HIS480 logo

PROCESSING IMAGES

Copyright 2006, 2011 by Dr. Jim Jones
of West Chester University


Go to the HIS480 Syllabus or Assignments.

So far, we have only considered written sources consisting of letters, numbers and punctuation marks. If pictures are in fact "worth a thousand words," then computers (attached to scannes) enable us to extract additional meaning from photographs, paintings, maps, and advertisements. All of them can all serve as sources of historical facts, as well as illustrations to enhance oral or written presentations.

Scanners look like part of a photocopy machine and they attach to your computer. They convert a two-dimensional image (i.e. a flat image) into a sequence of numbers, and then use software (plus some guidance from the user) to save them as files or convert them into something else -- like a wordprocessing file that can be edited.

The scanner operates by moving a tiny camera back and forth across picture from one end of the glass plate to another. The number of dots-per-inch you specify determines how many lines the camera follows and how many dots it records along each line. It also determines the size of the file that you'll need to save or upload, and that's important because bigger files require more time and space. Graphics files are MUCH larger than wordprocessing files. For instance, the file for this web page contains just over 10,000 characters, but the picture shown below occupies a file that is over three times as large. Having said that, you should try to scan images at a higher resolution that you expect to use, and the same goes for the number of colors. Unless you are interested in professional publishing, you will probably print images at 300 dpi and display them on web pages at 96 dpi, so scans at 600 dpi should suffice. For small objects like the postage stamp shown below, use 1200 dpi or even higher to make sure that the image looks smooth when you enlarge it.

HOW DOES A COMPUTER HANDLE GRAPHICS DATA?
All computer files consist of strings (sequences) of numbers. Text data uses numbers to represent individual characters such as 13 for HARD-RETURN, 65 for A, 66 for B, 67 for C, and so on. [REVIEW] Graphics files can do it in either of two ways: 1) by plotting points and lines on a background and by 2) breaking an image into dots. The second approach is more accurate, but also creates files that take up more space.

For example, scanning a half-inch square black box on a one-inch-square white background at 100 dpi will create a file with 100 X 100 -- a total of 10,000 -- "small" numbers which can be squeezed into about 1,200 bytes. You could convey the same information with two coordinates for each of four corners of the square, two numbers for the length and width of the background, a number for the thickness of the sides of the square, plus a number for the square and background colors -- about a dozen bytes in total.

The density of the dot patterns is measured in dots-per- inch (dpi). The higher the number, the more detail will be available from a scanned image. But doubling the dpi makes the file size four times larger, so a one-inch square picture would require 40,000 numbers (about 4,500 bytes) at 200 dpi.

postage stamp at
600dpi
Above: 600 dpi       Below: 96 dpi

postage stamp at 95 dpi
Here the most frequently-used formats for saving graphics:

Filename ending Attribute(s)
TIF (or TIFF) Most accurate format and works well with wordprocessors, but files are large
JPG Good way to compress photographs
GIF Good way to conpress images other than photographs
Other formats are acceptable.

As a general rule, you should save your scanned image in the largest file possible. If you are saving it to a ZIP disk (100 or 250 megabytes), CD-ROM (700 megabytes), flash drive (a.k.a. "travel drive") or hard drive (even bigger), then size is not an issue. If you have an older computer that still uses floppy disks (1.44 megabytes), you will have to reduce the size of file unil it fits, by reducing the physical size of the image, the number of dots per inch, or the number of colors.

Improving scanned images

There are several ways to enhance the appearance of your scanned image. For example, an old photograph on yellowed paper will not photocopy well, but if you scan it in color, you can remove the yellow and restore the contrast of the original photo. The following table shows a sequence of techniques used to improve the scanned image of a West African postage stamp.

Original photo / remove yellow tint
Original photo Remove yellow
Remove postmark / Enhance faded red
AOF stamp with postmark removed AOF stamp with red enhanced 
edges

An enhanced image can reveal new information. For example, I discovered the picture on the left, below, in a photo album that once belonged to my grandfather. From other photos, it is clear that he'd served in France in the military during World War I, and that he had been around airplanes. Using a magnifying glass, I spotted what looked like "drome" on the sign over the building's door and guessed it was part of the word aérodrome, French for "airport." A web search yielded a list of French air bases where Americans served. By enlarging and sharpening the area containing the sign the building, I was just able to recognize the first word -- Ourches -- an important base located west of Nancy near the Western Front. After consulting secondary sources on U.S. military aviation and locating more clues in other photos, I concluded that by the summer of 1918, my grandfather, an Iowa farmboy, was part of the ground crew of a DeHavilland DH4 two-seater observation plane operated by the U.S. Army Air Service.

Original photo detail of sign

A map shows where things are located in a way that is much easier to understand than a description of roads and place names. A sequence of pictures may illustrate change over time. Headshots of individuals give us an insight into their character (or at least that's what we would like to think). On the other hand, be careful. Pictures should not be used to "decorate" a research paper. In a research paper, pictures should only be used to present information that cannot be presented clearly in words. If a picture doesn't do anything to make your presentation clearer, do not include it.

Maps are a special case. They offer the largest amount of useful data per square inch, thanks to conventions that identify blue with water, cross-hatched lines with railroads, double or thick lines with major roads, and so on. Two or more maps can also show change over time, such as this pair depicting the French colonial capital of Bamako before and after a railroad was completed to the coast.

Map of Bamako, capital of French
Soudan in 1894 before the railroad arrived Map of Bamako, capital of French
Soudan in 1912 after the railroad arrived
The European section of Bamako expanded after the railroad to the coast was completed in 1904.

Scanned maps can also assist with site visits and even archaeology. By superimposing information from an old map onto a current map, you can locate historic features with enough accuracy to guide a visit. For a book on West Chester's railroads, I identified several modern buildings as former train stations using an 1838 map of the railroad and a 1995 road atlas. My technique consists of scanning both maps, removing irelevant material, and using streams and road intersections to align the two images.

You may devise your own techniques for using images in history, but if you plan to scan them, always remember to capture the most possible detail. You can always remove detail later to make a picture that will serve your purpose.


Go to the HIS480 Syllabus or Assignments.