30
4.1.4 Calibration
After adding a sensor to the system for the first time, and after the firmware upgrade, the app will ask
the user to calibrate the sensor to detect the floor area in the room before it can be used for fall
detection, as shown in the figure above.
Note: The current algorithm only sends alert when falling to the ground is detected. Falling on couch
or bed is not considered dangerous and therefore will not trigger alert.
Make sure that no people are in the view of the sensor during the calibration, in order to protect
user’s privacy. After that, every time the position or angle of the sensor is changed, the calibration
should be done again manually.
To do the calibration, tap the “Calibrate” command to enter the Calibrate page, as shown below. If the
sensor has not been calibrated, then when you enter the Calibrate page for the first time, the sensor will
automatically capture a background image and start to detect the floor in it. For the Cypress sensor, the
calibration takes about 15 seconds. For the Sentinare 2 sensor, the calibration takes about 30 seconds.
If the sensor was calibrated before, the App will first display the previous background image and floor
map. To update the calibration for the current sensor position, tap the “Recalibrate”, the sensor will
take a new background image and run the calibration algorithm to it. If the user is satisfied with the
floor calibration result, tap the “Save” button to save the result and exit the page.
If the sensor is installed upside-down (for Sentinare 2, the USB port is in the top if the sensor is installed
upside-down), the initial background image taken in the calibration page will be upside-down. To display
the correct image, please turn on the Upside-Down Installation toggle in the calibration page (Sentinare
2 only), and then recalibrate.
The Upside-Down Installation toggle is also available in the sensor’s main page (for both Cypress and
Sentinare 2). If the sensor is installed upside-down, you can set Upside-Down Installation toggle to ON in
the sensor page before entering the calibration page. This ensures that the first background image
automatically captured in the Calibrate page is correct. This is handy if you plan to add a batch of
sensors in the upside-down way.
Manual Calibration
Sometimes the auto calibration algorithm may mis-calibrate some non-floor areas as floor, or miss some
floor areas, which could cause false fall detection alarms or miss some falls. In this case, users can do
manual calibration. To do this, first tap the “Clear Floor” command to clear the automatic calibration
result, and then use a finger to redraw and fill up the correct floor area on the screen, as shown by the
second image below. After that, tap the “Save” button to save the result.
To Next Page
Loading...