Digital scanning refers to optical and electronic processes that capture and convert printed materials to digital format. Scanning is one component of a larger document imaging system that includes image-capture, storage, display, and retrieval capabilities. Document imaging systems typically differentiate between page imaging and text imaging, the most common types. Increasingly, however, the scanning industry has been leaning toward enhanced functionality, whereby single scanning systems are capable of scanning images, text documents, and even positive and negative film. Page scanners rely on bitmap images while text images rely on optical character recognition (OCR).
Bitmap images are arrays of horizontal and vertical dots or pixels that carry information about light and dark components of the image. A pixel in a simple black and white scanner carries one bit of information—whether the pixel is black-or-white. The number of available pixels or dots per inch determines the resolution of the image. The more dots per inch, the greater the resolution or level of visual sharpness of the image. These two critical concepts in digital scanning are called gray scale and resolution. Gray scale refers to the differentiated intensity of light and dark, while resolution refers to the level of detail available for display.
Scanning technology relies on photoelectric measures of light and dark to create bitmapped displays. The number of total photoelectric sensors and the amount of information contained about each pixel combine to create gray scale and resolution. The conversion of sensor data to digital format is obtained through the use of an analog-to-digital converter. The resulting digital information may be manipulated, stored, retrieved, or displayed on request as a digital mirror image of the original.
Scanner components typically include document input or reading devices, scan engines, and scanning software. Desktop digital scanners rely on either flatbed or sheet-fed operations to input hard-copy printed materials into digital form. Scanning engines incorporate cylinders and drums to record digital information, and frequently use charge-coupled devices (CCDs). Scanning software enables manipulation of both text and images. Using special scanning software, text recognition or optical character recognition (OCR) translates printed alphabetical symbols to digital words. These digital words may be edited or manipulated with a word processing software package.
Among the most significant, and long-awaited, developments in scanning technology is the ability to read handwritten data and transform it into digital format. Early uses of such technology included the scanning of addressed envelopes by postal departments for tracking mail, but the potential impact of this new technology is enormous. The latest versions improve on prototypes' poor resolution output by utilizing dual resonant lasers, which pivot along oscillating mirrors to capture optimal positioning and character recognition, producing superior digitization rates. Furthermore, the newest versions of these electro-optical systems use simple algebraic math for calculating the scan position and the angle of the laser beam's contact, rather than using the highly technical and specialized calculation functions employed by many other scanning systems, thereby resulting in relatively quick scan times.
Increasingly in the late 1990s, new developments in the scanning system market have focused less on breakthrough technology than on physical design, the aesthetics and utility of the user interface, and the integration of popular features and multi-imaging capabilities. At times, such developments range into the area of novelty, such as tiny scanning systems that can be worn on the hand or even fingers, or scanners the size of pens that can scan and store up to 3,000 pages of text. The latter, developed by the Swedish firm C Technologies AB, operates without a wire, and thus is highly portable, and can upload data to a standard personal computer. In the late 1990s, however, this technology was still in the experimental stages. In addition, many of the newer scanners come equipped with a wide range of software designed for sophisticated, often professional, editing, and with extensive archiving or storage capabilities, such as internal Zip drives.
Several considerations affect business use of scanners, including evaluation of needs, potential hazards, and maintenance activities. Advantages of scanning include reductions in both direct costs, such as materials required for manual reproduction of images and text, and indirect costs, such as the time an individual may spend, for example, retyping text rather than scanning it. Disadvantages include hardware investment and lack of industry standards. Purchasing a scanner involves assessing speed, resolution, gray scale, color, type, and special-features requirements.
In 1925, AT&T produced the wirephoto scanning service, and with it the first commercial image scanning system. Used by the news media, this service allowed photos taken around the world to be transmitted and printed in other newspapers. Additional experimentation and development resulted in the first color scanner patented by Alexander Murray and Richard Morse in 1937. Lacking digital processing and storage capabilities, however, scanning remained unchanged within the news media and undeveloped commercially. In the late 1960s, the National Aeronautics and Space Administration (NASA) spurred the use of image scanning in lunar explorations. Original lunar images were created and transmitted to Earth in analog (continuous) signals for later digitization. The Jet Propulsion Laboratory, under contract with NASA, developed a system to convert these images to digital form for computer processing. During the same period, analog facsimile scanners were developed for use in the business sector and, within ten years, were converted to digital facsimile scanners (fax). Medical uses of scanning increased and heralded the development of computerized tomography (CT scan) and magnetic resonance imaging (MRI) during the late 1960s. With the advent of personal computers in the early 1980s, scanning devices dropped in price and were actively marketed for use with home and business applications. A number of companies have incorporated scanning to improve insurance records and customer service, while others, such as Northwest Airlines, effectively use scanners in accounting and auditing. In 1993 almost 876,000 scanners were installed in the United States.
Document imaging systems facilitate the initial input, storage, retrieval, and display of digital images. Specialized image processing systems additionally provide for image enhancement, image restoration, image analysis, image compression, and image synthesis. Image enhancement activities include, for example, sharpening edges and adjusting contrast. Restoration activities, like photometric correction, adjust images to compensate for conversion errors. Image analysis may extract features or classify objects within an image, while image compression concerns itself with decreasing the overall size of a digital image file. Finally, image synthesis may incorporate activities like visualization and image mergers. Scanning software may incorporate features of an image processing system for user convenience and effectiveness.
Document imaging systems capture information based on full or partial pages of data or based on text or optical character recognition. Full or partial pages of information are converted to bitmap images using a digital process that creates software addresses for each small component of the image. OCR scanners map bitmaps to character symbols to convert text to digital format. In both cases, the beginning point of all document imaging systems is typically the initial input using a scanner.
Two of the most important concepts in digital scanning are resolution and gray scale. Resolution refers to the level of detail available in a printed image or the relative degree of visual sharpness. The number of pixels per inch or dots per inch (dpi) determines the quality of the image resolution. A common resolution for high-quality images is 300 dpi, though scanners have incorporated optical resolution as high as 5,600 dpi. Gray scale information for any pixel is a relative value of light intensity and is determined by the number of bits allowed for each pixel (a bit is a binary digit or the smallest element of the binary language). Frequent configurations include 4 bits per pixel (16 levels of gray scale) and 8 bits per pixel (256 levels of gray scale). Although gray scale creates better resolution, however, trade-offs to resolution include increased scan time and increased storage requirements. Many companies were competing for the greatest bit-depth technology in the late 1990s, though that trend is likely to slow for a very practical reason: the human eye is incapable of distinguishing any difference in the color palate beyond about 30 or 36 bits.
Gray scale is necessary to provide automatic scaling without loss or distortion. Scaling is the process of adding or removing pixels from an image. Because image resolution and image size are reciprocal functions, they are related by a scaling factor—scanner resolution multiplied by scanned size. Imaging continuous tone art and photographs requires gray scaling to accommodate shades. While print media represent shades with different sizes of dots, pixels are all the same size and must be manipulated by controlling the size or configuration of groups of pixels. Two methods (dithering and true gray-scaling) simulate shades. Black and white images may be converted to gray scale using a process called dithering. Dithering creates a simulated number of gray tones using geometric groupings of pixels that form patterns. These patterns represent shades of gray. Dithered images are often grainy and poor. True gray scaling, on the other hand, uses pixels that contain gray scale information. These pixels are grouped into symmetric patterns.
Scanners reflect light onto a printed page to illuminate light and dark areas of the page. These light and dark areas are recorded to a logical grid within the computer. Using a charge-coupled device (CCD), scanners record information by accumulating a charge proportional to the light intensity in a solid-state array of wells. Scanning and recording cylinders preserve photoelectric charges. The resolution at which an image can be scanned depends on the number of light sensors or CCDs in the scanner. Functionally, a CCD breaks up the scanned image into thousands of pixels. Each CCD photoreceptor cell converts light or dark into electrical voltage proportional to the light intensity. Exporting these voltages creates a bitmapped image. Raster scanning, line by line from top to bottom and left to right, yields a bitmap image.
A bitmap treats an image or document as a rectangular array of pixels by using a binary digital technique to represent the black-and-white pixels. Black pixels, represented by Is, and white pixels, represented by Os, are mapped to a grid to represent the light and dark areas of an image or document. When pixels hold information about a scale of light or dark (as opposed to simply black or white), they are consider to have gray scale definitions.
Digital scanners capture images (pictures and text) and convert them to computer files. These computer files represent the Os and l s of the binary language that the computer understands. Image scanners identify a picture as thousands and thousands of individual elements. These individual elements are known as pixels or pels and vary in density and pattern to accurately reproduce a graphic image. A picture element, or pel, is used when each element contains only black or white elements, while a pixel is used when the element contains intermediate shades of gray. Pixels are the smallest element of a display surface that hold information about color or light intensity. Bitmaps are the mapped pixel location and intensity necessary to recreate the original document or image.
There are three main subsystems to a digital scanner: the document feed, the scan engine, and the scanning software. First, the document feed system provides a means by which the printed material is entered into the computer. The scan engine consists of a light source, such as mirrors and lenses, a light intensity sensor, and recording medium. Finally, the scan control system is typically a software program that manages and directs the scanning process including resolution and gray scale detection.
Document feed systems ensure that paper documents or images enter the scanning device for digitization. Four scanner designs are available with different document feed formats. The flatbed scanner (or full-page scanner) resembles a photocopier with a flat glass area on which to lay documents. Almost any document type including books, heavy card stock, paste-ups, or other materials may be scanned on a flatbed scanner. By comparison, the document feed scanner can only handle single sheets and operates like a fax machine. While document feed scanners cannot scan a book, they can handle multiple sheets of paper automatically. Contemporary scanners frequently offer both flatbed and document feed in the same design. A third type of document feed system, the overhead scanner, is more specialized and used for three-dimensional objects. The scanning cylinder is usually encased above the scan bed and the light source points downward and is reflected upwards. The fourth, and increasingly popular, document feed system is the handheld scanner. Handheld scanners provide inexpensive scanning capabilities for small digitization activities. Most cover a four-inch swath of the document.
Scanning engines are the nuts and bolts of the digital scanner. Consisting of a light source (a moveable or fixed path of mirrors and lenses) and a light intensity sensor (a charge-coupled device), a document is scanned line by line. The light source, often fluorescent lamps, illuminates the document, and the reflected light is focused on a CCD by a mirror or prism. The resolution at which an image can be scanned depends on the number of CCDs in the scanner. For example, a 300-sensor-per-inch scanner can provide an image at 300 dots per inch. When the scan is in progress, either the sensors themselves or the document move at a fixed rate. As these sensors or CCDs are exposed to light, they generate a charge related to each pixel's level of light intensity or gray scale. One line at time, an image is produced that consists of an array of horizontal and vertical dots with varying intensity of light. An analog-to-digital converter generates digital information from the CCD's continuous analog signal. This information contains data about individual pixel elements and gray scale. The scan engine relays this information to the scanning software for processing, displaying, filing, or printing.
Scanning software frequently incorporates interactivity with the user. In addition to determining how the image is scanned, scanning software manipulates files, scales images, edits, rotates, and performs a wide variety of other functions including image enhancement and alteration. Scanning parameters such as page contrast, gray scale, thresholds, area dimensions, scaling, and resolution are all set up using scanning software, and are normally set according to the preferences of the user. In addition to the scanning set-up parameters, scanning software may also include programs to manipulate, edit, and save images. Specialized software is available to convert file formats for import or export. One special type of scanning software, optical character recognition (OCR), is specifically designed for use with textual materials.
OCR breaks down a bitmapped image into smaller bitmaps of individual character cells. Assuming that each character is unbroken and surrounded by space, OCR scanning software identifies text characters using pattern and feature recognition, and saves them as individual letters and words. In pattern recognition, a preexisting library of symbols is compared to the bitmapped character. The closest match determines the character code. In feature recognition, curves and lines and their relationships are derived from a sample character. Again, the closest match determines the character code. Errors occur when incomplete or unknown matches are encountered. Error rates in OCR range from 1 to 5 percent.
The simplest and most popular application of OCR is as a replacement for keyboard entry. In high-volume fields, such as law and business, OCR scanners speed document entry appreciably. Forms may also be used with OCR to capture questions and responses more accurately. OCR is a processor-intensive function best suited to high-end work stations. With the advent of inexpensive, powerful microcomputers, OCR applications are increasingly effective and available. Particularly as more paper texts are transferred into World-Wide-Web-readable HTML format, OCR applications play a vital role in the dissemination of information.
Color scanners require detecting, processing, and storing three pixels to accommodate red, blue, and, green colors within each range of gray scale. A four-bit color scanner provides only 16 colors. Typical color scan engines use an illumination system of fluorescent red, blue, and green lamps and filters. Balance and adjustment tables are provided by the computer and may be updated and altered using scanning software. Color scanners are processor-intensive and, while decreasing in price, remain somewhat specialized and expensive. For that reason, companies need to seriously consider their functional needs when investing in a color printer; a 24-bit scanner, for example, can reproduce 16.7 million colors, which can be quite useful for printers who produce glossy magazines, but would most likely be a waste in a firm that merely needs to reproduce simple bar graphs and pie charts.
Evaluating the potential and practical use of a scanner includes evaluation of needs, consideration of environmental hazards, and discussion of maintenance activities. Needs evaluation includes performance goals, user access, and identification of materials to be scanned. Environmental hazards such as temperature, humidity, dust, static electricity, and power supply present critical issues in the location and use of digital scanners. Even ensuring that the scanner rests on a level surface can be crucial, as the system can fall out of calibration. Cleaning and maintaining the scanning and recording cylinders, lenses, and mechanical components are additional factors in evaluating the potential use of scanners in the workplace.
Considerations in the purchase of a scanner include speed, resolution, gray scale, color, type, and special features. Speed refers to the scan speed of both black-and-white and gray scale images. Resolution refers to the optical resolution range given in dots per inch (dpi). The number of detectable gray levels and color capabilities impact both cost and capability. Many scanners now incorporate both flatbed and sheet feed types. Special features include paper size, supported printers, documentation and support, and image-editing options. The two initial steps in selecting a scanner are: (1) determining the type of scanning to be done (OCR versus image) and (2) determining the best scanner system. Price range and compatibility with current computer resources are additional factors to consider when selecting a scanner.
Image processing offers many advantages, although not all business will benefit all the time. Direct cost savings are available when scanning systems free storage space and permit reductions in workforce. These savings may be offset, however, by higher skill levels required for existing personnel. Fast retrieval, concurrent access, processing and distribution control, and reductions in lost documents can all contribute to improved productivity and competitive advantage. Nevertheless, initial costs for large-scale scanning operations can be expensive. Additionally, because image processing and scanning applications are relatively new applications, few experts or reputable vendors may exist. Incompatibility with current computerized resources and a lack of industry-wide standards may also create problems in installation, exchange, and use of digitized images. Companies in the credit industry such as American Express have capitalized on the advantages of scanning and minimized the disadvantages. By effectively using scanning technology, American Express Co. improved the aesthetic quality of the billing statement, reduced mailing time, reduced funding costs, and reduced document entry errors. Likewise, British Airways has improved cabin crew services by using scanning to facilitate the creation and entry of the voyage report. For these and other companies, image scanning and processing is a powerful tool in the management of critical information.
The 1990s were a good time for product sales and shipments, though not necessarily for firms. As prices have continually been forced downward, increased sales have often failed to keep pace with production and shipment costs, forcing many companies into bankruptcy or the sale of their product lines. Some major players in the late 1990s included Hewlett Packard, Microtek, Epson, and Umax Technologies. Industry analysts predict sustained growth in shipments well into the next decade, reaching about 38.9 million scanners, mostly flatbed, by 2003, according to International Data Corp.; this figure would be roughly three times the total shipments in 1998. However, by that time, vendors will scarcely be able to rely on large volume sales, as the market will be flooded with inexpensive flatbed scanners. As more consumers become familiar with scanning technology, firms will find their greatest hopes for achieving economies of scale in product innovation and upgrades.
[ Tona Henderson ]
Baxes, Gregory A. Digital Image Processing: Principles and Applications. New York: John Wiley & Sons, 1994.
Gann, Robert G. Desktop Scanning: Image Quality Evaluation. Englewood Cliffs, NJ: Prentice Hall, 1998.
Green, William B. Digital Image Processing: A Systems Approach. New York: Van Nostrand Reinhold, 1989.
Pepper, Jon. "Scanners Reborn." Informationweek, 26 October 1998, 97-100.
Richter, Uwe. "Scanning the Chaos." Novum, December, 1998, 23-7.
Szabados, David. "The Next Step in the Scanner Industry? It's Not More Bit Depth and Resolution.… Computer Technology Review, April 1998, 24-26.
Toth, Debora. "Desktop Scanners Gain Wide Acceptance" Graphic Arts Monthly, January 1999, 66-70.
Wetzler, Fred U. Desktop Image Scanners and Scanning. Silver Spring, MD: Association for Information and Image Management, 1989.