Activity biosensors have been used recently to measure and diagnose the physiological status of dairy cows. However, owing to the variety of commercialized activity biosensors available in the market, activity data generated by a biosensor need to be standardized to predict the status of an animal and make relevant decisions. Hence, the objective of this study was to develop a standardization method for accommodating activity measurements from different sensors. Twelve Holstein dairy cows were monitored to collect 12 862 activity data from four types of sensors over five months. After confirming similar cyclic activity patterns from the sensors through correlation and regression analyses, the gamma distribution was employed to calculate the cumulative probability of the values of each biosensor. Then, the activity values were assigned to three levels (i.e., idle, normal and active) based on the defined proportion of each level, and the values at each level from the four sensors were compared. The results showed that the number of measurements belonging to the same level was similar, with less than a 10% difference at a specific threshold value. In addition, more than 87% of the heat alerts generated by the internal algorithm of three of the four biosensors could be assigned to the active level, suggesting that the current standardization method successfully integrated the activity measurements from different biosensors. The developed probability-based standardization method is expected to be applicable to other biosensors for livestock, which will lead to the development of models and solutions for precision livestock farming.

Digital Livestock Technologies (DLTs) can assist farmer decision-making and promise benefits to animal health and welfare. However, the extent to which they can help improve animal welfare is unclear. This study explores how DLTs may impact farm management and animal welfare by promoting learning, using the concept of boundary objects. Boundary objects may be interpreted differently by different social worlds but are robust enough to share a common identity across them. They facilitate communication around a common issue, allowing stakeholders to collaborate and co-learn. The type of learning generated may impact management and welfare differently. For example, it may help improve existing strategies (single-loop learning), or initiate reflection on how these strategies were framed initially (double-loop learning). This study focuses on two case studies, during which two DLTs were developed and tested on farms. In-depth, semi-structured interviews were conducted with stakeholders involved in the case studies (n = 31), and the results of a separate survey were used to complement our findings. Findings support the important potential of DLTs to help enhance animal welfare, although the impacts vary between technologies. In both case studies, DLTs facilitated discussions between stakeholders, and whilst both promoted improved management strategies, one also promoted deeper reflection on the importance of animal emotional well-being and on providing opportunities for positive animal welfare. If DLTs are to make significant improvements to animal welfare, greater priority should be given to DLTs that promote a greater understanding of the dimensions of animal welfare and a reframing of values and beliefs with respect to the importance of animals’ well-being.

Ranked amongst the top five milk-producing countries globally, the Pakistan dairy industry can help to overcome food shortage and hunger, alleviate poverty and positively impact economic growth. This influencing role could potentially be more significant while the COVID-19 pandemic severely affects humanity, challenges the economy and increases the risk of global food shortage. However, its large national population of dairy livestock contrasts with Pakistan's top-five ranking, indeed, four to five Pakistani cows produce milk equivalent to one dairy cow of countries with a well-developed dairy industry. Low milk yield per cow negatively impacts the national production and compromises the development of an efficient processing sector, such that consumers are very often forced to use adulterated milk sold by local ‘milkmen.’ As a consequence, while committed to alleviating global hunger, Pakistan imports in excess of half a million tons of milk and milk-based products annually. Many studies have identified unproductive, inefficient and imprecise management issues combined with poor genetics and imbalanced nutrition as the leading barriers to improvement in the Pakistani dairy livestock sector. At an individual level, lack of awareness, affordability issues, illiteracy and low ambition of a large percentile of dairy farmers creates additional significant barriers. To address low productivity and poor genetics, Pakistani corporate farms and wealthier individual farmers import genetically improved breeds to attain high milk yields. However, they are then faced with the challenge of managing such breeds to attain sustainable and persistent milk yields under Pakistani climatic conditions, often risking excessive culling even to the point of business liquidation. In developed dairy industries, automated sensor-based livestock management systems are now available to help monitor, compute, and optimize procedures in real-time and are proven to increase productivity and profitability. The term precision livestock farming (PLF) is used to describe systems that monitor individual animals or groups of animals to overcome management deficiencies and optimize productivity. My stance in this Opinion Paper is that adopting and utilizing such precision technologies may support Pakistan in raising its livestock resources toward greater productivity, thereby helping to overcome the global food shortage.

The aim of this Research Communication was to apply the data mining technique to classify which environmental factors have the potential to motivate dairy cows to access natural shade. We defined two different areas at the silvopastoral system: shaded and sunny. Environmental factors and the frequency that dairy cows used each area were measured during four days, for 8 h each day. The shaded areas were the most used by dairy cows and presented the lowest mean values of all environmental factors. Solar radiation was the environmental factor with most potential to classify the dairy cow's decision to access shaded areas. Data mining is a machine learning technique with great potential to characterize the influence of the thermal environment in the cows' decision at the pasture.

Diversity of production systems and specific socio-economic barriers are key reasons explaining why the implementation of new technologies in small ruminants, despite being needed and beneficial for farmers, is harder than in other livestock species. There are, however, helpful peculiarities where small ruminants are concerned: the compulsory use of electronic identification created a unique scenario in Europe in which all small ruminant breeding stock became searchable by appropriate sensing solutions, and the largest small ruminant population in the world is located in Asia, close to the areas producing new technologies. Notwithstanding, only a few research initiatives and literature reviews have addressed the development of new technologies in small ruminants. This Research Reflection focuses on small ruminants (with emphasis on dairy goats and sheep) and reviews in a non-exhaustive way the basic concepts, the currently available sensor solutions and the structure and elements needed for the implementation of sensor-based husbandry decision support. Finally, some examples of results obtained using several sensor solutions adapted from large animals or newly developed for small ruminants are discussed. Significant room for improvement is recognized and a large number of multiple-sensor solutions are expected to be developed in the relatively near future.

This Research Reflection addresses the possibilities for Welfare Quality® to evolve from an assessment method based on data gathered on punctual visits to the farm to an assessment method based on sensor data. This approach could provide continuous and objective data, while being less costly and time consuming. Precision Livestock Farming (PLF) technologies enabling the monitorisation of Welfare Quality® measures are reviewed and discussed. For those measures that cannot be assessed by current technologies, some options to be developed are proposed. Picturing future dairy farms, the need for multipurpose and non-invasive PLF technologies is stated, in order to avoid an excessive artificialisation of the production system. Social concerns regarding digitalisation are also discussed.

Manually counting hens in battery cages on large commercial poultry farms is a challenging task: time-consuming and often inaccurate. Therefore, the aim of this study was to develop a machine vision system that automatically counts the number of hens in battery cages. Automatically counting hens can help a regulatory agency or inspecting officer to estimate the number of living birds in a cage and, thus animal density, to ensure that they conform to government regulations or quality certification requirements. The test hen house was 87 m long, containing 37 battery cages stacked in 6-story high rows on both sides of the structure. Each cage housed 18 to 30 hens, for a total of approximately 11 000 laying hens. A feeder moves along the cages. A camera was installed on an arm connected to the feeder, which was specifically developed for this purpose. A wide-angle lens was used in order to frame an entire cage in the field of view. Detection and tracking algorithms were designed to detect hens in cages; the recorded videos were first processed using a convolutional neural network (CNN) object detection algorithm called Faster R-CNN, with an input of multi-angular view shifted images. After the initial detection, the hens’ relative location along the feeder was tracked and saved using a tracking algorithm. Information was added with every additional frame, as the camera arm moved along the cages. The algorithm count was compared with that made by a human observer (the ‘gold standard’). A validation dataset of about 2000 images achieved 89.6% accuracy at cage level, with a mean absolute error of 2.5 hens per cage. These results indicate that the model developed in this study is practicable for obtaining fairly good estimates of the number of laying hens in battery cages.

Digitalisation is an integral part of modern agriculture. Several digital technologies are available for different animal species and form the basis for precision livestock farming. However, there is a lack of clarity as to which digital technologies are currently used in agricultural practice. Thus, this work aims to present for the first time the status quo in Swiss livestock farming as an example of a highly developed, small-scale and diverse structured agriculture. In this context, the article focuses on the adoption of electronic sensors and measuring devices, electronic controls and electronic data-processing options and the usage of robotics in ruminant farming, namely, for dairy cattle, dairy goats, suckler cows, beef cattle and meat-sheep. Furthermore, the use of electronic ear tags for pigs and the smartphone usage for barn monitoring on poultry farms was assessed. To better understand the adoption process, farm and farmer’s characteristics associated with the adoption of (1) implemented and (2) new digital technologies in ruminant farming were assessed using regression analyses, which is classified at a 10% adoption hurdle. The results showed clear differences in the adoption rates between different agricultural enterprises, with both types of digital technologies tending to be used the most in dairy farming. Easy-to-use sensors and measuring devices such as those integrated in the milking parlour were more widespread than data processing technologies such as those used for disease detection. The husbandry system further determined the use of digital technologies, with the result that farmers with tie stall barns were less likely to use digital technologies than farmers with loose housing systems. Additional studies of farmers’ determinants and prospects of implementation can help identify barriers in the adoption of digital technologies.

Precision feeding requires a mathematical model to estimate standardized ileal digestible (SID) lysine (Lys) requirements (SIDLysR) in real time. However, this type of model requires constant calibration updates. The objective of this study was to review the calibration of the model used to estimate the real-time Lys requirements of individual growing-finishing pigs. A digestibility trial (n = 10) was conducted to evaluate amino acids digestibility during the growing and finishing phases. Additionally, 120 pigs were used in two 28-day growth experiments conducted as completely randomized design with growing (25 ± 2.1 kg BW, n = 60; 10 pigs per treatment) or finishing barrows (68.1 ± 6 kg BW, n = 60; 10 pigs per treatment). In each experiment, the pigs were divided into six equal treatment groups and fed 60%, 70%, 80%, 90%, 100% or 110% of their estimated individual SIDLysR. The Lys requirement of each pig was estimated daily using a real-time model. Body composition was measured with dual-energy X-ray densitometry on day 1 and 28 of the experiments. Average daily feed intake increased quadratically (P < 0.05) during both growth phases. Maximum average daily gain (ADG) (0.98 kg) and maximum protein deposition (PD; 170 g/day) were observed in growing pigs fed 100% of the estimated SIDLysR (P < 0.001). During the growing period, PD in BW gain (17% to 19%) and N efficiency (52% to 65%) increased linearly (P < 0.01) with increasing inclusion rates of SID Lys. Finishing pigs had maximum ADG (1.2 kg/day) when they were fed 100% of the requirements. However, the amount of protein in BW gain (13% to 16%) and N efficiency (40% to 55%) increased linearly (P < 0.01) with increasing inclusion rates of SID Lys. In conclusion, the model proposed for precision feeding is correctly calibrated to predict SIDLysR that maximize PD and ADG of average pigs from 25 to 50 kg BW. Still, there is an opportunity to improve the estimation of SIDLysR and N retention in individual pigs by better representing the individual proportion of protein in BW gain and the factors controlling the efficiency of Lys utilization in individual pigs.

In this position paper, I shall summarise the current status of sensor technologies in ruminant livestock farming with emphasis on dairy cattle, outline the case for why I believe that sensor technologies could revolutionise global dairy farming in a positive way, describe the significant barriers that exist if that goal is to be achieved and highlight the benefits to animal wellbeing, profitability and sustainability that could result if the technologies are implemented to a significant extent. I shall not provide a comprehensive review of the sensor technology literature since that has been done before, but I intend to provide a sensible amount of background information and data that will allow the reader to obtain a picture not only of today’s sensor usage but, more importantly, the possible future direction of dairy animal-oriented sensor technologies, and I shall substantiate my claims and conclusions with relevant literature.

Worldwide, there is a trend towards increased herd sizes, and the animal-to-stockman ratio is increasing within the beef and dairy sectors; thus, the time available to monitoring individual animals is reducing. The behaviour of cows is known to change in the hours prior to parturition, for example, less time ruminating and eating and increased activity level and tail-raise events. These behaviours can be monitored non-invasively using animal-mounted sensors. Thus, behavioural traits are ideal variables for the prediction of calving. This study explored the potential of two sensor technologies for their capabilities in predicting when calf expulsion should be expected. Two trials were conducted at separate locations: (i) beef cows (n = 144) and (ii) dairy cows (n = 110). Two sensors were deployed on each cow: (1) Afimilk Silent Herdsman (SHM) collars monitoring time spent ruminating (RUM), eating (EAT) and the relative activity level (ACT) of the cow, and (2) tail-mounted Axivity accelerometers to detect tail-raise events (TAIL). The exact time the calf was expelled from the cow was determined by viewing closed-circuit television camera footage. Machine learning random forest algorithms were developed to predict when calf expulsion should be expected using single-sensor variables and by integrating multiple-sensor data-streams. The performance of the models was tested using the Matthew’s correlation coefficient (MCC), the area under the curve, and the sensitivity and specificity of predictions. The TAIL model was slightly better at predicting calving within a 5-h window for beef cows (MCC = 0.31) than for dairy cows (MCC = 0.29). The TAIL + RUM + EAT models were equally as good at predicting calving within a 5-h window for beef and dairy cows (MCC = 0.32 for both models). Combining data-streams from SHM and tail sensors did not substantially improve model performance over tail sensors alone; therefore, hour-by-hour algorithms for the prediction of time of calf expulsion were developed using tail sensor data. Optimal classification occurred at 2 h prior to calving for both beef (MCC = 0.29) and dairy cows (MCC = 0.25). This study showed that tail sensors alone are adequate for the prediction of parturition and that the optimal time for prediction is 2 h before expulsion of the calf.

Heat stress is one of the most critical issues jeopardising animal welfare and productivity during the warm season in dairy cattle farms. The global trend of increase in average and peak temperatures is making the problem more and more serious. Many devices have been introduced in livestock farms to monitor and control temperature-humidity index, as well as animal behaviour and production parameters. The consequent availability of collected databases has increasingly enhanced the research aimed to understand the consequences of heat stress in cattle, in relation to genetic, reproductive, productive and behavioural features. Moreover, these investigations laid the foundations for the development, calibration, validation and test of numerical models quantifying the individual responses to heat stress conditions. In this work, a generalised additive model with mixed effects has been developed to analyse the relationship between milk production, animal behaviour and environmental parameters based on data surveyed in 2016 in an Italian dairy farm. Each cow has been characterised in terms of her response to heat conditions, and the results led to define three classes of susceptibility to heat stress within the herd. These attributes have then been related to the various phenotypic parameters collected by the precision livestock farming devices used in the farm. The study provides a model to understand the effects of heat stress conditions on individual animals in relation to the main parameters describing their rearing conditions; moreover, the results contribute to improve the herd management by lending indications to define targeted treatments according to the cow’s characteristics.

We address the hypothesis that individual cow feed intake can be measured in commercial farms through the use of a photogrammetry method. Feed intake and feed efficiency have a significant economic value for the farmer. A common method for measuring feed mass in research is a feed mass weighing system, which is excessively expensive for commercial farms. However, feed mass can be estimated by its volume, which can be measured by photogrammetry. Photogrammetry applies cameras along the feed-lane, photographing the feed before and after the cow visits the feed-lane, and calculating the feed volume. In this study, the precision of estimating feed mass by its volume was tested by comparing measured mass and calculated volume of feed heaps. The following principal factors had an impact on the precision of this method: camera quality, lighting conditions, image resolution, number of images, and feed density. Under laboratory conditions, the feed mass estimation error was 0·483 kg for heaps up to 7 kg, while in the cowshed the estimation error was 1·32 kg for up to 40 kg. A complementary experiment showed that the natural feed compressibility causes about 85% of uncertainty in the mass estimation error.

Tail damage within the production of finisher pigs is an animal welfare problem. Recent research suggests that removal of known risk factors may not be enough to eliminate tail biting, especially in undocked pigs, thus a different strategy is worth investigating. This could be early detection of tail biting, using behavioural changes observed before tail damage. If these early stages of tail biting can be detected before tail damage occurs, then tail damage could be prevented by early interventions. The first step in developing such a strategy is to identify the types of behaviour changes that emerge during early stages of tail biting. Thus, the aim of the current study was to investigate whether pen level activity and object manipulation evolved differently during the last 7 days before the scoring of tail damage (day 0) for pens scored with tail damage (tail damage pens) and pens not scored with tail damage (matched control pens). The study included video recordings for twenty-four tail damage pens and thirty-two matched control pens. Activity level and object manipulation were observed the last 7 days before day 0 during the morning (0600 to 0800 h), afternoon (1600 to 1800 h) and evening (2200 to 2400 h, only activity level). Both activity level and object manipulation were analysed using generalised linear mixed effects models with a binomial distribution for activity level and a negative binomial distribution for object manipulation. The probability of being active was higher in tail damage pens compared to control pens during the afternoon the last 5 days before day 0 (P<0.001). This was seen due to a decrease in activity level in the control pens, which makes it difficult to identify future tail damage pens from this difference. Object manipulation was lower in tail damage pens compared to the control pens on all 7 days before day 0, but only in pens with undocked pigs (P<0.01). Thus, it is still unknown when this difference in object manipulation initiated. It was concluded that both activity level and object manipulation seemed related to ongoing tail biting and should be investigated through more detailed observations and for a longer time to establish the normal behaviour pattern for a particular pen. Thus, it is suggested that future research focusses on developing automatic monitoring methods for pen level activity and object manipulation and applies algorithms that establish and detect deviations from the normal behaviour pattern of the pen before tail damage.

Early detection of post-calving health problems is critical for dairy operations. Separating sick cows from the herd is important, especially in robotic-milking dairy farms, where searching for a sick cow can disturb the other cows’ routine. The objectives of this study were to develop and apply a behaviour- and performance-based health-detection model to post-calving cows in a robotic-milking dairy farm, with the aim of detecting sick cows based on available commercial sensors. The study was conducted in an Israeli robotic-milking dairy farm with 250 Israeli-Holstein cows. All cows were equipped with rumination- and neck-activity sensors. Milk yield, visits to the milking robot and BW were recorded in the milking robot. A decision-tree model was developed on a calibration data set (historical data of the 10 months before the study) and was validated on the new data set. The decision model generated a probability of being sick for each cow. The model was applied once a week just before the veterinarian performed the weekly routine post-calving health check. The veterinarian’s diagnosis served as a binary reference for the model (healthy–sick). The overall accuracy of the model was 78%, with a specificity of 87% and a sensitivity of 69%, suggesting its practical value.

Dairy cows are high value farm animals requiring careful management to achieve the best results. Since the advent of robotic and high throughput milking, the traditional few minutes available for individual human attention daily has disappeared and new automated technologies have been applied to improve monitoring of dairy cow production, nutrition, fertility, health and welfare. Cows milked by robots must meet legal requirements to detect healthy milk. This review focuses on emerging technical approaches in those areas of high cost to the farmer (fertility, metabolic disorders, mastitis, lameness and calving). The availability of low cost tri-axial accelerometers and wireless telemetry has allowed accurate models of behaviour to be developed and sometimes combined with rumination activity detected by acoustic sensors to detect oestrus; other measures (milk and skin temperature, electronic noses, milk yield) have been abandoned. In-line biosensors have been developed to detect markers for ovulation, pregnancy, lactose, mastitis and metabolic changes. Wireless telemetry has been applied to develop boluses for monitoring the rumen pH and temperature to detect metabolic disorders. Udder health requires a multisensing approach due to the varying inflammatory responses collectively described as mastitis. Lameness can be detected by walk over weigh cells, but also by various types of video image analysis and speed measurement. Prediction and detection of calving time is an area of active research mostly focused on behavioural change.

Low-cost feeding-behavior sensors will soon be available for commercial use in dairy farms. The aim of this study was to develop a feed intake model for the individual dairy cow that includes feeding behavior. In a research farm, the individual cows’ voluntary feed intake and feeding behavior were monitored at every meal. A feed intake model was developed based on data that exist in commercial modern farms: ‘BW,’ ‘milk yield’ and ‘days in milking’ parameters were applied in this study. At the individual cow level, eating velocity seemed to be correlated with feed intake (R 2=0.93 to 0.94). The eating velocity coefficient varied among individuals, ranging from 150 to 230 g/min per cow. The contribution of feeding behavior (0.28) to the dry matter intake (DMI) model was higher than the contribution of BW (0.20), similar to the contribution of fat-corrected milk (FCM)/BW (0.29) and not as large as the contribution of FCM (0.49). Incorporating feeding behavior into the DMI model improved its accuracy by 1.3 (38%) kg/cow per day. The model is ready to be implemented in commercial farms as soon as companies introduce low-cost feeding-behavior sensors on commercial level.