Tjøtta lambing dataset

Wearable inertial sensors have been employed to detect lambing in sheep, with findings indicating that the most noticeable indicator of lambing is the sheep repeatedly standing up and lying down due to discomfort. In this study, a group of sheep was monitored using an accelerometer sampling at 20Hz; while lambing events were recorded on video to accurately determine the timing of births. Video analysis enabled the recording of birth times, along with additional details such as the number of lambs born and whether nursing assistance was required. The data was organized by individual ewe and stored in separate files, supplemented with information such as the ewe's birth year, previous lambing history, and ultrasound records. The dataset also includes a video of one birth to demonstrate animal behavior during the process. Over the course of the trial, 61 ewes were monitored, resulting in the birth of 113 lambs. This dataset has the potential to train an algorithm for classifying birth-related behaviors, paving the way for an automated lambing detection system.