the tax stamp barcode in the same database row as it had previously stored the label ID. If the DataMan 500 cannot find a code then it sends a signal to the PLC. The PLC then triggers a reject mechanism to remove the bottle from the line at a downstream station.
"The reliability of the Cognex barcode reader is substantially higher than the other cameras we tried in this application," Chen said. "It rarely fails to detect the label even on bottles with the most confusing backgrounds and difficult lighting conditions. The barcode reader played a major role in the success of this application which saves a huge amount of time that was previously spent applying labels, manually scanning labels, and inputting the tax stamp information. The customer is very happy with the performance of the system and we have not received a single service call on the barcode reader."