Our empty event detection is based on a combination accelerometer trigger* and the fullness of the before and after pictures.
Reasons for missed empty events:
Obstructed images: An obstructed camera can prevent our AI from determining the fullness. If the image following an accelerometer trigger is obstructed and the fullness was not determined, then the empty event may not be detected till the next image with a determined fullness level.
Little change in fullness between two visible images: If the two images contain similar levels of fullness, then it may determine there was no empty event even if there was an accelerometer trigger.
Lift event determined but indicated on the wrong day: If an empty event is determined to have occurred between two images, but there was no accelerometer trigger between them or there is a series of obstructed images between them, the empty event will be indicated as having occurred at the average time between the two images. This may result in the empty event not being assigned to the day that it actually occurred.
* Note that accelerometer triggers may not occur if the container is lifted in a way that does not comply with the algorithm, that wakes the camera to take a post, which is used to identify that the motion the camera is experiencing is a lift.