Area and Resource Utilization of Group-Housed Horses in an Active Stable.

The aim of this study was to analyze the utilization of different stable areas of a total of 52 group-housed horses as well as their preferred stable parts and the use of resources. The study was situated in a "HIT Active Stable®" in Northern Germany for a period of 227 observation days. After dividing the whole farm area in a grid of 3 × 3 m, the dataset was examined with and without the pasture area. Furthermore, linear mixed models were applied. On average, horses used 53.2 ± 19 different squares per hour. The observation day (p < 0.001) and the covariate age (p < 0.001) had significant effects on the different squares visited per hour. No significant effects were found for sex (p = 0.30) and breed (p = 0.65) as only geldings and no stallions were part of the group and the distribution of the breeds was unfavorable. The random effect animal showed that the horse-individual estimates from -19.2 to 17.6 different squares visited per hour were quite large. Furthermore, it could be shown that the horses used resources such as feed stalls with a frequency of up to 0.14% more than other paddock areas without resources. Open lying halls with tarp skin were also preferred over the metal hall. The shelters were only partly popular. Use could be visualized with the help of heat maps. This study gives a good overview of the use of individual areas and resources and possible improvements.

Instance Segmentation with Mask R-CNN Applied to Loose-Housed Dairy Cows in a Multi-Camera Setting.

With increasing herd sizes came an enhanced requirement for automated systems to support the farmers in the monitoring of the health and welfare status of their livestock. Cattle are a highly sociable species, and the herd structure has important impact on the animal welfare. As the behaviour of the animals and their social interactions can be influenced by the presence of a human observer, a camera based system that automatically detects the animals would be beneficial to analyse dairy cattle herd activity. In the present study, eight surveillance cameras were mounted above the barn area of a group of thirty-six lactating Holstein Friesian dairy cows at the Chamber of Agriculture in Futterkamp in Northern Germany. With Mask R-CNN, a state-of-the-art model of convolutional neural networks was trained to determine pixel level segmentation masks for the cows in the video material. The model was pre-trained on the Microsoft common objects in the context data set, and transfer learning was carried out on annotated image material from the recordings as training data set. In addition, the relationship between the size of the used training data set and the performance on the model after transfer learning was analysed. The trained model achieved averaged precision (Intersection over union, IOU = 0.5) 91% and 85% for the detection of bounding boxes and segmentation masks of the cows, respectively, thereby laying a solid technical basis for an automated analysis of herd activity and the use of resources in loose-housing.

Distances Walked by Long Established and Newcomer Horses in an Open Stable System in Northern Germany.

Group housing is claimed to possibly provide horses with a species-appropriate movement possibility, and hence, better welfare. Thus, this study analyzed the daily walked distances of 51 horses held in one group in a "HIT Active Stable" (Hinrichs Innovation + Technik) in Northern Germany by using global positioning system (GPS) technology during a 7 ½-month time span. The daily walking distances of the whole group, as well as newcomers, were investigated. The horses traveled an average of 8.43 km/day. Linear mixed models were applied. The observation day had a significant effect on the daily walking distances (P < .01) due to season and the available area per horse. The age as covariate also had a significant effect (P < .01). The breed had no significant effect (P = .96). No significant differences were found in sex (P = .69), which can be explained by the fact that only mares and geldings were investigated, which do not show increasing locomotion caused by sexual behavior as stallions do. On six of the first nine days, new horses moved significantly more compared to the remaining 24 of the 30 observation days directly after individuals' inclusion. This is probably due to more exploration and rank-fighting behavior. Similar walking distances were seen among the horses on the single observation days because all horses had to travel the same distance to reach resources. Further, it is suspected that not all horses can sufficiently live out their urges to move, especially in winter, when pasture is inaccessible.

"HerdGPS-Preprocessor"-A Tool to Preprocess Herd Animal GPS Data; Applied to Evaluate Contact Structures in Loose-Housing Horses.

Sensors delivering information on the position of farm animals have been widely used in precision livestock farming. Global Positioning System (GPS) sensors are already known from applications in military, private and commercial environments, and their application in animal science is increasing. However, as trade-offs between sensor cost, battery life and sensor weight have to be made, GPS based studies scheduling long data collection periods and including a high number of animals, have to deal with problems like high hardware costs and data disruption during recharging of sensors. Furthermore, human-animal interaction due to sensor changing at the end of battery life interferes with the animal behaviour under analysis. The present study thus proposes a setting to deal with these challenges and offers the software tool "HerdGPS-Preprocessor", because collecting position data from multiple animals nonstop for several weeks produces a high amount of raw data which needs to be sorted, preprocessed and provided in a suitable format per animal and day. The software tool "HerdGPS-Preprocessor" additionally outputs contact lists to enable a straight analysis of animal contacts. The software tool was exemplarily deployed for one month of daily and continuous GPS data of 40 horses in a loose-housing boarding facility in northern Germany. Contact lists were used to generate separate networks for every hour, which are then analysed with regard to the network parameter density, diameter and clique structure. Differences depending on the day and the day time could be observed. More dense networks with more and larger cliques were determined in the hours prior to the opening of additional pasture.

Determination of Body Parts in Holstein Friesian Cows Comparing Neural Networks and k Nearest Neighbour Classification.

Machine learning methods have become increasingly important in animal science, and the success of an automated application using machine learning often depends on the right choice of method for the respective problem and data set. The recognition of objects in 3D data is still a widely studied topic and especially challenging when it comes to the partition of objects into predefined segments. In this study, two machine learning approaches were utilized for the recognition of body parts of dairy cows from 3D point clouds, i.e., sets of data points in space. The low cost off-the-shelf depth sensor Microsoft Kinect V1 has been used in various studies related to dairy cows. The 3D data were gathered from a multi-Kinect recording unit which was designed to record Holstein Friesian cows from both sides in free walking from three different camera positions. For the determination of the body parts head, rump, back, legs and udder, five properties of the pixels in the depth maps (row index, column index, depth value, variance, mean curvature) were used as features in the training data set. For each camera positions, a k nearest neighbour classifier and a neural network were trained and compared afterwards. Both methods showed small Hamming losses (between 0.007 and 0.027 for k nearest neighbour (kNN) classification and between 0.045 and 0.079 for neural networks) and could be considered successful regarding the classification of pixel to body parts. However, the kNN classifier was superior, reaching overall accuracies 0.888 to 0.976 varying with the camera position. Precision and recall values associated with individual body parts ranged from 0.84 to 1 and from 0.83 to 1, respectively. Once trained, kNN classification is at runtime prone to higher costs in terms of computational time and memory compared to the neural networks. The cost vs. accuracy ratio for each methodology needs to be taken into account in the decision of which method should be implemented in the application.

Temporal correlation coefficient for directed networks.

Previous studies dealing with network theory focused mainly on the static aggregation of edges over specific time window lengths. Thus, most of the dynamic information gets lost. To assess the quality of such a static aggregation the temporal correlation coefficient can be calculated. It measures the overall possibility for an edge to persist between two consecutive snapshots. Up to now, this measure is only defined for undirected networks. Therefore, we introduce the adaption of the temporal correlation coefficient to directed networks. This new methodology enables the distinction between ingoing and outgoing edges. Besides a small example network presenting the single calculation steps, we also calculated the proposed measurements for a real pig trade network to emphasize the importance of considering the edge direction. The farm types at the beginning of the pork supply chain showed clearly higher values for the outgoing temporal correlation coefficient compared to the farm types at the end of the pork supply chain. These farm types showed higher values for the ingoing temporal correlation coefficient. The temporal correlation coefficient is a valuable tool to understand the structural dynamics of these systems, as it assesses the consistency of the edge configuration. The adaption of this measure for directed networks may help to preserve meaningful additional information about the investigated network that might get lost if the edge directions are ignored.

Quality assessment of static aggregation compared to the temporal approach based on a pig trade network in Northern Germany.

Recent analyses of animal movement networks focused on the static aggregation of trade contacts over different time windows, which neglects the system's temporal variation. In terms of disease spread, ignoring the temporal dynamics can lead to an over- or underestimation of an outbreak's speed and extent. This becomes particularly evident, if the static aggregation allows for the existence of more paths compared to the number of time-respecting paths (i.e. paths in the right chronological order). Therefore, the aim of this study was to reveal differences between static and temporal representations of an animal trade network and to assess the quality of the static aggregation in comparison to the temporal counterpart. Contact data from a pig trade network (2006-2009) of a producer community in Northern Germany were analysed. The results show that a median value of 8.7 % (4.6-14.1%) of the nodes and 3.1% (1.6-5.5%) of the edges were active on a weekly resolution. No fluctuations in the activity patterns were obvious. Furthermore, 50% of the nodes already had one trade contact after approximately six months. For an accumulation window with increasing size (one day each), the accumulation rate, i.e. the relative increase in the number of nodes or edges, stayed relatively constant below 0.07% for the nodes and 0.12 % for the edges. The temporal distances had a much wider distribution than the topological distances. 84% of the temporal distances were smaller than 90 days. The maximum temporal distance was 1000 days, which corresponds to the temporal diameter of the present network. The median temporal correlation coefficient, which measures the probability for an edge to persist across two consecutive time steps, was 0.47, with a maximum value of 0.63 at the accumulation window of 88 days. The causal fidelity measures the fraction of the number of static paths which can also be taken in the temporal network. For the whole observation period relatively high values indicate that 67% of the time-respecting paths existed in both network representations. An increase to 0.87 (0.82-0.88) and 0.92 (0.80-0.98), respectively, could be observed for yearly and monthly aggregation windows. The results show that the investigated pig trade network in its static aggregation represents the temporal dynamics of the system sufficiently well. Therefore, the methodology for analysing static instead of dynamic networks can be used without losing too much information.

Adaption of the temporal correlation coefficient calculation for temporal networks (applied to a real-world pig trade network).

The average topological overlap of two graphs of two consecutive time steps measures the amount of changes in the edge configuration between the two snapshots. This value has to be zero if the edge configuration changes completely and one if the two consecutive graphs are identical. Current methods depend on the number of nodes in the network or on the maximal number of connected nodes in the consecutive time steps. In the first case, this methodology breaks down if there are nodes with no edges. In the second case, it fails if the maximal number of active nodes is larger than the maximal number of connected nodes. In the following, an adaption of the calculation of the temporal correlation coefficient and of the topological overlap of the graph between two consecutive time steps is presented, which shows the expected behaviour mentioned above. The newly proposed adaption uses the maximal number of active nodes, i.e. the number of nodes with at least one edge, for the calculation of the topological overlap. The three methods were compared with the help of vivid example networks to reveal the differences between the proposed notations. Furthermore, these three calculation methods were applied to a real-world network of animal movements in order to detect influences of the network structure on the outcome of the different methods.

Quantification of the effects of fur, fur color, and velocity on Time-Of-Flight technology in dairy production.

With increasing herd sizes, camera based monitoring solutions rise in importance. 3D cameras, for example Time-Of-Flight (TOF) cameras, measure depth information. These additional information (3D data) could be beneficial for monitoring in dairy production. In previous studies regarding TOF technology, only standing cows were recorded to avoid motion artifacts. Therefore, necessary conditions for a TOF camera application in dairy cows are examined in this study. For this purpose, two cow models with plaster and fur surface, respectively, were recorded at four controlled velocities to quantify the effects of movement, fur color, and fur. Comparison criteria concerning image usability, pixel-wise deviation, and precision in coordinate determination were defined. Fur and fur color showed large effects (η (2)=0.235 and η (2)=0.472, respectively), which became even more considerable when the models were moving. The velocity of recorded animals must therefore be controlled when using TOF cameras. As another main result, body parts which lie in the middle of the cow model's back can be determined neglecting the effect of velocity or fur. With this in mind, further studies may obtain sound results using TOF technology in dairy production.

Feasibility of automated body trait determination using the SR4K time-of-flight camera in cow barns.

As herd sizes have increased in the last decades, computerized monitoring solutions, which provide fast, objective and accurate evaluations of the herd status, gain more and more importance. This study analyzes the feasibility of a Time-of-Flight-camera-based system for gathering body traits in dairy cows for use under cow barn conditions. Recording, determination of body condition score on a 5 point scale by visual and manual inspection, and measuring the backfat thickness with ultrasound took place from July 2011 to May 2012 at the dairy research farm Karkendamm of the Institute of Animal Breeding and Husbandry, Christian-Albrechts-University Kiel (Germany) and between August 2010 and July 2012 at the Institute for Agricultural Engineering and Animal Husbandry of Bavarian State Research Center for Agriculture in Grub (Germany). The two breeds Holstein Friesian cows (Karkendamm) and Fleckvieh (Grub) were considered in this study. Software for recording, image sorting and evaluation, determining the body parts needed, and extracting traits from the images was written and assembled to an automated system. Sorting the images and finding ischeal tuberosities, base of the tail, and dishes of the rump, backbone, and hips had error rates of 0.2%, 1.5%, 0.1%, and 2.6%, respectively. 13 traits were extracted and compared to backfat thickness and body condition score as well as between breeds. All traits depend significantly on the animal and showed very large effect sizes. Coefficients of determination restricted to individual animals were reaching up to 0.89. The precision in measuring the traits and gathering backfat thickness was comparable. Results indicated that the application of Time-Of-Flight in determination of body traits is feasible.