Home    Products    Support  Downloads    Upgrades  Agents  Contact    About

 

Fairport Mapping and Image Digitising : Help Notes

Digitising is the process of describing an Aerial photo to the mapping system. To do this, you need a minimum of two points. From these two points, any location on the aerial photo can be determined.

Mapping needs to know how wide each pixel of an image is in real world coordinate.

For example: If you aerial photo shows an image 2300 metres wide, and the actual bitmap is 538 pixels wide, then it has a grid size of 4.275 metres per pixel. (Calculated by dividing the real world width by bitmaps width in pixels = 2300 metres / 538 pixels = 4.275 Mpp)

This figure is displayed at the bottom of the Digitising screen once you have entered two or more valid points:

In this example, the 1.80 MppV tells you how many metres each pixel is worth vertically, and 1.81 MppH is Metres Per Pixel Horizontally. For a correctly geolocated image, these values should be close to identical.

Some typical problems you may run into when geolocating you image:

Image is rotated

This is the most common. It may be out by 180 degrees (the image is expected to contain the northmost point at the top), or it could be out by just a few degrees. The Digitiser will accept angles of less than 1 degree without complaint, but will offer you the option of rotating images with an error between 1 and 45 degrees. Sometimes a rotation error may be detected, but what if the result of rotating the image still results in an error? This can happen if the image is warped or stretched….

Image is warped

Rare, but possible. This can happen when several aerial photo's have been 'stitched' together to make one big bitmap. The larger the area, the more potential for warping. Fixing warp errors is out of the scope of the program - you should contact whoever stitched together the images for advice.

How Much Error Is Too Much?

How much is too much? In the example above, there is an error of 1cm per pixel. (1.81-1.80). This error may have occurred by inaccurately specifying the pixel coordinates, or world coordinates when digitising, or it could be that the image is genuinely stretched.

If we assume that the image is stretched horizontally, then a 1 cm error per pixel means that every 100 pixels we shall accumulate an error of 1 metre. As our image is 210 pixels wide, this means that we could be from 0 to 2 metres out at the left or right of the image. Also, if the image is stretched horizontally at the top, but not at the bottom, then the error could be 1 to 2 metres at the top and gradually result in no error at the bottom of the image. How much error is acceptable is for you to decide based on your usage of the map.

Confirming Geolocation of Images

The best way to confirm geolocation is via real world coordinates collected with a GPS unit. Collect readings for fencelines, roads and other clearly visible land marks on your aerial photo and import them as points. When viewing these over your aerial photo you can zoom right in to verify accuracy of the aerial photo positioning. If the image is not correctly located, you can use the sizing and moving tools to adjust the location of aerial photo visually.