A Review of Machine Learning Models for Disease Prediction in Poultry Chickens

Abstract

Monitoring the bio-processes and bio-responses related to poultry farms is essential for the assessment and control of welfare-related parameters. In recent years, computer vision has emerged as a promising technique in the real-time automation of poultry surveillance systems due to its non-intrusive and non-invasive qualities, including its versatility to give a wide variety of information. Poultry farming is the activity of growing poultry birds which include chicks, geese, turkeys, and ducks for the production of meat and eggs to fulfill human needs. With the fast expansion of the poultry meat industry, stakeholders are concerned about the safety precautions taken to safeguard the birds’ well-being. In the chicken industry, early diagnosis of rising poultry disease outbreaks is crucial. The preventative phase ensures the health of broilers, and a number of methodologies are used to monitor and track the birds’ well-being. Chicken diseases are diagnosed via video surveillance, images, and sound observation using IoT-based devices. Chickens are susceptible to a number of ailments, including chronic respiratory disease (CRD), lameness, fever, respiratory disease, and others. The purpose of the paper is to review different methods to identify unhealthy chickens which can help farmers quickly administer the appropriate medication, preventing the diseased chicken from spreading throughout the farm and minimizing the financial loss. This article describes various chicken diseases, their preventative measures, and the various methods employed by researchers to pinpoint the disease. All of this has the advantage of allowing countermeasures to be undertaken ahead of time to prevent contamination of poultry farms or broilers.