Explore the importance of Ground Control Points (GCPs) for georeferencing point cloud reconstruction and learn how to upload GCPs in Reconstruct.
Section 1: The Importance of Ground Control Points (GCPs)
Reconstruct uses GPS location data from uploaded images to georeference point cloud reconstruction. Ground Control Points (GCPs) are landmarks with known geospatial positions (latitude, longitude, and altitude) that can be observed in one or more images. These points enhance the accuracy of the point cloud reconstruction.
Section 2: Preparing a GCP Text File
Users can upload a text file containing GCP information to Reconstruct before processing the point cloud. This text file should include GCP coordinates and their locations in images. To create this file, you can use a third-party app such as WebODM, following the instructions provided in Section 7 of this guide.
Section 3: GCP File Format
The GCP text file should have a specific format. The first line contains the name of the projection used for geo coordinates (e.g., WGS84). The following lines contain data for each GCP observation in this format:
<geo_x> <geo_y> <geo_z> <im_x> <im_y> <image_name>
Here, <geo_x> <geo_y> <geo_z> are the geospatial coordinates of the GCP, and <im_x> <im_y> are the pixel coordinates where the GCP is observed in the image. Set <geo_z> to NaN if the GCP's altitude is unknown.
Section 4: Supported Projections
Reconstruct supports various geospatial coordinate formats, including WGS84, UTM, and proj4. This section provides examples and explanations of these formats.
WGS84
This is the standard latitude, longitude coordinates used by most GPS devices. In this case, <geo_x> = longitude, <geo_y> = latitude, and <geo_z> = altitude.
UTM
UTM projections can be specified using a string projection string such as WGS84 UTM 32N, where 32 is the region and N is the hemisphere. In this case, <geo_x> = E, <geo_y> = N, and <geo_z> = altitude.
proj4
Any valid proj4 format string can be used. For example, for UTM 32N we can use +proj=utm +zone=32 +north +ellps=WGS84 +datum=WGS84 +units=m +no_defs.
Section 5: GCP File Requirements
For the GCP text file to work, it should contain information about at least 3 GCPs, with each GCP visible in at least 4 images (5 recommended GCPs and images per GCP). The file should have at least 12 lines (25 recommended) and use spaces, not tab characters. Ensure that the uploaded images have the same names as those used in creating the GCP file.
Section 6: Uploading the GCP Text File
Follow the instructions below upload the GCP text file to the system and process the point cloud.
Section 7: Generating GCP Text File Using WebODM Tool
WebODM is a free third-party tool that enables you to create your GCP txt file very easily.
What you need:
- A list of your GCPs with their name, LLA information, their coordinate system, and a visual clue that where they are placed on the job site. This list can be prepared by the Survey team. A sample of this looks like the below table.
- A set of drone capture images that each of your GCPs is visible in at least five different frames.
Steps:
Part 1: Preparing GCP Information
- Go to https://www.spatialreference.org/ref/epsg. Now search for your coordinate system name in which your survey data is captured (you can ask your surveyor). In this example, LLAs are in the PRS92 system.
- Click on the Proj4 link to get your coordinate system information in Proj4 format.
- +proj=tmerc +lat_0=0 +lon_0=121 +k=0.99995 +x_0=500000 +y_0=0 +ellps=clrk66 +towgs84=-127.62,-67.24,-47.04,-3.068,4.903,1.578,-1.06 +units=m +no_defs
- Create a txt file using Wordpad or Notepad with the following format (this is not your final txt file).
- Copy and paste the Proj4 code to the first line. Press Enter and go to the next line.
- Type the information of each of the GCPs in one line following format (separated by space): Name Easting Northing Elevation.
Part 2: Uploading GCPs and Images to WebODM
- Go to https://demo.webodm.org/plugins/posm-gcpi/.
- Click on "Load existing control points file" and select the txt file you prepared in the previous step to upload. After you confirm the information, click on "Load".
- Go to the folder that your drone capture images exist. Try to find five image frames in which the first GCP (in the first line of the txt file) is visible (in the example above, SC21).
- Go back to the WebODM page, click on "Choose images" and upload those five images to the interface.
Part 3: Marking GCPs on Images
- Double click on the first image to open. By default, you would see an arbitrary point added on the image side as well as the map side. Left-click on each and delete them.
- Now, on the image side (left screen), click on (+) to place a new point. Find the center of the marker and left-click on it. If you hold the left click, you can modify the position of the placed point.
- You have to decide which of the existing GCPs on the map this newly selected point on the image represents. This can be simply done by first left-clicking on the point on the image side to make it selected, then another left-click on one of the GCPs on the map. If you do this successfully, both points become green.
- Go back from this image and repeat this step for all the images that you uploaded.
- Upload a new batch of images in which another GCP is visible in them, then repeat steps 7-13.
Part 4: Exporting the GCP txt File
- After you are done, click on the "EXPORT FILE" button on the corner of the screen to get your TXT file.
Comments
0 comments
Article is closed for comments.