The integration of advanced technologies such as IoT and AI is revolutionising poultry farming, enabling real-time monitoring of chicken health. Through the use of sensors and predictive models, abnormal behaviours can be detected, and the spread of diseases can be prevented, enhancing the management and sustainability of poultry farms.
Poultry farming plays a crucial role in feeding the world’s growing population. However, abnormal bird behaviours can lead to health issues and the detection of diseases often relies heavily on observing these behaviours.
To address the challenge of monitoring appropriate behaviour and assessing the health status of poultry, a new intelligent monitoring system based on IoT sensors has been developed. The integration of the Internet of Things (IoT) technology within poultry farming can fundamentally transform the monitoring and management of poultry health.
This system allows for the detection and observation of chickens’ behaviour on farms, providing valuable data to operators for informed management decisions and accurate health assessments.
TAKEAWAYS
The importance of poultry farming
As early as ten years ago, the FAO highlighted in its specialised publication, “Poultry Development Review”, that the poultry sector is among the most dynamic and adaptable within livestock farming. In recent years, driven by strong global demand, this sector has seen substantial growth, expanding and becoming more globalised across countries of all income levels.
Livestock is a vital resource for the livelihood of approximately one billion of the world’s poorest people, and rural poultry plays a crucial role for many resource-limited farmers, often representing their sole asset. In low-income and food-deficit countries, rural poultry accounts for about 80% of poultry stocks and provides significant benefits to communities by:
- improving nutrition through the provision of nutrient-rich meat and eggs that are high in quality micronutrients
- generating modest income and serving as a form of savings, particularly for women, thus enhancing their financial resilience and reducing vulnerability
- producing manure that is beneficial for crops and home gardens
Beyond its economic and nutritional value, village poultry production also holds considerable socio-cultural and religious significance, recognised for its positive impact on agricultural communities.
Moreover, the United Nations’ Millennium Development Goals prioritised the reduction – and more specifically, the halving – of the number of people suffering from hunger by 2015, a formidable challenge given that in 2003, approximately 824 million people faced chronic hunger. The goals also included reducing child mortality (with an under-five mortality rate of 168 deaths per 1,000 live births in sub-Saharan Africa) and improving maternal health.
In practical terms, eggs and poultry meat play an essential role within this context, serving as excellent sources of high-quality protein, minerals, and vitamins such as B12.
Structure of poultry production
Global poultry production is divided into three main systems: large-scale commercial production, traditional village farming (especially in developing countries), and semi-commercial systems near urban areas. In developing countries, village-raised poultry constitutes a substantial part of the national poultry stock.
Village poultry plays an essential role in poverty reduction and food security, providing high-quality protein. In addition to offering nourishment, poultry products can be sold or bartered, often representing the only source of income for many families. A significant benefit of traditional farming is that women and children often manage the poultry, granting them a degree of economic independence.
Moreover, traditional farming supports genetic biodiversity, which holds potential for enhancing commercial breeds, although this potential remains under-researched and underutilised.
However, traditional poultry farming has its challenges. Despite its positive impact on poverty alleviation, there has been limited research aimed at improving its efficiency. Productivity has stagnated over the past 40 years, and competition with commercial operations is expected to increase.
The lack of infrastructure, such as essential nutrients and veterinary services, poses another obstacle, with the risk that poultry could become a reservoir for diseases, including zoonoses.
The importance of precision poultry farming
While controlling small domestic and family-run poultry farms poses challenges, more can be done in larger-scale agricultural operations through precision livestock farming (PLF). In an article published in Modern Poultry, Dr. Tom Tabler, a professor and poultry specialist at the University of Tennessee Extension Service, Department of Animal Science, discusses how PLF applies technology to monitor livestock and farming environments in real time, 24/7.
This approach provides early alerts when issues arise, enabling farmers to intervene promptly and prevent escalation.
Precision livestock farming relies on continuous data monitoring and reliable tools to support decision-making. Automated systems can weigh both animals and feed, track food intake, and use imaging and acoustic technologies to flag welfare issues or signs of disease. Sensors measure levels of ammonia, CO2, humidity, and greenhouse gases within farms.
One of PLF’s most significant contributions is optimising nutrient bioavailability. This technology helps farmers meet the daily nutritional needs of poultry without excess, reducing waste and minimising environmental impact. For commercial broiler chickens, for example, daily diets can be adjusted to align with specific nutritional requirements. Modern systems, equipped with silos and advanced technological equipment, allow precise mixing of protein and energy concentrates, ensuring optimal amounts of lysine, methionine, proteins, and energy are provided daily.
The Internet of Things supporting poultry farming
One of the most significant risks in poultry farming is the rapid spread of diseases within flocks. While the poultry industry is essential for meeting the global demand for low-fat, protein-rich foods, consumers are increasingly concerned with the quality and safety of products. Animal welfare is critical in ensuring consumer health and safety.
Reducing losses and controlling disease outbreaks can be achieved through the early identification of sick birds using technologies capable of tracking and categorising individuals based on behaviour.
The combination of the Internet of Things (IoT) and Artificial Intelligence (AI) now offers game-changing solutions for even smaller farms, with innovative approaches for early disease detection.
IoT sensors function as the eyes and ears of the poultry house, continuously monitoring the environment and detecting even subtle changes that human observation might miss. The data gathered by these sensors is processed by AI algorithms designed to recognise patterns and detect anomalies.
These early signals of illness enable timely interventions to prevent more severe issues. AI-based analyses help farmers pinpoint potential outbreaks before they escalate, ensuring prompt treatment and minimising losses.
The insights provided offer a comprehensive view of animal health and welfare, allowing farmers to make informed decisions regarding feed, housing conditions, and general management practices.
For instance, a study titled “An Approach towards IoT-Based Predictive Service for Early Detection of Diseases in Poultry Chickens”, conducted by researchers from the universities of Karachi and Kota Kinabalu in Malaysia, aimed to harness these technological advances for continuous health monitoring and tracking in poultry. IoT-based wearable devices, such as accelerometers and gyroscopes, collect data on animal movements and transmit it to the cloud for in-depth analysis. The goal is to swiftly identify signs of disease, facilitating early intervention and limiting the spread of infections.
This research utilised an IoT-based predictive framework that incorporated Generative Adversarial Networks (GANs) for generating synthetic data, achieving a 97% accuracy rate in classifying healthy and sick birds. Various machine learning algorithms, including Random Forest and Support Vector Machine, were tested to enhance prediction precision.
This technological blend, which also encompasses sound analysis and image processing for surveillance, allows for automated monitoring of poultry behaviour, essential for disease prevention.
The proposed system employs supervised learning techniques to classify birds and support proactive farm management, enhancing both sustainability and productivity. Implementing such technologies marks a step forward in precision farming that safeguards the health of both animals and consumers.
Sound analysis
Sound analysis, which examines energy distribution, frequency, and amplitude, can assess the health of chickens through their social and feeding behaviours. Multiple experiments have shown that sounds made during pecking can identify food intake with an accuracy of up to 95%.
Additionally, peak frequencies, which decrease as the chicken grows, enable monitoring of each animal’s growth rate. Using supervised neural networks and Support Vector Machine (SVM) models for sound analysis has also enabled the diagnosis of diseases like Newcastle disease and avian influenza with 100% accuracy.
Image processing
Image processing offers another cost-effective approach for evaluating poultry behaviour, health status, and weight prediction. Computer vision techniques can track movement around feeding areas and detect signs of stress or illness. Convolutional neural networks (CNNs) have achieved 99.17% accuracy in poultry monitoring.
Thermal cameras are used to check for heat stress, and posture analyses using SVM models have reached nearly 99.5% accuracy in disease classification. Continuous monitoring enables detection of infections as early as the fourth day, thus enhancing health management and overall animal welfare on farms.
IoT devices
Wearable monitoring devices, such as RFID microchips and accelerometers, are also revolutionising poultry farm management. RFID microchips allow tracking of chicken movements via signals sent to RFID readers, providing real-time observation of individual and group behaviour.
Additional sensors, such as those monitoring egg production, can identify anomalies early, including sternum fractures during egg-laying. These devices help track feeding and resting behaviour, categorising chickens as active, normal, or sick using clustering methods.
The integration of machine learning with IoT systems has improved data processing and decision-making capabilities in real-world settings. The IoT-Fog-Cloud ecosystem supports distributed processing, addressing latency and device heterogeneity issues. Generative models like CTGAN (Conditional Tabular GAN), designed to create synthetic tabular data, assist in overcoming data class imbalance, thereby enhancing machine learning model performance.
These synthetic datasets are used not only to test model effectiveness but also to ensure data privacy in health monitoring contexts. The combination of wearables and machine learning opens up new avenues for accurately monitoring and analysing poultry health and welfare, reducing the impact of infections, and improving overall farm management.
The study and its findings
In this study, data was collected over 20 weeks, monitoring 24 chickens divided into four different groups. The daily activities of each chicken, such as dustbathing, pecking, and preening, were tracked and analysed to better understand their correlations. Pecking is typically related to feeding, preening is essential for keeping feathers clean, and dustbathing helps remove parasites.
The study proposed an IoT-based predictive framework to monitor chicken movements and forecast their health in real time, addressing data imbalance issues through the use of deep generative models. Among the tested classification models, TabNet demonstrated the highest accuracy, while Decision Tree and Random Forest models also provided reliable predictions.
This research opens up new opportunities for developing IoT-based predictive services and suggests extending this work using Bayesian network models to enhance disease classification.
Data classification efforts
A similar approach was taken in a study recently published by a team of Egyptian researchers, including Mohammed Mostafa Ahmed from Sadat City University and Ehab Ezat Hassanien and Aboul Ella Hassanien from Cairo University. This study demonstrated how the use of IoT, image analysis, and other technologies enables farmers to monitor animal health in real time, gather production data, and improve overall surveillance.
The research focused on monitoring specific behaviours, such as feeding/pecking, preening, and dustbathing, which can increase when ectoparasite infections are present.
Special emphasis was placed on data collection and analysis, which included extracting statistical properties from raw data and building a complex hierarchy of features to analyse new or unknown cases.
The researchers aimed to overcome challenges associated with data management, such as data complexity, noise, and attribute loss, which can make it difficult to identify meaningful patterns during data mining processes.
One notable challenge in data classification is class imbalance, which adversely affects the accuracy and efficiency of analyses. Imbalanced datasets have significant disparities in sample numbers across categories, impacting the performance of machine learning algorithms.
To address class imbalance, various strategies were implemented, such as oversampling the minority class using the Synthetic Minority Oversampling Technique (SMOTE), which creates additional examples in the minority class through interpolation. Other sampling approaches balanced class distributions through positive, negative, or combined sampling.
The proposed system consisted of several phases: data preprocessing, feature extraction, feature selection, and behaviour detection using different classification algorithms. In this particular study, an optimised version of SMOTE was applied, enhanced by the Artificial Hummingbird Algorithm (AHA). The AHA simulates the unique flight skills and intelligent foraging strategies of hummingbirds, encompassing axial, diagonal, and omnidirectional flight patterns.
The experimental results highlighted that this optimised SMOTE approach achieved a 97% accuracy rate, outperforming other algorithms. Similar levels of accuracy were reached with the Random Forest algorithm, which attained a 98% accuracy in predicting poultry behaviours.
Glimpses of Futures
Over the past 20 years, interest in poultry and its products has grown significantly. Nearly every country in the world now has a poultry industry to meet an increasing demand, a demand that becomes particularly crucial in emerging economies where poultry farming is a vital source of sustenance for entire communities.
At the same time, consumers are showing greater awareness of sustainability, food safety, and ethical sourcing practices.
Poultry producers who adopt environmentally friendly and responsible practices, such as sustainable sourcing and animal welfare initiatives, are more likely to attract environmentally conscious consumers.
To anticipate potential future scenarios, we shall now analyse, using the STEPS matrix, the impacts that developments in these areas may have from a social, technological, economic, political, and sustainability standpoint.
S – SOCIAL: the FAO highlights in its research and publications that poultry farming plays a significant social role, especially in developing countries. Poultry provides an affordable source of high-quality protein, essential minerals, and vitamins such as B12, thereby improving nutrition in rural communities. It often represents one of the few income sources for impoverished families and is manageable even by women and children, fostering economic independence and providing a financial safety net. In many rural societies, poultry holds economic, social, and cultural value; it is used in religious ceremonies, as gifts, or in traditional practices, reinforcing communal bonds and a sense of belonging. Additionally, poultry manure serves as an organic fertiliser, enhancing crop productivity without the need for expensive chemical inputs.
T – TECHNOLOGICAL: the use of artificial intelligence (AI) in monitoring animal welfare has the potential to revolutionise poultry farming, making it more efficient and sustainable. Future applications may include increasingly autonomous and integrated systems that not only detect irregularities but suggest corrective actions in real time, enhancing operator response and reducing reaction times. The integration of AI with emerging technologies like the Internet of Things (IoT) and advanced neural networks could allow even more precise monitoring and customised animal management. Widespread and affordable implementation of these systems could benefit smaller farms as well, encouraging the broader adoption of ethical and sustainable practices. Nonetheless, challenges remain, such as cost reduction, staff training, and adapting technologies to varied production contexts. With continuous technological advancements and increased investment in research and development, AI could become an essential tool for ensuring optimal animal welfare and fostering a more responsible production system.
E – ECONOMIC: from an economic standpoint, the poultry industry undeniably faces significant challenges such as labour shortages, disease outbreaks, and animal welfare concerns, all of which negatively impact productivity and profitability. The adoption of advanced technologies like AI presents effective solutions to overcome these difficulties and meet the growing demand for poultry meat. By enhancing farm monitoring and management, AI promises to lower operational costs and boost efficiency, leading to higher profits. Traditional methods often fall short in comprehensively managing animal health, behaviour, and stress conditions, but AI-based systems can transform production by enabling continuous monitoring and timely interventions. Looking ahead, the integration of AI and IoT could lead to autonomous systems that optimise every aspect of farming, improving animal welfare and ensuring safer, more sustainable production. This would not only promote greater environmental sustainability but also positively impact the economy by reducing losses due to disease and improving overall product quality, aligning with the expectations of increasingly quality-conscious and safety-focused consumers. Furthermore, unlike large agricultural businesses, small farmers often lack access to veterinarians or advanced diagnostic tools. AI-based solutions offer an economically viable and easily accessible method for monitoring poultry health.
P – POLITICAL: discussing poultry farming requires balancing economic development with public health protection, food safety, and environmental sustainability. The intensification of farming activities has raised concerns about the impact of emissions and waste, such as the accumulation of poultry manure, which can trigger new disease strains and pose public health risks. In response, governments and regulatory bodies are enforcing stringent standards on animal welfare, food safety, and sustainable practices. Farms must comply with increasingly strict regulations, necessitating continuous staff training and impeccable documentation management to demonstrate compliance. The regulatory approach is evolving to include innovative technologies, such as digital tools for automating documentation and monitoring compliance, thereby improving management efficiency and operational transparency. In the future, the adoption of advanced genetics and breeding techniques, coupled with AI and IoT solutions, could support more sustainable management practices and better adherence to regulations.
S – SUSTAINABILITY: intensive poultry farming, if mismanaged, has a considerable negative impact on sustainability, harming both the environment and animal welfare. Poor waste management and high manure production can degrade air, water, and soil quality. Moreover, the production of feed, such as soy, contributes to deforestation in regions like Brazil, posing severe consequences for biodiversity and increasing greenhouse gas emissions. On the animal welfare front, conditions often do not allow natural behaviours, heightening stress and health issues for the animals. Therefore, it is crucial to integrate sustainable practices that respect the environment and improve animal welfare, contributing to more resilient and responsible agriculture. Early detection systems, like those described in this analysis, enable the swift isolation of sick animals, preventing the spread of disease within the farm and potentially to neighbouring agricultural areas. This is the direction needed to ensure poultry farming can play its part in achieving the United Nations Sustainable Development Goals.
