Feature: Portable Scanning Options

Lugging a scanner or copier to a client site or other location isn't exactly convenient. So I was curious: Is it possible to use a digital camera to take a picture of a printed page, then convert that image into an editable and/or printable document?

The question arose after I wrote about handheld scanners in a recent newsletter, in response to an e-mail query from Gregory Bilyeu, of Mount Vernon, Washington.

"I'm an attorney who spends time examining documents in the law library," Bilyeu wrote. "In this high-tech day, I'm still fishing for quarters for the copy machines. I've looked high and low for a handheld scanner that would work on books and other bound documents. An article on handheld scanners, both self-operating and those that must be attached to the notebook, would be very helpful."

In response to Bilyeu's e-mail, I searched the Internet for handheld portable scanners--the kind that you would sweep across a page in a bound document. My search turned up only pen-style hand scanners capable of digitizing a few lines at a time, which I wrote about in a follow-up article.

Try a Digital Camera

After that newsletter appeared, I received several e-mails from readers saying that they put their digital cameras to use as ersatz handheld scanners. For example, one reader wrote: "A digital camera works great for this. Just get one that has good close-focus capabilities and a good lens. Use a small tripod or fabricate a foldable stand. Try to get enough light without flash. ... In most cases the image will give you better output than those funky copy machines."

Intrigued, I ran this idea past Dave Johnson, who writes PC World's Digital Focus newsletter on digital photography. Sound interesting? You can sign up to get Dave's newsletter e-mailed to you once a week.

"It's possible [to convert digital camera images of printed pages into editable documents], but it's hard to get good results," Dave responded. "The problem is getting sufficient resolution and even lighting out of a digital camera. You absolutely need to mount the camera on a tripod over the paper to be scanned and use very bright, daylight-like lighting."

"If you get glare from a flash, you won't get good OCR results," Dave added. "A macro setting [an extreme close-up] usually works best, but that will depend upon the camera's lens characteristics--the key is to capture the full width of the page at the camera's top resolution to get about 300 dpi, and that may or may not be in the macro range of the lens when you position the camera at the right height."

An Easier Way?

That seems like a lot of work the average person wouldn't bother doing. But I decided to experiment using tools that the typical person might actually have.

I took pictures of book pages using my digital camera, the 3.2-megapixel Minolta Dimage XI. The Dimage XI can capture images as close as about 9 inches. Other digital cameras, such as the Canon PowerShot A80, have a macro setting that enables picture taking as close as 2 inches.

I didn't use a tripod, because I figured most folks wouldn't have one stashed in their briefcases. I snapped high-resolution images of pages from Bartlett's Familiar Quotations and other reference works under bright natural lighting conditions. Then I transferred the JPEG files from my camera to my notebook and opened them in ScanSoft's OmniPage Pro 14--a popular optical character recognition program. OmniPage Pro then converted the image files into Microsoft Word documents. For more details about OmniPage Pro 14 ($150) and OmniPage Pro 14 Office ($600), go to the ScanSoft site.

No Dice

Getting the full page into the camera frame was challenging. For example, when a book is left open, its pages don't lie perfectly flat. At least some of the text ended up being photographed at an angle, which made accurate OCR conversion difficult.

As good as OmniPage Pro is at most jobs, it could only do so much with my JPEGs. About 25 percent to 30 percent of each page I photographed was converted accurately into text. The bottom line: In my experience, photographing bound pages and running them through OCR software was a lot of work for very little result.

A Better Alternative

Dave Johnson tipped me off to a new product, the DocuPen ($200), from Planon System Solutions. According to the company, this pen-shaped handheld scanner can capture the full width of a page at once and lets you scan a page at a time. (Most pen-style scanners can only capture a few words at a time.)

The DocuPen scans text and graphics up to 200 dots per inch into its 2MB of flash memory and transfers files to a Windows PC via USB or serial connection. As of this writing, neither Dave nor I have evaluated the DocuPen. But products purchased from the Planon Web site can be returned within 30 days for a full refund, so you may as well give it a shot.

Before you run off to the library with your digital camera or DocuPen in hand, though, keep in mind that many published materials are protected by copyright. The Stanford University Library site provides a good primer on copyright law and fair use.