Scanning represents the first step toward knowledge acquisition from paper and microfilm. What features can make key differences in the post processes?
You can spend $100 for a little desktop Visioneer (www.visioneer.com) scanner to as much as $1 million for a large fully loaded Scan-Optics (www.scanoptics.com) 9000. And there are all sorts of scanners with different specifications in between. All fundamentally do the same job-convert a piece of paper to a digital representation in a short time, ranging from approximately 6 seconds for the $100 scanner to one-third of a second for the $1 million one.
Clearly there are much larger differentiators than just speed.
The term high volume means different things to different people. To the office manager who usually scans 200 pages a day, high volume may be 2,000 sheets of paper. But to a service bureau converting 250,000 pages every day, high volume means two or more shifts, and the need to look at every possible labor-saving device.
Usually we think of more than 5,000 sheets a day, over a sustained time as high volume. At that rate, the $100 scanner cannot hope to compete. Apart from the scanning time of such models, consider the following:
- Autofeeders are non-existent-each sheet must be individually fed.
- If the image is on both sides, you must turn the page over and rescan it.
- If the page is wider than letter sized, you can't scan it.
- If the paper is damaged or stapled, you can have difficulty scanning it.
- If the paper is thicker than bond, the rollers won't feed it.
- If the paper is thin, you run the risk of jamming and you can't get into the mechanism to get small torn pieces of paper out.
- The background is white, which means that any holes, tears or corners get filled in with white, possible resulting in loss of information without knowing it.
- The rollers are not designed to be cleaned.
- The lamps cannot be changed.
- Paper skew cannot be controlled.
- Image quality is totally dependent on the original.
- Each image must be viewed for quality before release.
- You cannot prove the paper was scanned.
- You have to index the documents from the image.
So what happens if you spend a little more? The first thing that you usually get is faster scanning (approximately 3 to 4 seconds), an autofeeder and a flatbed. That allows for partially unattended scanning (if the paper is clean bond), while small, damaged or thick documents can be scanned on the flatbed. Of course, scanning on the flatbed is time-consuming compared to autofeeding. This type of scanner costs $500 to $3,000 depending on speed and flexibility. The $500 models tend to be flatbed only and are primarily oriented toward desktop publishing, with the capability of 600-dpi resolution or more (with a substantial speed penalty). Most of these type of scanners are mass produced in the Taiwan or Korea.
At the next level, manufacturers start to tackle the issue of image quality and scanners become more autofeeder-oriented. Image processing (IP) hardware is included, which reduces the need for manual QC and rescanning. The IP sharpens the images, removes backgrounds, adjusts for variations in the image and removes speckles. The autofeeders start to handle a wider variety of material with roller cleaning kits and replacement parts, but transitions from thick to thin documents tend to cause double-feeds. Some scanners in this class start to have the ability to scan both sides of a sheet, while others offer 11-in. wide scanning. Resolutions tend to be capped at 400 dpi. This takes us to about the $10,000 level and represents the largest volume of true document management scanners. These scanners, which are mostly manufactured by Japanese companies-Canon (www.usa.canon.com), Fujitsu (www.fujitsu.com), Panasonic (www.panasonic.com) and Ricoh (www.ricoh.com)-are typically capable of handling up to 3,000 sheets/day. Unique to this market are VisionShape (www.visionshape.com), a U.S. manufacturer that positions itself as supporting a higher volume of pages, and Xerox (www.xerox.com), which sells a 600-dpi capable scanner aimed primarily at the republishing market.
After this, flatbeds disappear and scanners support less than one second per image scan times and/or two-sided scanning. Autofeeders carry more documents (up to 500) with manual override, easy jam clearance and double-feed detection, which stops the scanner. Some level of automated indexing (OCR or barcode) is supported at the high end, endorsers prove that scanning has occurred, and the scanner background is often black. Black backgrounds allow for edge detection of the paper, automated deskewing, better thresholding and improved QC. Duty cycles for the scanners are 5,000 pages per day and up. Onsite service is available with guaranteed response times. Manufacturers within this realm are typically based in the United States and include BancTec (www.banctec.com), Bell & Howell (www.bhscanners.com), Kodak (www.kodak.com) and PhotoMatrix (www.photomatrix.com); Fujitsu, Panasonic and Ricoh have entered this market, as well. Unique in this space are German-based CGK with a desktop color scanner, Scan-Optics with a hardware based in-line OCR and image scanner, and Kliendiest (Augsburg, Germany) with a multi-pocket full-page scanner designed for remittances. This takes us to about the $50,000 level.
At this stage we move to the true high-volume scanners, which typically are floor-standing devices designed to scan 10,000 or more pages/day in a multi-shift environment. These scanners-which scan in sub-second times-are manufactured and sold by companies such as BancTec, ElectroCom/AEG (www.electrocomimaging.com), IBML (www.ibml.com), Kodak, NCS (www.ncs.com) and Scan-Optics. They contain multiple value-add functions designed to reduce the hidden costs of conversion-data preparation, image quality control and indexing. Each of these options adds to the price, so a scanner that may start at $100,000 or so can easily end up costing $200,000.
The value-add options include one or more of the following, depending on the configuration or type of scanner:
- a variety of autofeeders that can feed most types of documents including specialty documents (checks), multiple endorsers,
- vacuum-assisted transports to feed and control even the most damaged papers at high speed,
- jam detection,
- inline patch card, OCR, ICR, barcode and/or OMR recognition,
- multiple pockets to outsort documents or to save the user from stopping the scanner to empty the hopper,
- grayscale-based electronic deskew,
- foot pedals to allow manual overrides without needing to stop production,
- touch screens for ease of control,
- multiple zone image read-outs,
- color-based pattern drop-out and/or color image capture, concurrent microfilming,
- simplex (one-sided) or duplex (two-sided) scanning