Laser bar code technology was the first step toward reaching the goal of total supply chain visibility in which a product''s identity and location are tracked from production to sortation to arrival at the customer''s door. However, while laser-based systems are reliable, their accuracy is affected by a number of conditions, including label quality, label size, label location, and transport speed.
Laser bar code scanners also have difficulty reading two-dimensional codes omni-directionally. These symbols can store up to 3,000 bytes of data, and they are widely used in such major industries as electronics, pharmaceuticals, automotive, and package distribution.
The second major limitation of laser scanners is scan speed. Typical scanning speed is 500-1,000 scans per second, which means that either conveyor speed must be limited to 560 fpm or label size must be increased. Either option results in lower throughput and additional non-value-added costs.
CCD (Charge Coupled Device) and CMOS (Complementary Metal Oxide Semiconductor) camera-based scanners were developed to overcome these problems. Compared to laser readers, camera-based systems provide highly reliable and accurate reading. They can read bar codes at any angle from 45-90 degrees from the surface as well as bar codes placed up to several meters from the camera.
Using camera images and sophisticated vision algorithms, the system can reliably read bar codes omnidirectionally and under difficult conditions including blurred, partially occluded or damaged, and "noisy" bar codes due to surface material. The camera sends imaging data to a decoding device which then transfers the data to a host controller. Camera-based systems also provide multiside inspection and optical character recognition (OCR) in addition to basic bar code scanning.
Camera Basics
Camera scanners differ from laser scanners in a number of ways. Laser scanners move a laser beam over a bar code, and all white areas are reflected while black areas are absorbed. This returns an analog representation of the bar code that is then digitized and sent to a decoder for processing. Most laser products have a scan rate of 500-1,000 scans per second.
Camera systems can be classified as either linear CCD arrays or CCD/CMOS area arrays. Linear arrays work similarly to laser scanners except they use gray scale imaging at scan rates as high as 18,000 kHz. Also, linear imaging systems need to be mounted so that the bar code to be read has to be in constant movement through the imaging area. The image is built through consecutive scans of the linear CCD sensor of the moving bar code.
Area arrays take an image similar to that of a digital camera. They capture entire frames or images of a bar code at rates of up to 60 frames per second.
Regardless of the type of array used, camera systems recognize bar codes by translating the symbol into a gray scale image. Gray scale image data ranges from 0- 255, with 0 being black and 255 being absolute white. This data is sent to a processor which first processes the data to determine a region of interest (ROI), an area where possible bar code data might be located. This part of the process represents the highest overhead in the processor. Once an ROI has been determined, that portion of the image data is then processed to determine if bar code data is present. Any bar data present is decoded by proprietary software.
Camera systems have improved greatly in recent years. The first systems contained a 4k linear sensor mounted in a large, bulky enclosure. It required costly high-power integrated sodium lighting for illumination. The focusing mechanism was crude and slow, so these systems did not represent a significant improvement over laser sensors.
Second generation systems included up to 6k sensors with separate imaging and decoding. Just like today''s camera systems, image data was sent to the decoder via a gray scale interface. However, these systems still required large, bulky enclosures and sodium lighting for illumination.
Today''s camera systems include up to an 8k sensor with less costly LED illumination. The illumination source can be mounted in the same plane as the camera, providing a compact installation. Image data is sent to the decoder via a "camera link" interface using an LVDS (Low Voltage Differential Signal) RS-644. Fiber optic interfaces are also available.
Cameras vs Lasers
Camera-based systems are cost-effective in the fastest and most challenging inspection applications such as postal sorting. These systems consist of high-speed cameras, powerful image processing, high-power lighting, and optical systems that can focus and read in as little as 30 milliseconds. Applications are fast; therefore, system demands are high.
The main advantage of camera bar code scanners is speed - in both bar code resolving and object transport. They typically can read bar codes on packages traveling at 600 fpm and spaced as little as 2 in. apart. Camera systems can accommodate box dimensions from 1-36 in. on a side and can read bar codes on any or all six sides of the package. Typical read rates are 99.7%.
In many high-speed applications, the challenge is to find bar codes in any orientation on multisided packages. Here, scanner read points are set up so that all sides of a six-sided parcel can be read. A camera is placed on each side of the cube. The transport medium then moves the parcel through the tunnel created by the scanner points. The application requires large fields of view and high transport speeds. Therefore, these systems are typically linear array cameras rather than area arrays. In these applications, cameras provide higher read rates than laser scanners. Cameras also can read bar codes regardless of whether the packages are right side up, upside down, or rotated.
Another advantage of camera readers is the ability to decode low aspect ratio codes, allowing for the use of smaller labels. This translates into savings in label paper.
Also, as noted above, cameras are better able to read two-dimensional and damaged codes because of their more sophisticated image capture diagnostics. This results in fewer no reads and less operator intervention to complete the sortation process.
The main disadvantage of cameras is cost. Camera systems can cost two to three times more than laser systems. In addition, camera systems require lighting and focusing, which can add to the cost.
As sorting systems add functionality, image-processing requirements increase, as do sensor speed and lighting requirements. At higher speeds, the sensor has less time to acquire an image. This typically means that more light has to be used for the camera to produce quality images. Many sorting applications have used high-pressure sodium lamps because of their high output, but LEDs are finding increased use because newer versions have sufficient power for high-speed, high-resolution sorting applications. Combined with Fresnel lenses, LED lights produce a strip of light instead of lighting an area. By placing the LED stripe coplanar to the linear CCD array, the light is directed exactly where it needs to be.
Camera System Design Considerations
Factors to consider when designing a camera bar code scanning system include:
- Bar code/label size
- Transport speed
- Transport width
- Package size
- Package spacing
Bar code/label size -- The overall size of the bar code is usually not important in camera systems but the width of the bar code''s narrow element is. A certain DPI (dots/pixels per inch) needs to be maintained to decode a bar code. A certain number of pixels are needed per narrow element to properly resolve the bar code so it can be decoded. The camera''s mounting distance is adjusted to produce the correct DPI. Camera mounting distance affects other application parameters, including transport width. To gain DPI, the camera must be mounted closer to the transport, but this will limit coverage.
Transport Speed -- Transport speed affects the selection of the sensor for the application. A 6k sensor has a higher scan frequency, or clock rate, than an 8k sensor, but the smaller sensor also limits transport coverage.
Transport Width -- To meet the transport width requirement, a combination of lens and sensor size needs to be made. Different lenses are used for different sensors.
Package Size and Spacing -- Package spacing depends on several factors, including whether the application is front read or back read and transport speed.
In a front or back read application, package spacing should be as large as the tallest box to avoid shadowing. Therefore, if the largest box is 20 in., package spacing should be 20 in. There is a fixed time associated with focusing the camera to read the bar code so package spacing cannot be less than the minimum focus time.
Beyond Bar Codes -- The Future of Camera Systems
Until recently, camera-based sorting systems were used only in high-end, high-speed sorting applications where increased read rates justified the extra cost quickly. Now, midrange applications are moving to machine vision because of the need to read 2D codes or because of other features that are beyond the capabilities of laser scanning systems. New CMOS technology offers the potential for lower cost cameras with faster read times.
Today, camera systems are moving into a number of applications where their unique characteristics offer advantages, including:
- Optical Character Recognition (OCR)
- Optical Character Verification (OCV)
- Dimensioning
- Video Coding
Optical Character Recognition (OCR) is finding increased use in applications where a bar code is not available. For instance, OCR software can be used to actually "read" the address block of a package or envelope label. Presently this technology is being used primarily in postal applications.
Dimensioning -- Camera systems provide the ability to determine package dimensions, which can be used to determine the optimum loading of a truck or to charge customers for oversize packages. In operation, a light curtain first determines box height. Then, length and width are determined by the pixel size of the box.
Video Coding is used in applications where the bar code cannot be read and the image can be sent to an operator who manually enters the correct data. Video coding can be used in any application but it is used most often in parcel and postal applications. In operation, the region of interest is selected, and the camera zooms in on this area. The operator then enters the pertinent data (ZIP code, for example) via a keypad. The data is sent to a processor for use in sortation or other operations.
Conclusion
Camera-based bar code scanning systems offer a number of advantages over laser-based systems, including improved accuracy, faster throughput, and the ability to read damaged codes. They offer advantages in package sortation systems requiring high speed and low or no misreads.