Application of precision livestock farming technologies in swine welfare management: What is possible today?
It is estimated that by the year 2050 the world population will be over 9 billion people and food production will need to increase up to 60% more to meet demand (FAO 2009). Therefore, livestock production would likely intensify increasing animal density and lowering the stockperson per animal ratio. This will result in less time available to monitor and manage individual animals properly, jeopardizing animal health and welfare. Currently, there is a growing interest to automate swine welfare assessment using precision livestock farming (PLF) which increases the farmer’s ability to keep contact with individual animals in the growing livestock production intensification.
This article will:
- Explain concept of PLF
- Discuss engineering advances in PLF
- Describe on-farm swine welfare assessment
- Provide examples of automatic swine welfare assessment
What is Precision Livestock Farming?
In modern livestock production systems, farmers need reliable and affordable technologies to assist them in daily management tasks which guarantees accurate and continuous individual animal monitoring. Application of engineering techniques in livestock farming to monitor, model, and manage animal production is called precision livestock farming (PLF). Its aim is to increase the farmer’s ability to keep contact with individual animals in the growing livestock production intensification (Vranken and Berckmans 2017).
Engineering Advances in PLF
Quantifying the parameters and procedures in an assessment methodology is a crucial and beneficial tool. Quantification provides the capability to analyze, compare and interpret the data for a better understanding of the situation and more precise decision making. Engineering tools and techniques such as sensors, metrology systems, data analysis, and machine learning methods have been a substantial help in many applications such as precision agriculture, livestock farming, and animal behavior.
Data management plan
The first step is to have a formal document that outlines data collection during a research project and data handling after project completion. This step aims to consider data management, preservation, and analysis before the project begins.
To quantify the problems, physical parameters (e.g. temperature, milk yield, body weight, etc.) relevant to the traits of interest need to be measured. Based on the physical parameter type to be measured, a variety of different sensors can be designed and implemented.
Image processing techniques
One of the specific types of physical parameters that can be measured or recorded is the object or system’s image. Each image consists of many smaller pieces called pixels. Each pixel contains a small amount of data from the whole picture. In other words, an image is a matrix of data, and hence can be analyzed and evaluated. The procedure of data analysis for an image is called image processing which provides information about the topographies in an image or enhances its resolution or clarity.
Complementary to the physical parameter measurement is the data analysis. A variety of statistical, mathematical, and analytical methods can be used to extract information from collected data. These techniques can be applied or developed based on the available theories for a particular application.
An algorithm is a set of specific instructions or a sequence of calculations that guides a computer through an activity such as processing data to create a graph or summary and allows automation of the activity. Developing a machine learning and data processing algorithm is the key to understanding and implementing the experimental data. For successful algorithm development, one must have a comprehensive understanding of data type (such as animals sound/voice, body gesture, etc.) and data trends such as changes over time (i.e. time-domain) or changes over the range of frequencies (i.e. frequency-domain). Next, in any algorithm is the evaluation of data based on “features”. Good features are the backbone of any machine learning and data processing algorithm. Appropriate feature creation and extraction often needs domain knowledge, creativity, and lots of time. Next step which we mostly tend to care about is, creating the algorithm modules and modeling. After creating the algorithm model comes the algorithm experimentation and validation.
On-farm Swine Welfare Assessment
There is no consensus on a unique definition of animal welfare within or across species where applicable. Definitions vary depending on different factors such as cultural, scientific, religious and political background (Swanson 1995). There are several proposed approaches to scientifically assess animal welfare (e.g. feelings-based approach, natural living, the five freedoms, among others) and each tends to emphasize different measurements. Curtis (2007) proposed that objective animal-based measures would provide a more accurate approach to assess farm animal welfare state because “what cannot be measured cannot be managed” (Curtis 2007). Animal-based measures address the actual welfare state of animals in terms of their behavior, fearfulness, health, or physical condition. Over the years, a series of pig welfare assessment protocols such as the Welfare Quality Protocol®, AssureWel, the Danish Animal Welfare Index and the Swine Welfare Assurance Program have been developed. They include several animal-based indicators that could be used to measure aspects of on-farm swine welfare by farmers, veterinarians, advisory service personnel, among others. A summary of the animal-based measures to assess pig welfare included in the Welfare Quality Protocol® is presented in Table 1.
|Welfare criteria||Animal-based measure||Assessment criteria|
of prolonged hunger
|Body condition score||0 – Pig with a good body condition;
2 – Lean pigs.
|0 – It takes firm pressure with the palm of the hand of the assessor to feel the hip bones and backbone; 1 – The hip bones and backbone are easily felt without any pressure on the palms, or the sow appears visually obese and it is impossible to feel the hip bones and backbone even by pushing down with a single finger;
2 – The sow appears visually very thin, with hips and backbone very prominent.
|Thermal comfort||Panting and/or huddling||0 – No pigs in the pen/group observed are panting and/or huddling;
1 – Up to 20% of pigs in the pen observed are panting and/or huddling;
2 – More than 20% of pigs in the pen observed are panting and/or huddling.
|0 – No panting and/or huddling;
2 – Evidence of panting and/or huddling.
|Absence of injury||Lameness||0 – Normal gait or difficulty in walking, but still using all legs; swagger of caudal body while walking; shortened stride;
1 – Severely lame, minimum weight–bearing on the affected limb;
2 – No weight–bearing on the affected limb, or not able to walk.
|0 – Normal gait, or the animal has difficulties walking but is still using all its legs, the stride may be shortened and/or there may be a swagger of the caudal part of the body when walking;
1 – The animal is severely lame; it put a minimum of weight on the affected limb (asymmetric walking);
2 – There is no weight–bearing on the affected limb, or the animal is unable to walk.
|Tail-biting||0 – No evidence of tail biting or indication of superficial biting along the length of the tail, but no evidence of fresh blood or of any swelling (red areas on the tail are not considered as wounds unless associated with fresh blood);
2 – Fresh blood is visible on the tail; there is evidence of some swelling and infection; part of the tail tissue is missing and a crust has formed.
|Ear-biting||0= unmarked ears;
1 or 2= scratched ears;
3= a new ear hematoma;
4= 1 rip through the ear;
5= > 1 ear rip.
|Absence of disease||Coughing and/or sneezing||Average frequency of coughs and/or sneeze per animal for 5 minutes.||0 – No evidence of coughing and/or sneezing;
2 – Evidence of coughing and/or sneezing.
|Pumping (labored breathing)||0 – Percentage of pigs with no evidence of labored breathing;
2 – Percentage of pigs with evidence of labored breathing.
|0 – No pigs in the litter display evidence of labored breathing;
1 – One pig in the litter displays evidence of labored breathing;
2 – More than one pig in the litter is displaying evidence of labored breathing.
|Expression of social behavior||Aggressive behavior||Number of sample points during which aggressive behavior was observed.||Number of animals showing aggressive behaviors.|
Table 1: Animal-based measures for pig welfare assessment, Adapted from (Welfare Quality 2009)
Automated Swine Welfare Assessment
It is important to choose the appropriate technology with correct application to measure animal welfare in order to successfully automate animal health and welfare monitoring (Schon and Meiering 1987). Additionally, the method selected for handling and storing the large amounts of data depends on the size and nature of data, the cost of storage and access over time, the time it takes to transfer the data, how the data will be used, and any privacy concerns (Hart et al. 2016). Sensors that are commercially available and technically feasible for detecting absence of prolonged hunger (e.g. body condition score – BCS), thermal comfort (e.g. panting, huddling, and body surface temperature), absence of injury (e.g. lameness, tail and ear-biting), absence of disease (e.g. coughing, sneezing, pumping, or labored breathing) and expression of social behavior (e.g. aggressive behavior) are discussed in the next section.
Absence of prolonged hunger
The authors did not find any published studies on automatic BCS assessment in swine. However, Bayer developed an app to automatically score BCS (for more information please visit https://animalhealth.bayer.com/sowdition). It is possible that the lack of studies regarding automated scoring of BCS is due to the fact that BCS is mainly used for the breeding herd and there are other more accurate and easy to use methods to assess body condition in sows such as the use of flank-to-flank measurements. Nonetheless, there are several studies regarding automatic estimation of body weight in grow-finish swine. Condotta et al. (2018) suggested that the Kinect depth sensor and image processing would be a reasonable approach to continuously monitor pig mass (Condotta et al. 2018). Kashiha et al. (2014) found that video imaging of fattening pigs was promising for real-time weight and growth monitoring with an accuracy of 97.5% at group level and 96.2% at individual level (Kashiha et al. 2014a). Kollis et al. (2007) designed and implemented a system using off-the-shelf hardware to estimate pigs’ weight with an average error of 5% (Kollis et al. 2007). Kongsro et al. (2014) presented a prototype for pig weighing based on the Microsoft Kinect camera technology, utilizing the infrared depth map images and estimated pig’s weight with an error estimate of 4-5% of mean weight (Kongsro 2014). Wang et al. (2018) developed a novel portable and automatic measurement system for pig body size using two depth cameras. The average relative errors for body width, hip width, and body height were 10.30%, 5.87% and 7.01% respectively, which demonstrates the efficacy of the proposed system (Wang et al. 2018). Wu et al. (2004) developed a stereo imaging system with six high-resolution cameras (3032 × 2028 pixels) and three ﬂash units to capture the three-dimensional shapes of live pigs with good quality (Wu et al. 2004). There are already commercial systems using camera technology to estimate body weight in pigs in real time (e.g. ProGrow from Skov; for more information please visit https://www.skov.com/en/news/Pages/ Camera-weighing-for-weaners-and-porkers.aspx).
Take home message: Video imaging of pigs seems promising for real-time on-farm weight and growth monitoring. However, further study needs to develop speciﬁc algorithms to account for gender and genotype effect on body surface area and body weight as well as other traits which affect the model parameters for weight estimation.
Shao (1997) evaluated the feasibility of automatically classifying swine comfort behavior via image processing/analysis and neural networks. He established algorithms of machine vision techniques to provide a foundation for identification of the thermal comfort behavior of the pigs (Shao 1997). Shao and Xin (2008) developed a real-time image processing system to detect thermal comfort state of group-housed pigs based on their resting behavioral patterns (Shao and Xin 2008). Xin (1999) explored an automated image analysis system to assess the thermal comfort of swine and to improve the accuracy of the postural image-based assessment of the thermal comfort state (Xin 1999). Xin and Shao (2002) investigated interactive assessment and control of swine thermal comfort by real-time computer image analysis of pig resting patterns (Xin and Shao 2002).
Take home message: Real-time image processing prototype systems successfully detect pigs’ motion and classiﬁes the thermal comfort state of the pigs into cold, comfortable, or warm category. Nonetheless further testing and reﬁnement of the system are needed for commercial production settings.
Absence of injury
Lameness. Assessment of lameness in sows could also involve an analysis of postural time budgets since lame animals have more difficulty lying down and their ability to change postures may be impacted. Ringgenberg and Bergeron (2010) developed and validated an automated method of detecting postures and stepping behaviour in sows using accelerometers. They observed that sensitivity for detecting postural changes was high (>91%) for standing, ventral, and lateral lying, but only 50% for sitting position. However, the authors suggested that individual calibration would increase the accelerometer performance for sitting and it can be used to detect postures and the number of hind limb steps in sows. To date, accelerometers are not used in a commercial setting for lameness detection (Ringgenberg and Bergeron 2010). Stavrakakis et al. (2015) distinguished non-lame from lame pigs by video recording and tracking neck region elevation during walking (Stavrakakis et al. 2015a). In another study Stavrakakis et al. (2015) showed that step to stride length ratio could identify up to 74% of the pigs which subsequently developed lameness (Stavrakakis et al. 2015b). Lind et al. (2005) and Kongsro (2013) found that a digital video-based tracking system for automatically tracking pigs’ locomotor behavior was inexpensive, highly reliable and accurate (Lind et al. 2005; Kongsro 2013). Kashiha et al. (2014) showed that pig’s locomotion in a group can be determined using image analysis with an accuracy of 89.8% (Kashiha et al. 2014b). Sun et al. (2011) developed a microcomputer-based force plate system for measuring lameness in sows. The authors reported that the system was able to identify sow lameness by measuring separately the weight of each leg (Sun et al. 2011). Meijer et al. (2014) recommended a pressure mat as an objective, highly replicable tool to analyze and quantify porcine gait (Meijer et al. 2014). Pluym et al. (2013) developed a detection system, based on force stance variables derived from force plate analysis and visual stance variables derived from image processing. She indicated that this system has very promising potential to support future objective research in sow lameness (Pluym et al. 2013). McNeil et al. (2019) reported that force plate measurements could detect lameness almost 5 days earlier than using a visual scoring system (McNeil et al. 2019).
Take home message: Accelerometer, digital video-based tracking system, pressure mat and force plate are accurate, reliable methods to assess lameness in sows saving manpower on farm. However, more measurements and especially delicate software are needed to have a robust system for a commercial farm environment.
Tail and ear-biting. Blömke et al. (2020) developed a camera-based system for tail biting assessment with accuracy of 99.5%, sensitivity of 77.8%, and specificity of 99.7% (Blömke et al. 2020). D’Eath et al. (2018) demonstrated the potential for a three-dimensional machine vision system to automate tail posture detection and provide early warning of tail biting on farm with accuracy of 73.9%, sensitivity of 88.4%, and specificity of 66.8% (D’Eath et al. 2018). Diana et al. (2019) developed an ethogram for ear biter and bitten pig using video images. The authors identified six biter pig behaviors and seven bitten pig behaviors with a precision of 83.2% suggesting their potential to develop a PLF tool to monitor ear biting behavior (Diana et al. 2019).
Take home message: Three-dimensional machine vision system has the potential to automate tail posture detection and provide early warning of tail and ear-biting on farm.
Absence of disease
Chung et al. (2013) proposed an efficient data mining solution for the detection and recognition of pig wasting diseases using sound data in audio surveillance systems. The results showed that the method is cheap and 94% accurate (Chung et al. 2013). Vandermeulen et al. (2015) developed a detection method with 72% sensitivity, 91% specificity and 83% precision to detect screams based on sound features with physical meaning and explicit rules (Vandermeulen et al. 2015). Ferrari et al. (2008) described acoustic features of cough sounds originating from a lung infection to develop a real-time cough classiﬁcation algorithm, based on sound feature analysis, to acquire a monitoring system for automatic and continuous infection in pig houses (Ferrari et al. 2008). Exadaktylos et al. (2008) proposed a real-time method for identifying sick pig cough sounds. An 85% overall correct classiﬁcation ratio was achieved with 82% of the sick cough sounds being correctly identiﬁed (Exadaktylos et al. 2008). Finger et al. (2014) used a microphone-operated system (Pig Cough Monitor) to detect coughing in a fattening pig herd. The system enabled an objective measurement of any deviation from normal respiratory pattern and could be clearly linked to PRRSV infection (Finger et al. 2014). Genzow et al. (2014) used pen-based oral fluid samplings in combination with monitoring cough in fattening pigs by means of the Pig Cough Monitor which enabled an objective measurement of “respiratory health” or cough, respectively (Genzow et al. 2014). Schön et al. (2004) developed a system using a combination of sound analysis and artificial neural networks to detect stress vocalisation of pigs in noisy pig units with <5% recognition errors (Schön et al. 2004).
Take home message: Automatic detection and recognition of pig diseases using sound data can be an efficient and economical solution. We suggest considering the multi-modality of video and audio data in a pig house surveillance system which can incorporate the automatic detection and recognition of a pig’s vocalization.
Expression of social behavior
Viazzi et al. (2014) extracted the number of moving pixels on a given pig pen as a proxy to aggression intensity and to extract the ratio of moving pixels to the group pixels as an occupation index to detect aggression (Viazzi et al. 2014). Lee et al. (2016) used a Kinect depth sensor to measure the speed of standing pigs and the distance between pigs to detect aggression and classify both knocking and chasing behaviors (Lee et al. 2016). Chen et al. (2019) developed a depth image analysis method to automatically detect aggressive behaviors of group-housed pigs. Their proposed algorithm detected aggressive behaviors with an accuracy of 97.5%, a sensitivity of 98.2%, specificity of 96.7%, and precision of 96.8% (Chen et al. 2019). Oczak et al. (2014) tested a method to automatically detect aggressive behavior in pigs, by using an activity index and a multilayer feed forward neural network. The results revealed that artiﬁcial neural network classiﬁed high aggression events with a sensitivity of 96.1%, speciﬁcity of 94.2%, and an accuracy of 99.8% whereas medium aggression events were classiﬁed with a sensitivity of 86.8%, speciﬁcity of 94.5%, and an accuracy of 99.2% (Oczak et al. 2014).
Take home message: It is possible to use machine vision techniques in order to automatically detect aggressive behaviors among pigs under commercial farm conditions.
Precision livestock farming increases the farmer’s ability to keep contact with individual animals in the growing livestock production intensification. Using PLF, a large amount of data can be collected in a short period of time which can improve swine welfare monitoring. However, it should be noticed that the enthusiasm for automated recording does not lead to more emphasis on certain behaviors in welfare assessment. We need to choose behavioral measures according to their relevance to animal welfare and then develop automatic recording methods, rather than choosing measures for their ability to be recorded automatically (Rushen et al. 2012). Additionally, future larger scale studies and data combination from different sources in a multivariate approach are needed to develop a robust on-farm welfare monitoring system.
References and Citations
Blömke, L., N. Volkmann, and N. Kemper. 2020. “Evaluation of an Automated Assessment System for Ear and Tail Lesions as Animal Welfare Indicators in Pigs at Slaughter.” Meat Science 159: 107934. https://doi.org/10.1016/j.meatsci.2019.107934.
Chen, C., W. Zhu, D. Liu, J. Steibel, J. Siegford, K. Wurtz, J. Han, and T. Norton. 2019. “Detection of Aggressive Behaviours in Pigs Using a RealSence Depth Sensor.” Computers and Electronics in Agriculture 166: 105003.
Chung, Y., S. Oh, J. Lee, D. Park, H.H. Chang, and S. Kim. 2013. “Automatic Detection and Recognition of Pig Wasting Diseases Using Sound Data in Audio Surveillance Systems.” Sensors 13 (10): 12929–42. https://doi.org/10.3390/s131012929.
Condotta, I., T. Brown-Brandl, K. Silva-Miranda, and J. Stinn. 2018. “Evaluation of a Depth Sensor for Mass Estimation of Growing and Finishing Pigs.” Science Direct 3: 2–9.
Curtis, SE. 2007. “Performance Indicates Animal State of Being: A Cinderella Axiom?” The Professional Animal Scientist 27 (6): 573–83.
D’Eath, R., M. Jack, A. Futro, D. Talbot, Q. Zhu, D. Barclay, and E. Baxter. 2018. “Automatic Early Warning of Tail Biting in Pigs: 3D Cameras Can Detect Lowered Tail Posture before an Outbreak.” PLoS ONE 13 (4): 1–18. https://doi.org/10.1371/journal.pone.0194524.
Diana, A., L. Carpentier, D. Piette, L. Boyle, D. Berckmans, and T. Norton. 2019. “An Ethogram of Biter and Bitten Pigs during an Ear Biting Event: First Step in the Development of a Precision Livestock Farming Tool.” Applied Animal Behaviour Science 215: 26–36. https://doi.org/10.1016/j.applanim.2019.03.011.
Exadaktylos, V., M. Silva, J. M. Aerts, C. J. Taylor, and D. Berckmans. 2008. “Real-Time Recognition of Sick Pig Cough Sounds.” Computers and Electronics in Agriculture 63 (2): 207–14. https://doi.org/10.1016/j.compag.2008.02.010.
FAO. 2009. The State of Food and Agriculture 2009 ; Livestock in the Balance. Rome, Italy. https://doi.org/10.18356/6e4ebb75-en.
Ferrari, S., M. Silva, M. Guarino, J. Aerts, and D. Berckmans. 2008. “Cough Sound Analysis to Identify Respiratory Infection in Pigs.” Computers and Electronics in Agriculture 64 (2): 318–25. https://doi.org/10.1016/j.compag.2008.07.003.
Finger, G., M. Hemeryck, C. Duran, and M. Genzow. 2014. “Practical Application of the Pig Cough Monitor in a German Fattening Pig Herd with PRDC.” In Proceedings of the 23rd IPVS Congress, 207. Cancun, Mexico.
Genzow, M., C. Duran, K. Strutzberg-Minder, G. Finger, and M. Hemeryck. 2014. “Monitoring of a Commercial Fattening Herd by Means of the Pig Cough Monitor and Oral Fluid Diagnostics.” In Proceedings of the 23rd IPVS Congress, 116. Cancun, Mexico.
Hart, Edmund M., Pauline Barmby, David LeBauer, François Michonneau, Sarah Mount, Patrick Mulrooney, Timothée Poisot, Kara H. Woo, Naupaka B. Zimmerman, and Jeffrey W. Hollister. 2016. “Ten Simple Rules for Digital Data Storage.” PLoS Computational Biology 12 (10): 1–12. https://doi.org/10.1371/journal.pcbi.1005097.
Kashiha, M., C. Bahr, S. Ott, C. Moons, T. Niewold, F. Ödberg, and D. Berckmans. 2014a. “Automatic Weight Estimation of Individual Pigs Using Image Analysis.” Computers and Electronics in Agriculture 107: 38–44. https://doi.org/10.1016/j.compag.2014.06.003.
Kashiha, M., C. Bahr, S. Ott, C. Moons, T. Niewold, F. Tuyttens, and D. Berckmans. 2014b. “Automatic Monitoring of Pig Locomotion Using Image Analysis.” Livestock Science 159 (1): 141–48. https://doi.org/10.1016/j.livsci.2013.11.007.
Kollis, K., C. S. Phang, T. Banhazi, and S. Searle. 2007. “Weight Estimation Using Image Analysis and Statistical Modelling: A Preliminary Study.” Applied Engineering in Agriculture 23 (1): 91–96. https://doi.org/10.13031/2013.22332.
Kongsro, J. 2013. “Development of a Computer Vision System to Monitor Pig Locomotion.” Journal of Animal Sciences 3 (3): 254–60. https://doi.org/10.4236/ojas.2013.33038.
Kongsro, J. 2014. “Estimation of Pig Weight Using a Microsoft Kinect Prototype Imaging System.” Computers and Electronics in Agriculture 109: 32–35. https://doi.org/10.1016/j.compag.2014.08.008.
Lee, J., L. Jin, D. Park, and Y. Chung. 2016. “Automatic Recognition of Aggressive Behavior in Pigs Using a Kinect Depth Sensor.” Sensors 16: 631.
Lind, N., M. Vinther, R. Hemmingsen, and A. Hansen. 2005. “Validation of a Digital Video Tracking System for Recording Pig Locomotor Behaviour.” Journal of Neuroscience Methods 143 (2): 123–32. https://doi.org/10.1016/j.jneumeth.2004.09.019.
McNeil, B., K. Stalder, J. Calderon Diaz, J. Stock, T. Parsons, D. Beam, and A. Johnson. 2019. “Development of Sow Lameness Classification Trees Using an Embedded Microcomputer-Based Force Plate in a Commercial Setting” 3 (1): 1–6. https://doi.org/10.21767/2577-0594.10006.
Meijer, E., C. Bertholle, M. Oosterlinck, F. van der Staay, W. Back, and A. van Nes. 2014. “Pressure Mat Analysis of the Longitudinal Development of Pig Locomotion in Growing Pigs after Weaning.” BMC Veterinary Research 10 (1): 1–11. https://doi.org/10.1186/1746-6148-10-37.
Oczak, M., S. Viazzi, G. Ismayilova, L. Sonoda, N. Roulston, M. Fels, C. Bahr, et al. 2014. “Classification of Aggressive Behaviour in Pigs by Activity Index and Multilayer Feed Forward Neural Network.” Biosystems Engineering 119: 89–97. https://doi.org/10.1016/j.biosystemseng.2014.01.005.
Pluym, L., D. Maes, J. Vangeyte, K. Mertens, J. Baert, S. Van Weyenberg, S. Millet, and A. Van Nuffel. 2013. “Development of a System for Automatic Measurements of Force and Visual Stance Variables for Objective Lameness Detection in Sows: SowSIS.” Biosystems Engineering 116 (1): 64–74. https://doi.org/10.1016/j.biosystemseng.2013.06.009.
Ringgenberg, N., and R. Bergeron. 2010. “Validation of Accelerometers to Automatically Record Sow Postures and Stepping Behaviour.” Applied Animal Behaviour Science 128 (1–4): 37–44. https://doi.org/10.1016/j.applanim.2010.09.018.
Rushen, J., N. Chapinal, and A. M. De Passillé. 2012. “Automated Monitoring of Behavioural-Based Animal Welfare Indicators.” Animal Welfare 21 (3): 339–50. https://doi.org/10.7120/096272126.96.36.1999.
Schon, H., and A. G. Meiering. 1987. “Computer-Aided Control Improves Livestock Operations.” Agricultural Engineering 68 (7): 15–18.
Schön, P. C., B. Puppe, and G. Manteuffel. 2004. “Automated Recording of Stress Vocalisations as a Tool to Document Impaired Welfare in Pigs.” Animal Welfare 13 (2): 105–10.
Shao, B., and H. Xin. 2008. “A Real-Time Computer Vision Assessment and Control of Thermal Comfort for Group-Housed Pigs.” Computers and Electronics in Agriculture 62 (1): 15–21. https://doi.org/10.1016/j.compag.2007.09.006.
Shao, J. 1997. “Classification of Swine Thermal Comfort Behavior by Image Processing and Neural Network, Retrospective Theses and Dissertations.” Iowa State University.
Stavrakakis, S., J. H. Guy, I. Syranidis, G. Johnson, and S. Edwards. 2015. “Pre-Clinical and Clinical Walking Kinematics in Female Breeding Pigs with Lameness: A Nested Case-Control Cohort Study.” Veterinary Journal 205 (1): 38–43. https://doi.org/10.1016/j.tvjl.2015.04.022.
Stavrakakis, S., W. Li, J. H. Guy, G. Morgan, G. Ushaw, G.R. Johnson, and S.A. Edwards. 2015. “Validity of the Microsoft Kinect Sensor for Assessment of Normal Walking Patterns in Pigs.” Computers and Electronics in Agriculture 117: 1–7. https://doi.org/10.1016/j.compag.2015.07.003.
Sun, G., R. F. Fitzgerald, K. J. Stalder, L. A. Karriker, A. K. Johnson, and S. J. Hoff. 2011. “Development of an Embedded Microcomputer-Based Force Plate System for Measuring Sow Weight Distribution and Detection of Lameness.” Applied Enginnering in Agriculture 475–482: 475–82.
Swanson, JC. 1995. “Farm Animal Well-Being and Intensive Production Systems.” Journal of Animal Science 73: 2744–51.
Vandermeulen, J., C. Bahr, E. Tullo, I. Fontana, S. Ott, M. Kashiha, M. Guarino, et al. 2015. “Discerning Pig Screams in Production Environments.” PLoS ONE 10 (4): 1–15. https://doi.org/10.1371/journal.pone.0123111.
Viazzi, S., G. Ismayilova, M. Oczak, L.T. Sonoda, M. Fels, M. Guarino, E. Vranken, J. Hartung, C. Bahr, and D. Berckmans. 2014. “Image Feature Extraction for Classification of Aggressive Interactions among Pigs.” Computers and Electronics in Agriculture 104 (2): 57–62.
Vranken, E., and D. Berckmans. 2017. “Precision Livestock Farming for Pigs.” Animal Frontiers 7 (1): 32–37.
Wang, K., H. Guo, Q. Ma, W. Su, L. Chen, and D. Zhu. 2018. “A Portable and Automatic Xtion-Based Measurement System for Pig Body Size.” Computers and Electronics in Agriculture 148: 291–98. https://doi.org/10.1016/j.compag.2018.03.018.
Welfare Quality. 2009. “Welfare Quality Assessment Protocol for Pig.” Lelystad, The Netherlands.
Wu, J., R. Tillett, N. McFarlane, X. Ju, J. Siebert, and P. Schofield. 2004. “Extracting the Three-Dimensional Shape of Live Pigs Using Stereo Photogrammetry.” Computers and Electronics in Agriculture 44 (3): 203–22. https://doi.org/10.1016/j.compag.2004.05.003.
Xin, H. 1999. “Assessing Swine Thermal Comfort by Image Analysis of Postural Behaviors.” Journal of Animal Science 77 Suppl 2: 1–9. https://doi.org/10.2527/1999.77suppl_21x.
Xin, H., and B. Shao. 2002. “Real-Time Assessment of Swine Thermal Comfort by Computer Vision.” In Proceedings of the World Congress of Computers in Agriculture and Natural Resources, 362–69. Iguacu Falls, Brazil. https://doi.org/10.13031/2013.8353.