Friendly Reminder: For the most up-to-date information regarding this article, please use the following link: https://confluence.sportradar.ag/display/SK/Scoreboard+Readings%3A+Arena+reader+analyze+interface.
Calibration is the process of configuring the decoding from an OCR Reader Camera (Optical Character Recognition) pointed at the Scoreboard in an arena.
Re-calibration might be needed due to camera movement, physical issues at the arena or the changing layout of the scoreboard.
When we notice that there is a calibration issue during a live game (Score, shotclock or game clock are not matching to the ones shown on the scoreboard in the arena), the first thing we need to do is open the “Arena” tab in WebAdmin and open the details for that arena.
After we open the details for that particular arena, the next thing we need to do is open the “Readers” tab located in the top right corner of the page within WebAdmin.
After clicking on the “Readers” tab, you will an example of the image below. What you need to do is select the reader in which the issue occurs (usually its reader 1 for shot clock and reader 3 for the scoreboard).
Pro Tip: By default, reader1 captures from Camcourt1, reader2 from Camcourt2, and reader3 from Camscore1. If you want to modify these defaults, you can define which camera the reader captures from in the director.conf file within WebAdmin.
For example, for reader1 you can enter #define READER1 C_STREAM3 if you want it to capture from CAMCOURT3 or #define READER1 S_STREAM2 if you want it to capture from CAMSCORE2.
After opening the desired reader, a page will be displayed like the one below (New interface as of August 2023).
The image above displays a list of previously captured images from the reader. This list of images is used to:
- Adjust the calibration by selecting an image to be used in the Scoreboard calibration tool to define the different regions where each piece of information should be decoded (Examples: scores, clocks or quarter)
- Run regression tests to ensure that the system continues to correctly decode the numbers/digits in a set of images from the previous recordings. In order to run such test on an image, a ground truth must be set for the image. The new interface allows us to set the ground truth values.
Capturing a new image and adding it to the "Reader's records" list
After the initial creation of the reader or after any modification leading to incorrect decoding of the information, you will need to capture and add new images to the "Reader's records" list.
By default, the interface is proposing, in the "Candidate records" list, images captured in the venue every hour in the past 24 hours.
Note: You might need to expand the "Candidate records" section to see the list of images.
In case none of these automatically collected images is relevant, you can trigger the extraction of additional images from a past recording using the "Capture" button. You can specify the exact game ID and time within the recording for accurate frame extraction. Try to extract frames where all information (scores, clocks,..) is being displayed. The extracted images will be added to the "Candidate records" list.
In order to review an image in the "Candidates records" list and decide if you want to select it for an addition to the "Reader's records", you need to click on the "Original" selector in front of the image name.
In order to add the image to the "Reader's records", click on the "Add image to the list" button.
Beware that images listed in the "Candidate records" list are deleted after 24 hours and replaced by newer ones.
Correcting calibration when motion was detected
See page "ReaderX : Scoreboard (or camera) has moved"
Launching the Scoreboard calibration tool on an image
In order to open the scoreboard calibration tool with an image, click on the "OCR" button in front of the image.
Running Regression Tests and Setting Ground Truth on Images in the "Reader's records" List use the Link Below
Was this article helpful?
That’s Great!
Thank you for your feedback
Sorry! We couldn't be helpful
Thank you for your feedback
Feedback sent
We appreciate your effort and will try to fix the article