AutoGCPs detection functionality

Access: This feature is only available on Pix4Dcloud Advanced.

Image and GCP requirements

For the AutoGCPs detection algorithm to be successful, the following requirements regarding the images and the GCPs have to be met.

Image requirements

  • Image quality has to be good (blurry or low contrast images make the algorithm fail). 
  • Images have to be geolocated (5-10 m accuracy is enough) in WGS84 (EPSG:4326).
  • Images have to be taken nadir. Oblique images are not supported.

GCP requirements

targets.png

  • Only squared and diagonal targets are supported.
  • Targets should not be smaller than 5 times the average GSD of the project.
  • Black and white targets are supported. Other colors may work in some circumstances.
  • GCPs must be placed at a distance of at least 10 m from each other.
  • GCPs must have a projected coordinate system whose definition is included at https://epsg.io.
Warning: If the AutoGCPs detection algorithm fails, the processing will run with no GCPs.

Upload the GCP file

  1. Follow the instructions to create a new dataset.
  2. Once at step 3. GCPs automation, select Upload file.
  3. Browse the CSV file which contains the GCPs' details and click Open.
  4. The GCPs are imported.
Important: The format of the file is described here.
  • First column (required): Label.
  • Second column (not required): Type (either GCP or CHECKPOINT with capital letters). The default value is GCP.
  • Third column (required): X coordinate.
  • Fourth column (required): Y coordinate.
  • Fifth column (required): Z coordinate.
  • Sixth column (not required): XY accuracy.
  • Seventh column (required if the sixth column is defined): Z accuracy.
Fields should be separated by commas. See the following example.
Example
001_,GCP,4392829.089,5386534.037,424.524
002_,GCP,4392772.823,5386545.0894,424.582
003_,GCP,4392732.981,5386516.197,424.723
004_,CHECKPOINT,4392761.505,5386468.789,422.646
005_,GCP,4392755.707,5386431.546,426.298
006_,GCP,4392832.252,5386456.906,427.001

Define the GCP coordinate system and target size

Once the CSV file is uploaded, specify the additional properties for the GCPs.

Important: Notice that only projected coordinate systems can be selected.
Important: If the GCPs are given in feet, the .prj file which defines the coordinate system must also be given in feet.
  • Diagonal target size: Specify the diagonal length of the targets in centimeters or inches.

See and edit the defined control points

After the coordinate system definition, click See and edit control points to access the control points Table.

Control points Table

The user can check the coordinates, change the type of the control points and swap Easting with Northing coordinates, by clicking one of the two words in the table header.

GCPtable.png

Control points Map

By clicking Map, an OpenLayers map is displayed with the GCPs and Checkpoints. This is a good way to check that the coordinate system and values have been well defined. If there is no coordinate system defined, the GCPs and CPs can not be displayed on the Control points Map.

mceclip1.png

Click on a point to switch from GCP to CP or the other way around.

Click Save to save the changes.

(Optional) Click Replace file to upload a new CSV geolocation file.

(Optional) Click delete.pngto delete the CSV file.

Start processing

Click Process to conclude the dataset creation and to start uploading the input files for processing.

  • The dataset starts to be processed.
  • The AutoGCPs detection algorithm will run. In case the GCPs are not found, an email is sent to the user email address explaining that no GCPs were found and the processing will be done without using them.
  • When the process is finished, an email is sent and the Quality Report of the project is attached.
  • In the Files > GCPs section of the project, if the GCPs were detected by the algorithm, the details of the GCPs will be shown. 

AutoGCPs report

An AutoGCPs report (PDF) is also generated and available for download. This report shows the detected marks for each point and it is intended to be used to ensure the GCPs have been correctly marked automatically. More information on how to download this report in How to download results from the Cloud.

An example of report for AutoGCPs marking follows.

AutoGCP_report_example.png

Was this article helpful?
2 out of 2 found this helpful

Article feedback (for troubleshooting, post here)

2 comments

  • Rio Hondo GIS Program

    Would ground control points with labels for for this workflow?

  • Alice (Pix4D)

    Hi,

    They are not officially supported in Pix4Dcloud Advanced but they have been recognized in some test projects we made. 

Please sign in to leave a comment.