Together, fGIS and HyperCube can be used to rectify and georeference imagery. HyperCube, freeware produced by the US Army Corps of Engineers Topographic Engineering Center, includes functions to analyze, filter, warp, mosaic, reformat, calibrate, and combine multi and hyper-spectral imagery and data. It comes as a single "exe" file, which does not require installation. Just unzip the file and use it.

The following technique is helpful if you have scanned hand-drawn maps, paper maps, printed photos or any digital images that you would like to warp or rubber-sheet to conform to a known projection.

There are four basic steps:

1. Scan or copy and save a geographic image (A). Either a TIFF or JPEG format will do.
2. Use fGIS to create a georeferenced target image (B) and world coordinate file from previously rectified vector or graphic files.
3. Load the source and target images into HyperCube. The warp (C) is accomplished by designating common control points on both the source and target images. HyperCube will scale, rotate, warp and crop (A) as needed so that it is the same size and shape as (B).
4. Finally, using Windows Explorer, the world coordinate file from step 2 is copied from (B) to (C).

Detailed Procedures

With Your Photo/Scanning Software:

If you scan a paper map or line drawing (such as a timber-type overlay) on a flatbed scanner, capturing the image at 150 DPI to 200 DPI is generally sufficient.

Paper aerial photographs are typically 9" x 9", whereas most consumer type scanners are limited to an 8.5" width. Inexpensive scanners may introduce distortions or noise in addition to aberrations caused by terrain, optics or aircraft orientation. That said, a consumer-grade flatbed scanner can produce acceptable results for smaller aerial photo conversion projects. 

The following table provides three typical aerial photo scales and the resulting pixel resolution and image output file sizes for various DPI scan settings for a 9" x 9" photo.

Table Source: http://www.ncgia.ucsb.edu/education/curricula/cctp/units/unit18/18.html

In fGIS:

If you don't know how to set up a view with fGIS, see the Quick Start Guide. The fGIS view can include a previously registered aerial photo and/or any vector layers that might provide good landmarks. Save the view with 'Map>Export View to Image' or use the fGIS image export tool. Save the image in either a TIFF or JPEG format (accepted by HyperCube). fGIS will write a world coordinate file for the saved target view, which will be the key to geo-referencing the source image.

Selecting common control points in the warping process will be much easier if you make the source and target images approximately the same dimensions and pixel resolution. For example, let's say you scan a paper map or photo and the resulting source image is 3,000 pixels wide. If your screen resolution is 1024 x 768, you could increase the output image size for the target image in fGIS by 3 (shown in the dialog below, assuming a full-screen view) to produce a target about 3,000 pixels wide.

Making the source and target images about the same size also means that you are less likely to lose pixels (details) when the source image is warped.

In HyperCube:

1. Load in the "source" and "target" image. Use "File => Open" and specify the file type you want to open from the drop-down. If either image is in  indexed color (e.g. a standard 256 color DRG USGS topo), you'll have to use Image=>Convert Image to change it to true color.

2. Select Functions=>Warp

3. Choose the source and target images using the drop-down menus, and also specify source and target (1=>2 or 2=>1). Note that you must specifically choose the images (click on their names from the drop down for Image 1 and Image 2) even if  the names are already visible in the drop-down menus.

4. When you've successfully chosen the two images, the "Instructions" section in the Warp window will now tell you to select corresponding points in the two images by  shift-clicking on the same geographic point in each window. The minimum number of points will depend on the Transform you've selected, although at least three points are needed for the Affine Transform (which is generally the most useful). HyperCube will allow you to select any number of reference points. When you're done, press Warp and the source image is warped into the shape and dimensions of the target image. The title bar of the new image includes a code (Such as "A_N" for Affine/Nearest Neighbor) to help you keep track of which Transform and Interpolation method was used. You can change the parameters and experiment with the results. The control points can also be saved for re-use.

5. You can save the warped image as a TIFF, JPEG, or several other formats, and then use your original world file to georeference it (i.e., copy and rename the world coordinate file from the target to the warp).

If the target image is a GeoTiff, you can also use Hypercube to georeference the new image. Select the georeferenced target image, and choose  Edit=>Options=>Image to Ref Coords. This will bring up a  window with the georeferencing information; write down all  the numbers and settings here. Then select the new image  you've created by Warp, bring up the Image to Ref Coords  window for this image, and then enter the georeferencing  information and settings exactly as you copied it in the  previous step. Now when you save this image as a TIFF, it will embed the georeferencing info directly into it.

Note: HyperCube relies on you to type in the file extension suffix. If you want to save the warped image as a TIFF, make sure you append ".tif" to the name of the file.

HyperCube has excellent documentation in PDF format. See pages 75-76 for more information about the image warping options.

If the Army Corp of Engineers HyperCube page is down, you can also download the program and documentation from here:

• HyperCube.zip (797 KB)

• HyperCube Pictorial User's Guide (11.8 MB PDF)

To reproject an image from one coordinate system to another, also see DGWarp.