Georeferencing with Manually Recorded GCPs
Use manually recorded GPS locations to align your photos with the world.
What is georeferencing?
The goal behind our Manual Georeferencing Workflow is to correlate pixels in the imagery with their
corresponding real-world GPS coordinates. The map above is an example of aerial imagery being layed over a public basemap. Zoom in and see the difference in the available resolution. Use the control in the top right to hide the georefereced layer to see how things line up.
This alignment is acheived by the user creating correlations between a few images and manually recorded GPS locations for objects in the photos. These correlations are called Ground Control Points, or GCPs. In creating these GCPs, a bit of precision is required to carefully associate objects in the images with the manually recorded points.
It is important to get accurate data when recording your GCPs in the field. Map will only be as accurate as their GPS points are. Also, when associating the objects where the points were recorded with the recorded data, be sure to select the exact location where the reading was taken. A little bit of approximation is OK, but try to get as close as possible. Being off by a couple of pixels will likely still work out, but more than 5 pixels or so is not. Remember: garbage in, garbage out.
- Zoom In - Scroll or click
- Zoom Out - Scroll or click while holding Shift key
- Define Point - Right Click
A manual GCP is created by defining a marker on an image and entering the latitude and longitude in decimal degrees optionally the alititude of the object in meters.
If the points are being taken on a GPS receiver that cost less than $1000, it is probably better not to enter an altitude. In this case, an altitude will be pulled from the best available public elevation database. If the area being surveryed is heavily modified (construction or excavation), the manually recorded point will likely be better but take care to reference the GPS receiver to a known elevation before making the recordings.
An Anchor Point consists of two GCPs that share an exact GPS location. This means that an object that is present in at least two images should be defined and assigned the EXACT same location coordinates. The "Reuse this location" button is a shortcut for this. Our process requires the definition of at least three Anchor Points in addition to at least one standard GCP.
Anchor Point Location Selection
Feature selection when out in the field is an important component to creating accurate maps.
Choose points within the area being mapped that will be easy to find in imagery.
Select locations that are the following:
- Going to be present in at least two images, not on exteme edges.
- Features on the ground, NOT man-made structures or vertical features like trees.
- Reasonably spread throughout the mapped area (not right next to each other.)
- Not in a straight line.
- Permanent so they can be reused for subsequent visits.
- Inside the flight path envelope. GCPs that are outside the interior or the area defined by the flight path risk being cropped out of the map.
Anchor Points Creation
Create pairs of GCPs that share a identical GPS location.
Anchor Points give Maps Made Easy the basis for aligning a created map with the world. The created map will only be as accurate as the Anchor Point entry. Sloppy Anchor Point entry can result in dramatically skewed results.
For each pre-selected object:
- Click "Add New GCP".
- Click Choose File and select one of the images in which the object the measurement is present.
- Zoom in and right click on the object in the image where the measurement was taken. A marker will be drawn on the basemap as a visual check.
- Enter the Latitude, Longitude and Elevation values for the location designated in the image.
- If everything looks good, click "Save".
- Click "Add New GCP" again to create the other GCP to be used in the anchor point.
- The Latitude, Longitude and Elevation numbers that were defined on the previous GCP are automatically populated.
- Click Choose File and select another image (not the same as before) in which the object is present.
- Press the orange "Reuse this location" button to use the EXACT location again.
- Zoom in and right click on that same object in the image. Again, make sure to select the exact area of the object that was selected before.
- If everything looks good, click "Save".
- The table should be shaded green to show that these points represent on of the 3 required anchor points.
- Repeat this process for at least two more anchor points.
Note: more Anchor Points or GCPs is not better. Anchor Points and GCPs do not affect the reconstruction and are only used to align the reconstructed model with the real world. 3 really good anchor points are better than 10 OK ones.
This is what a properly assigned GCP table should look like:
It sounds complicated but it really isn't. Check out our videos to see how quickly it can go. The workflow wizard will keep track of when enough GCPs and Anchor Points have been created and allow you to advance when all criteria is met. Good luck!