Anne-Charlotte Olsson

Posture detection targeted towards providing assessments for the monitoring of health and welfare of pigs has been of great interest to researchers from different disciplines. Existing studies applying machine vision techniques are mostly based on methods using three-dimensional imaging systems, or two-dimensional systems with the limitation of monitoring under controlled conditions. Thus, the main goal of this study was to determine whether a two-dimensional imaging system, along with deep learning approaches, could be utilized to detect the standing and lying (belly and side) postures of pigs under commercial farm conditions. Three deep learning-based detector methods, including faster regions with convolutional neural network features (Faster R-CNN), single shot multibox detector (SSD) and region-based fully convolutional network (R-FCN), combined with Inception V2, Residual Network (ResNet) and Inception ResNet V2 feature extractions of RGB images were proposed. Data from different commercial farms were used for training and validation of the proposed models. The experimental results demonstrated that the R-FCN ResNet101 method was able to detect lying and standing postures with higher average precision (AP) of 0.93, 0.95 and 0.92 for standing, lying on side and lying on belly postures, respectively and mean average precision (mAP) of more than 0.93.

In literature, piglet mortality is described as a multifactorial complex influenced by factors as litter size, age and health of the sow, farrowing system, management etc. In this study, a parallel comparison was made between two farrowing systems; a temporarily confined (TC) (farrowing – 3 days after) versus a loose sow (L). On average, 0.4 more pigs per litter survived until weaning if the sow was temporarily confined (TC) compared with being loose (L). Cause of death was recorded according to a strict template. Underweight and crushing was the most common causes. For crushing, during 1–3 days after birth, an interaction between sow age and farrowing system was observed, with differences between farrowing systems only for intermediate-aged and older sows (>parity 2). There were no significant differences between systems regarding farrowing duration or number of stillborn pigs, but a significant increase in farrowing problems was recorded for TC-sows.

The effect of feeding environment on individual pigs (small, medium and large barrows) housed in groups of 16 pigs was studied. The experiment comprised the following four treatments: trough feeding, four dry feeders, one dry feeder and one wet/dry feeder per 16 pigs, respectively. Performance, skin injuries, plasma cortisol, agonistic behaviour, ingestive behaviour and activity were studied. The effect of competition for feed was most clearly evident for the small pigs. The small animals in the one dry feeder treatment were forced to eat more during the night, had more forced withdrawals from the feeder, and tended to have more scratches and skin injuries than did the small animals in the four dry feeder treatment. It was concluded that this competition led to impaired well-being in the smallest pigs. No significant differences were detected in plasma cortisol levels, when comparing the different feeding systems within each pig size category. Trough feeding greatly affected pig feed intake pattern: during the growing period when pigs had ad libitum access to feed, the pigs fed in a trough ate mostly at feeding times, while the pigs in the four dry feeder treatment ate at times spread out over the entire 24 h period. It was concluded that there was a confounding effect between pig size category (small or large) and feeding system with regard to ingestive behaviour and aggressive interactions at feeding.