Growth recorded automatically and continuously by a machine vision system for finisher pigs

Abstract

Conventional weighing methods in the livestock industries require direct contact with the animal. Due to this contact, conventional weighing methods are both physically demanding and hazardous for those involved. Alternatively the live weight of an animal can be estimated from its body dimensions using non-invasive methods. This paper presents the recent improvements in the ongoing development of a completely automatic, two dimensional computer vision system, designed to obtain critical dimensions of the body of pigs in order to estimate their live weight. Results from validation trials (comparing conventional livestock weighing results and the results generated by the vision-based system) are reported for pigs in their 'finisher' stage. Currently average group weights are predicted with ± 0.7 kg precision under commercial farm conditions. Recording the continuous live weight change of livestock (growth) is important as it can be used to measure the animal’s response to various factors such as the surrounding climate, housing environment and nutrition. Assessing the animal’s response to these conditions is essential in improving the efficiency and welfare of livestock in both research and commercial settings.