Cluster index for estimating thermal poultry stress (gallus gallus domesticus)

Abstract

Reducing production losses due to thermal stress is a challenge for poultry. The cost of operating HVAC systems is high, and precision in identifying critical situations for further action is necessary. Making decisions based on parameters inherent to animals has been widely studied and the analysis of animal behavior has advantages in terms of the lower cost of this monitoring and the need to not manipulate animals. This work proposes an index to estimate the thermal comfort of poultry birds (gallus gallus domesticus), based on the analysis of the birds' agglomeration behavior, called cluster index. This index is calculated from variables measured through the analysis of digital images, obtained in the filming of poultry groups from a surveillance camera installed on the roof of the farm. The image processing and analysis algorithm was developed on MatLab, considering that the filming was done with a camera installed on top of the birds, it can be applied to any video as long as the poultry have white plumage and are on a bed of dark wood shavings. The cluster index was tested on videos from two different experiments: one with three breeds of broiler breeders (Hybro, Cobb, and Ross) and one with pruning chickens (Nick Chik), bred on wood shavings and with known ambient temperature conditions. The results showed that poultry, whether laying or from heavy breeding, clump less in conditions of high temperature, and that among heavy breeding breeds, there is a significant difference in cluster behavior under conditions of comfort and thermal stress. It is concluded that the proposed cluster index was able to distinguish differences in the agglomeration of birds (gallus gallus domesticus) under different temperature conditions, from the automatic analysis of images and it is suggested that the cluster index associated with a remote monitoring system can be used to determine early thermal stress in poultry bred on shavings.