Automated Tracking Systems for Assessment of Farmed Poultry: History
View Latest Version
Please note this is an old version of this entry, which may differ significantly from the current revision.
Subjects: Agriculture, Dairy & Animal Science
Contributor:
Suresh Neethirajan

One of the most commonly farmed livestock is poultry and their significance is felt on a global scale. Current poultry farming practices result in the premature death and rejection of billions of chickens on an annual basis before they are processed for meat. This loss of life is concerning regarding animal welfare, agricultural efficiency, and economic impacts. The best way to prevent these losses is through the individualistic and/or group level assessment of animals on a continuous basis. On large-scale farms, such attention to detail was generally considered to be inaccurate and inefficient, but with the integration of artificial intelligence (AI)-assisted technology individualised, and per-herd assessments of livestock became possible and accurate.

  • poultry behaviour
  • target tracking
  • deep learning
  • precision livestock farming
  • poultry production systems

1. Introduction

Today’s demands for increased livestock production result in various challenges for the animals they pertain to. A balance is needed between the quantity and quality of poultry production. However, farmers must worry about maximizing profits, a need that has promoted a prioritization of production over aspects such as welfare. Flock size and growth are commonly maximised in minimal spaces to offset low margins for farmers. Societal pressures towards sustainability also influence minimal inputs for poultry farming aspects such as land, labour, and natural resource usage. These efforts may lead to increased poultry production with decreased production time and resource usage, but they have also unintentionally led to the proliferation of harmful genetic alterations and the increase in associated diseases. The solution to these complex agricultural needs is to assist farmers with automated surveillance of the animals. Through the continuous and automated monitoring of animals, farmers are able to detect welfare and production concerns in a manner that is both quick and reliable. With the integration of modern technological advances, poultry farming has the opportunity to grow in terms of production quantity and animal care quality with minimal added expense.
On a global scale, 69 billion chickens are raised for meat production every year [1], but not all of them make it to people’s plates. In the UK alone, over a 3-year period between 2016 and 2019, about 61 million chickens were rejected for human consumption due to defects and diseases in slaughterhouses [2]. The threshold set by the Spanish agency for Food Safety and Nutrition on the rejection of chickens before processing in slaughterhouses, is 2% per annum [3]. In a Turkish study [4], it was estimated that approximately 0.4% of broiler chickens are dead-on-arrival before the process of slaughtering under commercial conditions. Globally, several million chickens do not survive the rearing process, and are possibly rejected at the slaughterhouse because of illnesses, scratches, bruises, and other signs of welfare failures. Considering the difference between food accessibility and hunger for some people, and for farmers, rejection of chickens at slaughterhouses can be a great source of profit loss. This statistic also makes a huge difference for the animals, as it suggests that millions of chickens bred for meat suffer from unmanaged, painful, and possibly deadly medical (pathological) conditions each year. With better diagnostics and agricultural management, fewer resources would be wasted, more chickens could be produced, and less suffering would be faced by these animals.
There is a need to increase agricultural capabilities to detect anomalies in chicken behaviour and health, and thereby welfare, without increasing a need for manual labour, and for that, automated systems are needed. Automated systems have been studied and proven to be capable of accurately collecting data related to the following needs:
  • Individual tracking, even in large groups of animals that are condensed in a confined space.
  • Phenotype assessment and analysis for the non-invasive understanding of genotypes, which are important for resilient breeding methods.
  • Identification of the needs of individual animals in relation to welfare.
  • Continuous data-collecting capabilities that cannot be replicated by humans.
  • Assessment of activity and changes on a flock level.
  • Early direction of behavioural and physical shifts in comparison to past flocks.
  • Analysis of nuances related to welfare-focused farming, such as the preferences in light intensity for individuals or groups of agricultural birds.
  • Long-range use for the non-disruptive observation of fearful and free-range livestock.
  • Bone fracture assessment for immediate intervention.

2. Need for Automated Poultry Surveillance

Poultry and eggs are a major source of dietary protein for people across the globe [5]. As a result, these animal food sources must be produced in a way that minimises their cost and maximises their availability if poultry is going to remain a major food source as the human population continues to grow. The profitability and productivity of commercial poultry farming depend on regular monitoring of the birds, and minimal human labour to maintain its affordability [6]. Recently, in June 2021, the European Commission set a goal to phase out the use of cages for farmed animals by the year 2027. This created a need for redesigning the poultry housing systems using an enhanced understanding of the range of behaviour and locomotion of laying hens and broilers. The modern solution to this issue can be found in technological advances that are both growing in accuracy and decreasing in price. Introducing artificial intelligence (AI) in poultry farming and management has the potential to improve multiple aspects of the industry. With the ability to accumulate data that triggers informed actions, this technology has the potential to improve animal welfare, minimise the spread of disease, improve breeding standards, and reduce waste [7]. With so many promising implications, it should be no surprise that automated poultry surveillance is receiving plenty of attention in the realm of research.
Many of the production economics of poultry farms depend upon visually accessible aspects, such as the size, weight, and appearance of poultry eggs and meat. This is precisely why computerised, video-based systems are becoming a popular real-time automated tool for poultry processing. It is praised as a non-intrusive and non-invasive option for flock assessment that seriously reduces, and even eliminates, events of unnecessary stress, which are commonly caused by human observation. This aspect makes it a beneficial tool for presenting a wide range of data on animals within a flock and for the sorting and grading of poultry-related products [8].
The detection and prediction of abnormal behaviour and poultry diseases can be accurately managed using automated tracking platforms [9]. These systems are capable of recording data and analysing poultry farming focuses, including flock density, flock floor distribution, heat stress, feeding and drinking behaviours, optical flow patterns, activity, and the detection and counting of laying hens [10].
With the continuous focus on enhancing the welfare of chickens and mounting of new evidence towards chicken cognition and emotions [11][12][13][14], there is a dire need for considering the individual needs of chickens. This demands a change in poultry management from a flock-level perspective to an individual bird’s needs. AI technologies enable the identification of individual broilers [15], or laying hens, among hundreds of birds via videos irrespective of similar sizes, shapes, and colours of the feathers. This unique ability enables automated monitoring systems to offer welfare-centred intervention decisions. This technology also permits the use of robust detection of eggs, which will make the tedious and time-consuming task of floor egg collection easier for farmers [16]. Behavioural issues in group-housed turkeys, such as cannibalism, can be rapidly detected and consequently addressed through deep learning techniques [17]. Some systems are even developed to find the location of chickens on a farm for simplified assessment and treatment by farmers [18].

3. Artificial Intelligence in Poultry Monitoring

Currently, the role of AI in various aspects of society is becoming increasingly obvious to the public, so it is no surprise that this method of management is making its way into food production systems. Computerised monitoring technology promises to fulfil the growing criteria for improved poultry production management, including conversion of the feed ratios and profitability [19].

3.1. Computer Vision Technology

Welfare factors related to farm management can be better understood by monitoring poultries’ natural processes and responses. Computer video systems can assess and determine a wide range of data at a time, including housing management, weight measurement, behaviour, detection of diseases, slaughtering processes, egg quality, and carcass quality checking [6][19].

3.2. Components of Machine Vision for Poultry Tracking

A computer tracking and monitoring technology for various poultry processes consists of two main parts [19]. The first of which is the hardware. Hardware is recognised as the physical components of these systems, which include computing instrumentation, data-acquisition hardware, lighting, wiring, and other tangible components. Advancements in hardware are the primary reason for the development of vision technology in poultry farming. There are three key components of functional hardware in computer vision systems:
  • Cameras and various lenses suited to the environment and assigned task.
  • Lighting units.
  • Mounts that allow for the full view of an observed farming space, without interrupting normal poultry functions.
The second component is software. This includes the programs and other operating information needed for the hardware to perform its specified function. Software is specially designed for data acquisition and data analysis, especially in the field of agriculture, where it must be altered to suit the species of interest. A data-acquisition software system performs its role in the storage and selection of good quality images (or videos) that are produced by the cameras. Data-analysis platforms help in the processing of images using algorithms suited to the data and research needs [19].

3.3. Types

Computerised, visual-analysis systems exist in two major forms, which are identified as machine learning-based systems and deep learning-based systems.

3.3.1. Machine Learning-Based Systems

Machine learning-based computer vision systems follow specific image analysis protocols and a specially designed algorithm. The basic workflow of a machine learning-based system for poultry monitoring is as follows [6]:
  • Acquisition of image: focused on depth or RGB images.
  • Pre-processing of image: normalization, resizing, and colour-space transformation.
  • Region of interest (ROI) segmentation: background removal or subtraction, ellipse modelling, and other focus-enhancing alterations.
  • Features extraction: optical flow meter, locomotor, and morphological features.
  • Modelling: machine learning-based algorithms.
  • Regression: monitoring of bioprocesses and bioresponses.

3.3.2. Deep Learning-Based Systems

Deep learning-based computer vision systems are a recent advancement in automatic livestock observation, which simplify associated data processing. Various processes of machine learning systems, such as segmentation, feature extraction, and selection, are time consuming and subjective laborious tasks. It is also important to note that the performance of these algorithms must change in relation to sensor sensitivity [6]. The most important feature of a deep learning system is its ability to directly process the image, thus eliminating older laborious processes by using a deep neural network (DNN). These deep learning models generally provide higher accuracy than machine learning, making them better suited to the observation of large flocks [6]. Deep learning systems also solve the common complications with multiple object tracking when using a single camera. This is a revolutionary advancement for researchers and farmers since it minimises equipment costs [20].

3.4. Applications

Computer vision systems can be adapted to suit a variety of applications, including the following, which are geared towards poultry farming [6]:
  • Recognition and identification of images: checking for the presence of poultry in every image.
  • Detection of object: locating the exact position of poultry in every image.
  • Classification of image: classifying the identified poultry as sick or absent.
  • Segmentation: identifying the watering and feeding structures in every image.
  • Recognition of specific objects: noting the behaviours exhibited by members of a flock.

4. Milestones in the Field of Automated Poultry Tracking

Different techniques and methods have been developed in the poultry industry to ensure improved production rates [21]. Among the poultry species farmed globally, chickens are the most common and are produced in the largest quantities. Recent advancements of automated poultry monitoring tools is shown in Table 1.
Table 1. An overview of current research advancements of automated poultry monitoring tools.

Applications

Used Tools and Platforms

Solved Poultry Problems

References

Counting of individual broilers

Camera, TBroiler

Abnormal behaviour; patterns

[9]

Broiler movement

Camera

Various among individuals

[7]

Productivity in broilers

Camera, sensors

Advance treatments for healthy growth

[22]

Behaviour at different feeders

Camera

Choice of feeder design

[23]

Detection of disease

Camera, Improved Feature Fusion Single Shot Multibox Detector (IFSSD)

Outbreak prevention

[24]

Sick broiler assessment

Camera

Disease management

[25]

Keel bone fracture

Infrared receivers

Timely treatments

[26]

Laying hen light preference

Camera, tracking algorithm

Layer detection in cages

[27]

Pecking in turkeys

Camera, microphone, and metallic balls

Assessment of cannibalism

[17]

Tracking in pigs

Camera, sensors

Individual behaviour

[28]

Poultry movement and range behaviour assessment

AI-based algorithms and cameras (multi-object tracking algorithm and single shot multibox detector algorithm)

Group-level poultry movement

[29]

Turkey behaviour identification

Video analytics, multi-object tracking

Turkey health status and behaviour identification

[30]

Thermal comfort of poultry birds

Camera, computer vision

Unrest index and locomotion

[31]

Laying hen behaviour

Camera, AI algorithms

Cluster and unrest behaviour

[32]

Adult free-range hen behaviour investigation

Camera, sensors, AI algorithms

Range use and fearfulness behaviour

[33]

Stocking density of broilers

AI algorithms, machine vision cameras

Relationship between stocking density and feeding/drinking of broilers

[34][35][36]

This entry is adapted from the peer-reviewed paper 10.3390/ani12030232

References

  1. Ritchie, H.; Roser, M.; Meat and Dairy Production. Our World in Data. 2017. Available online: https://ourworldindata.org/meat-production#number-of-animals-slaughtere (accessed on 1 November 2021).
  2. Food Standards Agency, UK. 2019. Available online: https://data.food.gov.uk/catalog/datasets/599f1f49-f6b9-4e8e-851a-6710ef8f7a32 (accessed on 1 November 2021).
  3. Spanish Agency for Food Safety and Nutrition. 2015. Available online: http://www.aecosan.msssi.gob.es/AECOSAN/docs/documentos/seguridad_alimentaria/gestion_riesgos/Procedimiento_deteccion_post_mortem_bienestar_pollos_engorde.pdf (accessed on 1 November 2021).
  4. Bulent, T.E.K.E. Survey on dead on arrival of broiler chickens under commercial transport conditions. Large Anim. Rev. 2019, 25, 237–241.
  5. Lin, C.Y.; Hsieh, K.W.; Tsai, Y.C.; Kuo, Y.F. Automatic monitoring of chicken movement and drinking time using convolutional neural networks. Trans. ASABE 2020, 63, 2029–2038.
  6. Okinda, C.; Nyalala, I.; Korohou, T.; Okinda, C.; Wang, J.; Achieng, T.; Wamalwa, P.; Mang, T.; Shen, M. A review on computer vision systems in monitoring of poultry: A welfare perspective. Artif. Intell. Agric. 2020, 4, 184–208.
  7. Gebhardt-Henrich, S.G.; Stratmann, A.; Dawkins, M.S. Groups and Individuals: Optical Flow Patterns of Broiler Chicken Flocks Are Correlated with the Behavior of Individual Birds. Animals 2021, 11, 568.
  8. Nyalala, I.; Okinda, C.; Kunjie, C.; Korohou, T.; Nyalala, L.; Chao, Q. Weight and volume estimation of poultry and products based on computer vision systems: A review. Poult. Sci. 2021, 100, 101072.
  9. Fang, C.; Huang, J.; Cuan, K.; Zhuang, X.; Zhang, T. Comparative study on poultry target tracking algorithms based on a deep regression network. Biosyst. Eng. 2020, 190, 176–183.
  10. Geffen, O.; Yitzhaky, Y.; Barchilon, N.; Druyan, S.; Halachmi, I. A machine vision system to detect and count laying hens in battery cages. Animal 2020, 14, 2628–2634.
  11. Ferreira, V.H.B.; Guesdon, V.; Calandreau, L. How can the research on chicken cognition improve chicken welfare: A perspective review? Worlds Poult. Sci. J. 2021, 77, 679–698.
  12. Marino, L. Thinking chickens: A review of cognition, emotion, and behavior in the domestic chicken. Anim. Cogn. 2017, 20, 127–147.
  13. Greenacre, C.B.; Morishita, T.Y. Backyard Poultry Medicine and Surgery: A Guide for Veterinary Practitioners; John Wiley & Sons: Hoboken, NJ, USA, 2021.
  14. Neethirajan, S. The use of artificial intelligence in assessing affective states in livestock. Front. Vet. Sci. 2021, 8, 879.
  15. Neethirajan, S. ChickTrack—A quantitative tracking tool for measuring chicken activity. TechRxiv 2021, 1–19.
  16. Li, G.; Xu, Y.; Zhao, Y.; Du, Q.; Huang, Y. Evaluating convolutional neural networks for cage-free floor egg detection. Sensors 2020, 20, 332.
  17. Nasirahmadi, A.; Gonzalez, J.; Sturm, B.; Hensel, O.; Knierim, U. Pecking activity detection in group-housed turkeys using acoustic data and a deep learning technique. Biosyst. Eng. 2020, 194, 40–48.
  18. Rufener, C.; Berezowski, J.; Sousa, F.M.; Abreu, Y.; Asher, L.; Toscano, M.J. Finding hens in a haystack: Consistency of movement patterns within and across individual laying hens maintained in large groups. Sci. Rep. 2018, 8, 12303.
  19. Abd Aziz, N.S.N.; Daud, S.M.; Dziyauddin, R.A.; Adam, M.Z.; Azizan, A. A Review on Computer Vision Technology for Monitoring Poultry Farm–Application, Hardware, and Software. IEEE Access 2020, 9, 12431–12445.
  20. Ciaparrone, G.; Sánchez, F.L.; Tabik, S.; Troiano, L.; Tagliaferri, R.; Herrera, F. Deep learning in video multi-object tracking: A survey. Neurocomputing 2020, 381, 61–88.
  21. Ichiura, S.; Mori, T.; Horiguchi, K.I.; Katahira, M. Exploring IoT based broiler chicken management technology. In Proceedings of the 7th TAE 2019, Prague, Czech Republic, 17–20 September 2019; pp. 205–211.
  22. Abdoli, A.; Alaee, S.; Imani, S.; Murillo, A.; Gerry, A.; Hickle, L.; Keogh, E. Fitbit for chickens? Time series data mining can increase the productivity of poultry farms. In Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, San Diego, CA, USA, 6–10 July 2020; pp. 3328–3336.
  23. Taki, M.; Ajabshirchi, Y.; Ranjbar, S.F.; Rohani, A.; Matloobi, M. Modeling and experimental validation of heat transfer and energy consumption in an innovative greenhouse structure. Inf. Process. Agric. 2016, 3, 157–174.
  24. Dawkins, M.S.; Roberts, S.J.; Cain, R.J.; Nickson, T.; Donnelly, C.A. Early warning of footpad dermatitis and hockburn in broiler chicken flocks using optical flow, bodyweight and water consumption. Vet. Rec. 2017, 180, 499.
  25. Zhuang, X.; Bi, M.; Guo, J.; Wu, S.; Zhang, T. Development of an early warning algorithm to detect sick broilers. Comput. Electron. Agric. 2018, 144, 102–113.
  26. Rufener, C.; Abreu, Y.; Asher, L.; Berezowski, J.A.; Sousa, F.M.; Stratmann, A.; Toscano, M.J. Keel bone fractures are associated with individual mobility of laying hens in an aviary system. Appl. Anim. Behav. Sci. 2019, 217, 48–56.
  27. Li, G.; Li, B.; Shi, Z.; Zhao, Y.; Ma, H. Design and evaluation of a lighting preference test system for laying hens. Comput. Electron. Agric. 2018, 147, 118–125.
  28. Zhang, L.; Gray, H.; Ye, X.; Collins, L.; Allinson, N. Automatic individual pig detection and tracking in pig farms. Sensors 2019, 19, 1188.
  29. Yang, Q.; Xiao, D. A review of video-based pig behavior recognition. Appl. Anim. Behav. Sci. 2020, 233, 105146.
  30. Ju, S.; Erasmus, M.A.; Zhu, F.; Reibman, A.R. Turkey behavior identification using video analytics and object tracking. In Proceedings of the 2021 IEEE International Conference on Image Processing (ICIP), Anchorage, AK, USA, 19–22 September 2021; pp. 1219–1223.
  31. Del Valle, J.E.; Pereira, D.F.; Neto, M.M.; Gabriel Filho, L.R.A.; Salgado, D.D.A. Unrest index for estimating thermal comfort of poultry birds (Gallus gallus domesticus) using computer vision techniques. Biosyst. Eng. 2021, 206, 123–134.
  32. Fernandes, A.M.; de Lucca Sartori, D.; de Oliveira Morais, F.J.; Salgado, D.D.A.; Pereira, D.F. Analysis of Cluster and Unrest Behaviors of Laying Hens Housed under Different Thermal Conditions and Light Wave Length. Animals 2021, 11, 2017.
  33. Bari, M.S.; Allen, S.S.; Mesken, J.; Cohen-Barnhouse, A.M.; Campbell, D.L. Relationship between Range Use and Fearfulness in Free-Range Hens from Different Rearing Enrichments. Animals 2021, 11, 300.
  34. Kim, H.; Kang, H.; Kang, B.; Kim, C. Analysis of Behavioral Characteristics of Broilers by Feeding, Drinking, and Resting Spaces according to Stocking Density using Image Analysis Technique. J. Korea Acad. Ind. Coop. Soc. 2020, 21, 558–569.
  35. Lee, D.H.; Kim, A.K.; Choi, C.H.; Kim, Y.J. Study on image-based flock density evaluation of broiler chicks. J. Korea Inst. Inf. Electron. Commun. Technol. 2019, 12, 373–379.
  36. Li, G.; Hui, X.; Chen, Z.; Chesser, G.D., Jr.; Zhao, Y. Development and evaluation of a method to detect broilers continuously walking around feeder as an indication of restricted feeding behaviors. Comput. Electron. Agric. 2021, 181, 105982.
More
© Text is available under the terms and conditions of the Creative Commons Attribution (CC BY) license; additional terms may apply. By using this site, you agree to the Terms and Conditions and Privacy Policy.
This entry is offline, you can click here to edit this entry!
ScholarVision Creations
Feedback