Acquisition of broiler dimensions through depth image analysis

Abstract

Chicken is one of the most consumed proteins in the world, which has made the productive system intensive. The intensification of the rearing system can be a source of stress to animals, making it necessary to search for alternatives to manage and monitor the physical condition of animals in a less invasive way and to promote an environment that allows genetic optimization for weight gain and deposition of musculature in the carcass and the free expression of the natural behaviors belonging to the species. The use of remote monitoring tools and less contact with animals is an alternative to eliminate possible stressors to the animal, which promotes a better way of handling and management and will promote a better-quality meat, while respecting the principles of animal welfare. Such measures allow the adhesion and acceptance of the product by consumers who are concerned with how animals are treated during the production system. Thus, the use of image analysis may be an alternative for monitoring the body development of animals throughout the production. The present search aims to automatically obtain the body weight of broilers through depth image. The KinectAzure® sensor will be used for image collection, data analysis and subsequent prediction of body dimensions will be performed using an algorithm developed in the mathematical software MATLAB® (R2018a), so that a weighting method will be developed. It is expected to develop an effective and practical program to monitor the body mass of broilers at a distance during the production system. It is also hoped to develop a method for obtaining commercial bird sizes, which will assist in new poultry plant and equipment designs.