To improve efficiency by removing manual steps, Site Scan can automatically detect specific types of ground control point (GCP) targets within the images, and pretag them to the corresponding GCPs.
Site Scan uses computer vision to detect checkerboards and crosses in aerial imagery. Esri recommends using checkerboards for automatic GCP detection. For optimal detection, the targets should provide a significant contrast from the background on which they are set.
Automatic GCP detection is turned off by default for all projects, and must be explicitly turned on if you want to use it. To turn on automatic GCP detection, complete the following steps:
- Open the project for which you want to use automatic GCP detection.
- Browse to Output Settings.
- Scroll down and check the Automatic GCP Detection check box.
- Click Save.
Considerations for automatic GCP detection
The following are some things to consider for automatic GCP detection:
- Automatic GCP detection will run only if no GCPs have been tagged manually.
- Manually tagging any photo will prevent you from using automatic GCP detection for that mission.
- Ensure that the desired GCP set is selected in the mission's Ground Control Points layer. If you have multiple GCP sets, select the one you want Site Scan to automatically detect for you.
- Process the flight normally using the Process button.
- You can select a lower-quality setting when processing the first time to speed up the detection. After GCPs are detected, you can reprocess them at a higher-quality setting of your choice.
Once the detection has finished running, you will receive an email notification. Click Review GCPs under the Ground Control Points layer of the mission.
After selecting the Review GCPs option, the following can be observed:
- All of the mission's image thumbnails with GCPs autodetected, or estimated
- The total number of GCPs detected
- The name of each GCP, and the number of photos that have been detected or estimated for each GCP
- The date of the upload for the project's GCP set
Types of detection
There are two types of detection in the Site Scan GCP detection workflow: detected and estimated.
All images with detected GCPs are automatically selected for processing.
Images with estimated GCPs are deselected; by default, they will not be used in processing, but they can be selected by checking the check box and adjusting the crosshair location if needed. Estimated indicates that the detection confidence was low, but a GCP could be present in the image.
Adjust and process GCPs
If necessary, adjust the center of the detected GCPs in the images to precisely select their exact survey location. To adjust GCPs, scroll to the image in the thumbnail and click the center of the GCP to align the crosshairs.
To finish, click the Process button in the upper right window to process the flight with the tagged GCPs, or click Save for Later to review the GCPs later. You can also add more tags through the manual GCP tagging workflow by clicking Manually Tag GCPs.
Add more manual tags
If some GCPs were not automatically detected or not detected in enough (5+) photos, use the Manually Tag GCPs option to manually tag more photos. The process is identical to manually tagging GCPs (see the Use manual tagging topic).