1. Introduction
One of the most significant problems in cattle is lameness. The percentage of lameness prevalence among cattle that were exposed to the same environmental risk factors was 36.8% for U.K. [
1] and relatively lower (18.7%) but also high for Greece [
2]. Lameness is the declination from normal limb motion, usually with the presence of pain.
Figure 1 presents some of the most frequent forms of infections in cattle hoofs that cause lameness.
Lameness impacts cattle welfare and health negatively; thus, a non-preventive attitude constitutes a violation of the five principles of cattle welfare [
3]. In addition to this, lameness causes great economic loss to cattle farms, as it relates to lower milk production, milk quantity rejection, and extra medical expenses and labour costs [
4] and is by far the greatest cause of death among cattle, with a rate of 20% [
5].
Modern lameness detection methods are based on visual and clinical observation [
6] and usually classify subjects on a scale from 1 to 5, according to the severity of observed lameness prevalence, where 1 is for light lameness and 5 is for severe lameness. In fact, these methods rely on human factors (lack of experience, subjectivity, and non-repeatability) and in most cases result in an understatement of the problem at an early stage. It has been mentioned that only one in four light lameness cases are detected in dairy cattle farms and only one in three cases of lameness is correctly classified as per the scale of severity [
7].
The purpose of this project is the development of a prototype of a system that is objective, reliable, and automatic and will detect lameness in cattle at an early stage of prevalence. Early detection of the problem helps to establish early treatment, successively resulting in more effective and rapid treatment, improved welfare and health for cattle, and reduced financial loss for farmers. Meanwhile, existing lameness detection methods, that are based on visual observation, fail to assess the problem in an early stage, as is graphically presented in
Figure 2.
2. Methods
The system prototype is simplistically presented in
Figure 3 and generally consists of the following described basic parts, which are force plates in a walkway arrangement, an RFID ear tag system with a receiver and tags in all cattle ears, and a PC running purpose-built software that includes a user interface and a machine learning algorithm that supports the decision-making process.
The force plates acquire the ground reaction forces of the vertical axis Z of all limbs of cattle passing along the walkway one by one. The force plates were developed specifically to be suited to the dirty, dusty, and moist environment of where they would be installed, including the need to wash away mud from the top surface, the requirement of the levelling regulation of the system on top of non-flat surfaces such as soil, the constraints in surface materials that can be used due to injury hazards of cow hoofs, and the necessity of easy replacement and availability in the market of all walkway and force plate mechanical parts.
The electronic parts of the force plates were also designed and built specifically for the project.
Figure 4 presents a typical load cell of nominal weight of 500 kg, like the ones that are used in force plate assembly, and the two types of custom PCBs—the analogue and the digital circuit, respectively. The PCBs were purpose-designed and -built in order to meet the main project challenges, which meant acquiring a number of analogue channels in a scale of magnitude of 100, with 24-bit precision, and a sampling rate of 1000 Hz, producing packages of synchronised data.
The RFID ear tag system is a widespread commercial technology that covers the needs of identification of cattle that pass along a walkway one by one. The incorporation of the RFID system is in line with the guidance of the software, so that data are gathered separately for each distinct cow and any fluctuations in already-recorded motion patterns of limbs are also identified via the functioning machine learning algorithm.
3. Results and Discussion
Early visual detection of lameness is by default a difficult task. Thus, the pattern recognition algorithm training proved to be a challenging process, as many cattle that were tagged as completely healthy by visual observation were actually in a very early stage of lameness prevalence which could have been detected by the system, but not visually. This contrast had been creating a complication in lameness prevalence recognition from the very beginning.
In order to overcome this challenge, field data acquisition lasted much longer than was initially planned. Increased measured trials served the purpose of acquiring adequate gait patterns of healthy cattle, until meeting the aim of detecting slight gait pattern changes in some of them. This process allowed for safe segregation of cattle that were healthy and those that were in a very early stage of lameness prevalence, following the creation of a proper dataset for machine learning algorithm training.
An approximate outcome of the trials is that gait pattern changes that were recognised by the system resulted in 80% actual early-stage lameness prevalence.
4. Conclusions
The main goal of this research and development project was the development of a system that detects lameness in the very early stages, long before a visual observation can be effective. Initial field data acquisition showed encouraging results, as most of the cattle that slight gait pattern changes were detected in actually developed early-stage lameness that could be visually observed.
Author Contributions
Conceptualization, G.G., D.T.; methodology, G.B. and C.K.; software, C.K. and I.L.; practical implementation, P.P., E.V., S.M., T.P., N.T., S.P. and G.T.; writing—original draft preparation, G.B.; writing—review and editing, G.B.; All authors have read and agreed to the published version of the manuscript.
Funding
This project was funded by the Hellenic General Secretariat of Research and Innovation, project code T2EDK-00758.
Institutional Review Board Statement
Not applicable.
Informed Consent Statement
Not applicable.
Data Availability Statement
Data are contained within the article.
Conflicts of Interest
The authors George Bellis, Paris Papaggelos, Evangeli Vlachogianni, Ilias Laleas and Stefanos Moustos are employed by Biomechanical Solutions (BME). The author Thanos Patas, Sokratis Poulios and Nikos Tzioumakis are employed by Polytech S.A. George Tsiogkas is employed by TSIOGKAS FARM. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as potential conflicts of interest.
References
- Barker, Z.E.; Leach, K.A.; Whay, H.R.; Bell, N.J.; Main, D.C. Assessment of lameness prevalence and associated risk factors in dairy herds in England and Wales. J. Dairy Sci. 2010, 93, 932–941. [Google Scholar] [CrossRef] [PubMed]
- Katsoulos, P.D.; Christodoulopoulos, G. Prevalence of lameness and of associated claw disorders in Greek dairy cattle industry. Livest. Sci. 2008, 122, 354–358. [Google Scholar] [CrossRef]
- Webster, A.J. Farm animal welfare: The five freedoms and the free market. Vet. J. 2001, 161, 229–237. [Google Scholar] [CrossRef] [PubMed]
- Bruijnis, M.R.; Hogeveen, H.; Stassen, E.N. Measures to improve dairy cow foot health: Consequences for farmer income and dairy cow welfare. J. Dairy Sci. 2010, 93, 2419–2432. [Google Scholar] [CrossRef] [PubMed]
- NAHMS. Part I: Reference of Dairy Cattle Health and Management Practices in the United States, 2007. In USDA Animal Health Report; USDA: Washington, DC, USA, 2007; p. 95. [Google Scholar]
- Sprecher, D.J.; Hostetler, D.E.; Kaneene, J.B. A lameness scoring system that uses posture and gait to predict dairy cattle reproductive performance. Theriogenology 1997, 47, 1179–1187. [Google Scholar] [CrossRef] [PubMed]
- Espejo, L.; Endres, M.I.; Salfer, J.A. J. Scientific report on the effects of farming systems on dairy cow welfare and disease. Dairy Sci. 2006, 89, 3052–3058. [Google Scholar] [CrossRef] [PubMed]
| Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).