Making the correct decision regarding image resolution is critical.

I provided an overview of digital orthophotography in my last column (August 2001.) Digital orthophotos, which are accurate photographic-based computer maps, have become very popular in all areas of the surveying and engineering arena.

That column briefly addressed the important topics of image resolution and file size. Making the correct decision regarding image resolution early in the planning stage is critical to the success of any orthophoto project. The resolution should be high enough to ensure the necessary detail is available in the imagery. For example, a utility project might require imagery that clearly shows manholes and utility poles. In this case, a resolution of at least 1 foot should be specified to ensure these features are clearly visible.

You must be careful, however, and not specify an overly high image resolution. Doing so can substantially increase a project’s cost. Moreover, while a really high resolution image can be aesthetically pleasing, the corresponding file sizes at ultra high resolutions can create significant problems when working with and storing the images. Therefore it is important to select a resolution that will provide the necessary features while striking a balance with project cost and file size.

Resolution and Features

Digital images are made up of fundamental pieces of information; each piece relates to the visual characteristics of a small part of the image. In the computer world, these individual pieces of information are known as pixels. Pixel is short for “picture element.” Each pixel in a digital orthophoto represents a specific area of the ground captured in the photo.

The ground dimension along each side of the pixel is known as the spatial resolution of the image. For example, an orthophoto is said to have a resolution of 2 feet when the ground distance along each side of the pixel measures 2 feet.

Each pixel contains the predominant radiometric, or visual, characteristics of the area on the ground represented by the pixel. For example, a pixel that falls predominantly on a concrete sidewalk will be very light in color to reflect the light color of the concrete. Similarly, a pixel that falls on a new asphalt roadway will be correspondingly dark to reflect the characteristics of the new asphalt.

Higher resolutions are gained by cutting up the ground into smaller squares in the digital image. Notice the sample digital images under the Determining File Size portion of this article. Resolutions of 1-, 2- and 3-feet are illustrated. Look closely at the detail in the image captured at the 1-foot resolution. The cars parked along the side of the street appear very crisp and clear in the imagery. The individual trees located along the street are also easy to see. Now take a look at these same features in the imagery at resolutions of 2- and 3-feet. There is a substantial difference in the clarity of these images.

Determining File Size

Cost is an important consideration during the planning of any orthophoto project. As a general rule, the cost of a project increases as the resolution is increased. It is also very important to have a fundamental understanding of the final file sizes that should be expected at the different image resolutions considered for the project.

A black and white digital image has a single byte of storage for each pixel. That byte of information contains an integer value between 0 and 255, which relates to a grayscale varying from pure white to pure black. Black and white orthophotos are most often used today, however color images are gaining more popularity. A color image makes use of three bytes of storage for each pixel, one byte each (again ranging in value from 0 to 255) for red, green and blue bands. The combination of these three bands allows all colors to be accurately represented in the image. Hopefully it is apparent that a color orthophoto will be three times the size of a black and white orthophoto covering the same area.

The number of pixels in any image can easily be determined by dividing the project area by the ground area covered by each pixel.

Number of Pixels = Ground Area / Area Covered by Each Pixel

Be careful with the units; they must be the same. Since the image resolution is the dimension along one side of a square, the ground area covered by each pixel becomes the square of the resolution (technically the product of the width and height.) The result is the ground area in square feet when the resolution is specified in feet.

Area Covered by Each Pixel = Resolution2
(Resolution in feet; area in sq feet)

Since the units of both the numerator and denominator must be the same, the ground area needs to be converted to square feet. Most often you will know the area in either square miles or acres. Therefore it becomes necessary to apply the following conversion factors to the known area to convert to the units of square feet.

1 Acre = 43,560 Square Feet
1 Square Mile = 27,878,400 Square Feet

And finally, most computer images are given in terms of megabytes (MB). In general terms, a megabyte is one million bytes. More technically, however, a megabyte is equal to 220 bytes, which works out to 1,048,576 bytes. These calculations are for uncompressed images. Compressed image formats will yield smaller file sizes. The exact reduction in size will depend on the type of image compression used and the quality factor selected during compression.

Now let’s take a look at a couple of numerical examples to illustrate the calculations.

Example 1

A new subdivision is planned for an area that totals 500 acres. Planimetric and topographic mapping has been completed for the area. Black and white digital orthophotos are being considered at alternate resolutions of 0.5 or 1.0 feet. Determine the total size of the imagery for each option.

Number of Pixels = Ground Area / Area Covered by Each Pixel

At a resolution of 0.5 feet
Number of pixels = 500 acres * 43,560 sq ft per acre / 0.52
= 21,780,000 sq ft / 0.25 sq ft
= 87,120,000 pixels
= 87,120,000 pixels x 1 byte per pixel (for B&W)
= 87,120,000 bytes / 1,048,576 bytes per MB
= 83 Megabytes

At a resolution of 1.0 feet
Number of Pixels = 500 acres * 43,560 sq ft per acre / 1.02
= 21,780,000 sq ft / 1.0 sq ft
= 21,780,000 pixels
= 21,780,000 pixels x 1 byte per pixel (for B&W)
= 21,780,000 bytes / 1,048,576 bytes per MB
= 21 Megabytes

Example 2

A new roadway is planned along a cross-country route. Digital orthophotos will be used to aid in the planning, design and public involvement processes. An image resolution of 2.0 feet has been selected for a corridor width of 8,000 feet along the length of the project. Determine the file size per mile of corridor length for both black and white and color orthophotos.

Each mile of corridor will have an area of 8000 ft x 5280 ft, or 42,240,000 sq ft.

Number of Pixels = Ground Area / Area Covered by Each Pixel
= 42,240,000 sq ft / 2.02
= 42,240,000 sq ft / 4.0 sq ft
= 10,560,000 pixels

Since a black and white image requires 1 byte of storage per pixel, the file size per mile of corridor will be

= 10,560,000 pixels x 1 byte per pixel / 1,048,576
= 10 MB per mile of corridor

We know that a color image uses three bytes of information per pixel. Therefore the file size will be three times that of a black and white orthophoto, or

= 10 MB per mile of corridor x 3
= 30 MB per mile of corridor

Conclusion

It is very important to carefully consider the options when planning a digital orthophoto project. You must select a resolution that will allow you to see the features on the ground that are needed for the project. Be careful, however, and develop a clear understanding of the project cost and file sizes that will result based on this decision. Even though the computer technology available to us today is significant, you must be sure that you can store and display these images in acceptable ways.