Animal Identification System Research Articles

Individual Buffalo Identification Through Muzzle Dermatoglyphics Images using Deep Learning Approaches

Animal identification is essential for routine farm operations, residue traceback, insurance, and ownership management. Owing to their uniqueness, incorrigible nature, tamperproof over time, environment-friendly, and pain-free, visual biometrics-based animal identification has recently gained momentum over traditional animal identification methods. Among visual biometrics-based cues, muzzle identification is a simple and relatively low-cost method. Therefore, to address the inherent significant limitations of conventional animal identification systems, we undertook this investigation to collect a database of digital images of muzzles that works as a benchmark, to apply deep learning frameworks to identify individual buffaloes from their muzzle images, and to compare their accuracy in terms of their identification capabilities. Muzzle images of 198 Surti buffaloes were subjected to transfer learning and fine-tuning processes in deep-learning neural networks. The performance was recorded for each pre-train model (ResNet50, InceptionV3, VGG16, AlexNet) with different hyperparameters of the epoch, batch size, and learning rate. A perusal of the data revealed that ResNet50 has the highest train accuracy (99.8%) and test accuracy (99.69%) among all four models used. AlexNet has the lowest train accuracy (90.8%) among the models. The findings concluded that all these four models could be applied to identify individual buffaloes; however, ResNet50 had the highest accuracy, and deep learning applications have great potential for individual buffalo identification and are promising tools for precision livestock farming

Rural Banditry and Economic Crimes: Exploring the Perceptions of Male Pastoralists’ Lived Experience in Kwara State, North-Central Nigeria

Armed banditry and insurgency are among the principal insecurity challenges in present-day Nigeria. This study investigated the lived experience of male pastoralists on rural banditry and associated economic crimes in Kwara State, north-central Nigeria to fill identified gaps in the literature. A sample of 20 pastoralists’ hamlets was selected for the study. Data were collected through focus group discussions and key informant interviews. Qualitative content analysis was used to analyse the data. Data analysis showed that study participants were aware of the problem of rural banditry. Major banditry-related economic crimes reported by participants were robbery and burglary, livestock rustling, farmers-herders conflicts with attendant destruction of pastoralists’ assets, among others. Through facilitation by pairwise ranking exercise within focus group sessions, study participants identified livestock rustling as the most important banditry-related economic crime. Participants carried out social profiling of livestock rustlers. Results further suggested that the key factors influencing livestock rustling were economic deprivations and social dislocation in Nigerian society, criminal networks on livestock rustling and illegal trading on rustled livestock, and access to light weapons by the livestock rustlers, among others. Participants reported the immediate and long-term effects of livestock rustling on the crime victims and their communities. The study recommends the rollout of the government-planned National Animal Identification and Traceability System, the enactment of state law that would prohibit indiscriminate livestock transportation and, the creation of a local security system to complement the state-based policing system to reduce armed banditry and livestock rustling among the study population.

218 Weaning weights associated with improved performance of grow-finish pigs

Abstract This study was conducted to evaluate the downstream impact of wean weight (WW) on finishing pig performance. Individual body weights (BW) were recorded at the time of weaning using the LeeO individual animal identification system. Finisher performance evaluation started at d 57 post-weaning. Pigs (n = 1,276; PIC 800 x DNA 241 F1) were blocked by wean age (20.6 ± 1.5 d) and sex, ranked in order of individual WW (6.05 ± 1.43 kg), and divided into BW quartiles to create four WW treatment groups [WW Quartile (Q) 1, 2, 3, and 4; Table 1]. Pigs (n = 21) in the same Q were housed in each pen yielding 15 replicate pens per quartile. Pigs were fed common diets formulated in a 4-phase sequence from the time of placement in the finisher through marketing, which occurred on a fixed-BW basis targeting 131.5 kg. Pen BW and feed disappearance were recorded every 3 wk and at marketing, and individual BW were recorded on d 134 post-weaning. Data were analyzed using linear mixed effects models in the lme4 package of R. Differences in mean BW of each Q were evident at weaning (P < 0.001), start of finisher evaluation (d 57, P < 0.001), and through d 121 (P < 0.001); by d 144, the BW between Q3 and Q4 did not differ (126.6 vs 129.0 kg, respectively, P > 0.05). Overall average daily gain (ADG) and average daily feed intake (ADFI) were greater (P < 0.001) for finishing pigs of either Q3 or Q4 compared with Q1, but no difference (P > 0.10) in gain to feed ratio (G:F) was observed between quartiles from 28.0 to 131.5 kg BW. Pigs in Q3 and Q4 required fewer (P < 0.001) days on feed than pigs in either Q1 or Q2, which culminated in an overall linear (P < 0.001) effect such that a 0.45-kg advantage in WW resulted in 2.25 d less to reach a market BW of 131.5 kg. Coefficient of variation (CV) in BW at d 134 post-weaning was greater (P < 0.001) for pigs in Q1 compared with pigs in Q2, Q3, or Q4. Backfat depth was thicker (P < 0.05) for Q1 pigs compared with Q4 pigs but did not differ from pigs in Q2 or Q3. A decrease (P < 0.001) in carcass ADG was seen for pigs in the lightest 50% of the population compared with those in Q4. In conclusion, even though feed efficiency was not impacted by WW, heavier WW resulted in fewer days to reach desired market weight while simultaneously resulting in improved carcass characteristics.

217 Weaning weight associated with improved nursery pig growth rate

Abstract Heavier body weight (BW) at weaning is associated not only with heavier birth weight, but also with heavier downstream BW at nursery exit and at fixed-time marketing. To further address management implications of wean weight (WW), this study evaluated WW impact on nursery pig start-up, gain to feed ratio (G:F), and performance variation from 6 to 20 kg BW. Individual BW were collected at the time of weaning using the LeeO individual animal identification system. Pigs (n = 1,054; PIC 800 x DNA 241 F1) were blocked by wean age (20.1 ± 0.8 d) and sex, ranked in order of individual WW (5.94 ± 1.35 kg), and split into eight WW groups each representing 12.5% of the weaning population (P < 0.05; Table 1). Similar sized pigs were housed together (12 pigs/pen, 11 replicate pens/WW group) and fed a common 4-phase nursery diet for 42 d post-wean. Pen weights and feed disappearance were recorded at d 7, 14, 21, and 42, and individual BW were collected at d 42. Data were analyzed using linear mixed effects models in the lme4 package of R. Body weight differences (P < 0.05) between each WW group were evident on d 0 and through d 7 post-weaning. The lightest 12.5% of the weaning population had the greatest (P < 0.001) individual pig BW coefficient of variation (CV) at weaning and d 42 whereas no differences (P > 0.05) in d 42 BW CV were evident among the other WW groups. During the first week post-weaning, average daily feed intake (ADFI) diverged (P < 0.05) between the lightest 12.5% (113 g/d) and the heaviest 25% (134 g/d) of the wean population. From d 0 to 42, the lightest 12.5% WW pigs had slower average daily gain (ADG; P < 0.001), less ADFI (P < 0.001), and reduced exit BW (P < 0.001) than the other WW groups. Neither ADG nor ADFI differed (P > 0.05) among the median 50% of the population from d 0 to 42. No differences (P > 0.10) in overall nursery G:F between any WW groups were observed. Across the population, regression analysis indicated a 0.45 kg increase in WW resulted in a 1.0 kg increase (linear, P < 0.001) in nursery exit BW. Moreover, both ADG and ADFI as a percentage of BW increased as WW decreased, suggesting lighter WW pigs have competitive growth capacity to their heavier WW counterparts through 20 kg BW. Data demonstrates WW influences nursery exit weight of pigs fed a common diet. Future research should explore solutions to increase growth rate of light WW pigs.

Development of a 3D tracking system for multiple marmosets under free-moving conditions

Assessment of social interactions and behavioral changes in nonhuman primates is useful for understanding brain function changes during life events and pathogenesis of neurological diseases. The common marmoset (Callithrix jacchus), which lives in a nuclear family like humans, is a useful model, but longitudinal automated behavioral observation of multiple animals has not been achieved. Here, we developed a Full Monitoring and Animal Identification (FulMAI) system for longitudinal detection of three-dimensional (3D) trajectories of each individual in multiple marmosets under free-moving conditions by combining video tracking, Light Detection and Ranging, and deep learning. Using this system, identification of each animal was more than 97% accurate. Location preferences and inter-individual distance could be calculated, and deep learning could detect grooming behavior. The FulMAI system allows us to analyze the natural behavior of individuals in a family over their lifetime and understand how behavior changes due to life events together with other data.

R-CNN Based Deep Learning Approach for Counting Animals in the Forest: A Survey

This review paper delves into the pivotal realm of animal classification using images obtained through diverse techniques in forest environments. A robust framework is introduced, employing Transfer Learning (TL) within a Convolutional Neural Network (CNN) and leveraging the power of the Region-based Convolutional Neural Network (R-CNN) model for the construction of an automated animal identification system. This innovative framework is adeptly applied to analyze and identify focal species within captured images, contributing to the advancement of wildlife monitoring technologies. The dataset under scrutiny comprises 6,203 camera trap images featuring 11 distinct species, including Wild pig, Barking deer, Chital, Elephant, Gaur, Hare, Jackal, Junglecat, Porcupine, Sambhar, and Sloth bear. The inclusion of this diverse set of species ensures the robustness and applicability of the proposed methodology across a broad spectrum of wildlife scenarios. The integration of Transfer Learning withinthe Region-based Convolutional Neural Network (R-CNN) emerges as a crucial element, showcasing outstanding performance in species classification.Notably, the proposed model achieves a remarkable accuracy rate of 96% on the test dataset after a mere 18 epochs, employing a batch size of 32. This breakthrough holds the potential to expedite research outcomes, foster the evolution of more efficient and dependable animal monitoring systems, and consequently, alleviate the time and effort invested by researchers.In line with ethical considerations, the authors maintain anonymity in theircontribution, focusing on the significant strides made in the classification andanalysis of camera trap images within the observed site. This paper positions itself as a noteworthy and impactful contribution to the broader field of wildlife research and technology

An automated approach for human-animal conflict minimisation in Assam and protection of wildlife around the Kaziranga National Park using YOLO and SENet Attention Framework

Human-animal conflict in Assam, India's north-eastern state, is rising continuously. Because it occurs year-round, it damages agricultural productivity and kills people and animals, including elephants. When a herd of wild elephants emerges from a deep forest and enters human-inhabited territory around the Kaziranga National Park (KNP) in Assam, an alert must be sounded for the neighbourhood residents and forest workers to prevent conflicts. Another concern is that many wild animals die near the KNP while crossing the national highway NH-37 which traverses the area. During floods, animals flee to the highlands for food and shelter. An automated animal identification and warning system near the KNP may reduce human-animal confrontations. This paper reports the design of a system that attempts to address the above concerns. Artificial Intelligence (AI)-based strategies are utilized to recognize wild animals from live video sequences, provide warnings to avoid encounters, and protect humans and animals. Deep learning models and YoloV5 with the SENet attention layer are used to recognize wild animals in real-time. This model is trained using a public and customized dataset of animal species. Cameras attached to the cloud-based AI system take photographs from several KNP locations to confirm the model. The model's 96% accuracy in animal photographs and videos taken day and night and in feed from contemporaneous location has shown its utility. The model also improves reliability by 1–13% over previous methods.

Tracking individual broilers on video in terms of time and distance

Tracking group-housed individual broilers using video can provide valuable information on their health, welfare, and performance, allowing breeders to identify novel or indicator traits that aid genetic improvement. However, their similar appearances make tracking individual broilers in a group-housed setting challenging. This study aimed to analyze broiler tracking on video (number of ID-switches, tracking time and distance) and examined potential tracking errors (ID-losses - location, proximity, kinematics) in an experimental pen to enable broiler locomotion phenotyping. This comprehensive analysis provided insights into the potential and challenges of tracking group-housed broilers on video with regards to phenotyping broiler locomotion. Thirty-nine broilers, of which 35 noncolor marked, were housed in an experimental pen (1.80×2.61 m), and only data at 18 d of age were used. A YOLOv7-tiny model was trained (n=140), validated (n=30), and tested (n=30) on 200 annotated frames to detect the broilers. On the test set, YOLOv7-tiny had a precision, recall, and average precision (@0.5 - Intersection over Union threshold) of 0.99. A multi-object tracker (SORT) was implemented and evaluated on ground-truth trajectories of thirteen white broilers based on 136 min of video data (1-min intervals). The number of ID-switches varied from 5 to 20 (mean: 9.92) per ground-truth trajectory, tracking times ranged from 1 (by definition) to 51 min (mean: 12.36), and tracking distances ranged from 0.01 to 17.07 meters (mean: 1.89) per tracklet. Tracking errors primarily occurred when broilers were occluded by the drinker, and relatively frequently when broilers were in close proximity (within 10 cm), with velocity and acceleration appearing to have a lesser impact on tracking errors. The study establishes a 'baseline' for future research and identified the potential and challenges of tracking group-housed individual broilers. The results highlighted the importance of addressing ID-switches, identified potential tracking algorithm improvements, and emphasized the need for an external animal identification system to enable objective, simultaneous and semi-continuous locomotion phenotyping of group-housed individual broilers.

RESAJANA ZOO – Aplikasi Pengenalan Hewan

Many people do not know the name of the animal if they do not see directly the characteristics of the animal, this is caused by a lack of knowledge about the characteristics of the animal and the lack of references regarding these characteristics even though knowing the characteristics of animals is very important for knowledge if someone is in the forest, at sea and in the air With these problems, an animal identification system is made by knowing the characteristics of the animal

English

Development of the livestock sector in South Africa has the potential to contribute significantly to national development. The aim of this study was to explore challenges and propose recommendations to support livestock sector development in South Africa using a stakeholder-driven approach. Interviews with high-level stakeholders across major livestock industries were conducted to gather insights. Responses were verified by a literature review and supplemented with an event study analysis on parameters for livestock production and exports. Interview respondents revealed livestock disease management in the face of increasing disease burden, high input costs, and limited government resources as a major challenge to national livestock-sector development. Implementation of national and international regulations on controlled diseases severely limits profitability. Emerging farmers are discouraged from investing in livestock disease management and are thus excluded from entering and remaining in the commercial value chains. Inclusive growth of the sector can be achieved through strengthening public-private partnerships through joint mentorship, training, and awareness-raising programmes. Evidence- and risk-based approaches to the national and international regulations and official guidance on outbreak management are needed. National animal identification and traceability systems and wider adoption of compartment systems for biosecurity, growth stimulants and poultry cages can enhance disease management and exports. Key words: African swine fever, event study analysis, foot-and-mouth disease, highly pathogenic avian influenza, livestock disease, livestock sector development, South Africa, stakeholder-driven approach.

Current Livestock Identification Systems

Traditional livestock identification and tracking techniques have been used for a while, but their performance is constrained by their susceptibility to theft, fraud, and security issues. This review discusses bioactive identifiers, biometric identifiers, barcodes, radio frequency identification and others as current means of identifying livestock and the potential benefits of using biometric identification systems such as improved accuracy and efficiency, as well as their ability to provide more reliable data than traditional identification methods. We also discussed the challenges and concerns associated with the use of biometric identification systems. Selecting the best method from the classical approaches involves various considerations, such as the size of the farm, the cost, and the required functions of the identification process. The article concludes by discussing the future of biometric identification systems in the livestock industry. We recommend that farmers should consider these factors in choosing the best animal identification system for their farms

ПРИМЕНЕНИЕ ЦИФРОВЫХ ТЕХНОЛОГИЙ ДЛЯ СНИЖЕНИЯ УГЛЕРОДНОГО СЛЕДА В ЖИВОТНОВОДСТВЕ

Today it is a generally accepted fact that the livestock industry has a negative impact on the environment. With the development of industrial production of livestock and poultry products, this impact has increased significantly. The risks of environmental pollution in animal husbandry are associated with the disposal of manure and manure, as well as greenhouse gas emissions during the entire production cycle. Therefore, one of the main tasks of modern animal husbandry is to reduce emissions. The purpose of the study is to develop a digital livestock management model aimed at improving production processes to reduce the harmful effects of the activities of an economic entity on the environment. As part of the study, a model was proposed that is a set of digital systems (reproduction system, animal identification system, microclimate system, feeding system, veterinary system, waste system, product control system) based on geographic information technologies. Digital systems cover all technological stages of animal breeding, care and fattening, production output. Robotic systems for feeding, milking, weighing, caring for animals perform their work more accurately and efficiently, which reduces feed losses, reduces the cost of animal treatment, increases their productivity and insemination rate. From the point of view of reducing greenhouse gas emissions, the use of drones for the delivery of feed and other materials used in animal husbandry is beneficial. The result of the introduction of digital technologies in the process of feeding animals is feed savings and efficient consumption, and as a result, a reduction in the carbon footprint due to reduced energy costs

PSXI-20 Satellite analysis and monitoring of rangelands of the Republic of Kazakhstan

Abstract Kazakhstan ranks 5th in the world in terms of pasture area, with 30% of land used as pasture. Evaluating pasture resources is relevant for improving the pasture lands regulation system. Since 2018, National Space Agency has carried out the Ministry of Agriculture space pastures monitoring program. The main task is a seasonal satellite evaluation of the degradation productivity degree pasture vegetation. Thematic analysis remote sensing data from KazEOSat-1, 2 satellites and territory geoinformation made it possible to determine, with a high accuracy: land boundaries and areas, seasonal state and average vegetation type pastures productivity evaluation, integration with land cadastre data identified unaccounted pastures, and animal identification system made it possible to assess pastures load. Seasonal state pastures evaluation done by satellite parameter choice. The main factor influencing vegetation spectral properties is biomass increase. Spectral reflectance decreases in spectrum’s red part and near-IR range increases, vegetation index value changes. The vegetation index was used to create correct linear weighing system: vegetation parameters values (the green plant biomass volume projective leaf cover of the soil). To calculate seasonal productivity, GNDVI vegetation index used to more chlorophyll concentration than NDVI index. GNDVI showed best correlation calculating productivity, concidering exposed soil atmosphere influence. Satellite evaluation showed digital maps (pasture.gharysh.kz). The area was counted according to condition degradation and productivity. In Karaganda region’s Zhanaarkinsky district brown soils are widespread, with little humus, natural and man-made desertification is observed. The minimum productivity of pastures was 0–3 c / ha, average - 7.7 c / ha, maximum - 12–15 c / ha. In 2020, 73% (4,364.0 thousand hectares) pastures were poor assessed, 18.4% (1,104.3 thousand hectares) - very poor. The reason is the location in the desert and semi-desert zones, the annual rainfall is 150 mm. Degraded pastures in 2020 amounted to 449.1 thousand hectares (93.8% degradation of a low degree, 5.8% of a medium degree, 0.4% of a strong degree).

Herd-level versus animal-level variation in methane emission prediction in grazing dairy cattle

In response to the increased concern over agriculture’s contribution to greenhouse gas (GHG) emissions, more detailed assessments of current methane emissions and their variation, within and across individual dairy farms and cattle, are of interest for research and policy development. This assessment will provide insights into possible changes needed to reduce GHG emissions, the nature and direction of these changes, ways to influence farmer behavior and areas to maximize the adoption of emerging mitigation technologies. The objectives of this study were to (1) quantify the variation in enteric fermentation methane emissions within and among seasonal calving dairy farms with the majority of nutritional requirements met through grazed pasture; (2) use this variation to assess the potential of new individual animal emission monitoring technologies and their impact on mitigation policy. We used a large database of cow performance records for milk production and survival from 2 398 herds in New Zealand, and simulation to account for unobserved variation in feed efficiency and methane emissions per unit of feed. Results showed an average of 120 ± 31.4 kg predicted methane (CH4) per cow per year after accounting for replacement costs, ranging 8.9–323 kg CH4/cow per year. Whereas milk production, survival and predicted live weight were reasonably effective at predicting both individual and herd average levels of per cow feed intake, substantial within animal variation in emissions per unit of feed reduced the ability of these variables to predict variation in per animal methane output. Animal-level measurement technologies predicting only feed intake but not emissions per unit of feed are unlikely to be effective for advancing national policy goals of reducing dairy farming enteric methane output. This is because farmers seek to profitably utilize all farm feed resources available, so improvements in feed efficiency will not result in the reduction in feed utilization required to reduce methane emissions. At a herd level, average per cow milk production and live weight could form the basis of assigning a farm-level point of obligation for methane emissions. In conclusion, a comprehensive national database infrastructure that was tightly linked to animal identification and movement systems, and captured live weight data from existing farm-level recording systems, would be required to make this effective. Additional policy and incentivization mechanisms would still be required to encourage farmer uptake of mitigation interventions, such as novel feed supplements or vaccines that reduce methane emissions per unit of feed.

Influence of inbreeding and some environmental factors on calving intervals in N'Dama cattle of Southwest Nigeria

Records of 4184 N'Dama calves born between 1947 and 1984 at Fashola Stock Farm, Oyo State, Nigeria was used to compute inbreeding coefficient and determine its effects on calving intervals of cows. Only 273 of the 4184 calves (6.53%) were inbred with an average inbreeding coefficient of 9.71%, while five out of 293 sires had inbreeding coefficient of 9.25% and 43 out of 1849 dams had inbreeding coefficient of 9.95%. The inbreeding coefficient for the entire population was 0.63%, respectively 0.70% and 0.58% for male and female calves. The average calving interval was 445.34 ± 2.68 days within the herd while year of birth, parity of dam, age of dam and dam's age at first calving all significantly (P<0.05) impacted calving intervals of cows. Season of birth and inbreeding coefficient of dam did not exert statistical (P>0.05) influence on calving intervals of dams. Although the number of inbred animals in the population is relatively low, however the inbreeding level of these animals is high, which may be attributed to the fact that breeding policy on the farm was not strictly adhered to, and lack of proper and adequate registration system coupled with inconsistent animal identification system on the farm may be responsible for the inbreeding observed in the N'Dama population at Fashola Stock Farm.

РАЗРАБОТКА ИНТЕЛЛЕКТУАЛИЗИРОВАННОГО СТАНКА ДЛЯ РАЗДЕЛЕНИЯ ПОТОКА ЖИВОТНЫХ ПО ЗАДАННЫМ ПРИЗНАКАМ

The Federal scientific Agroengineering center VIM has developed technical tools, algorithms and software for the intelligent automatic control system for milking animals “Stimul” on the “Herringbone” milking unit in three versions. The created system does not include automatic selection gates for effective management of zootechnical and veterinary services of animals. (Research purpose) The research purpose is in developing an intelligent machine for automatic sorting of animals for servicing and managing the herd according to specified characteristics. (Materials and methods) The article presents the development of control and management systems in dairy farming based on the conceptual principles of digital transformation. The digital control system is based on a multifunctional panel controller. The created control unit has a port for connecting to the RS 485 network and provides support for network functions via the Modbus Protocol. The programming of the control unit has been made in the SMLogix tool environment, which supports the FBD function block language. (Results and discussion) The article presents an intelligent machine for automatic sorting of animal flows for servicing and managing the herd according to specified characteristics with the unification of hardware, software modules and interface. The article describes the necessary parameters for the automatic remote animal identification system, the basic component of the control system of an intelligent machine for sorting animals according to specified characteristics. (Conclusions) The machine allows to automatically identify, sort and send animals to the specified areas for individual service.

Prototype of Integrated Livestock Recording Application with Animal Identification and Certification System in Kebumen

The Department of Agriculture and Food of Kebumen has utilized information technology to register livestock and provide identity cards to livestock with the Livestock Card Database system. Unfortunately, between the Livestock Card Database System and the Recording and Certification System was not integrated, so there is one job carried out by two officers. Data on making livestock cards comes from forms filled out by data officers in the village area from livestock, while data recording comes from recording officers from members of herding group / designated officers. The output of this research is in the form of an Android-based recording application prototype design, which is integrated with the livestock card database system as an identification system, and the recording and certification system as a system for animalcertification.

ENSURING THE MAINTENANCE OF THE AUTOMATED INFORMATIONAL SYSTEM “STATE ANIMAL REGISTER” IN THE REPUBLIC OF MOLDOVA

The goal of this work is to analyze and reflect the importance of developing the Integrated Agricultural Register for the Republic (IAR) of Moldova, which is based on the implementation of the process of Electronic Transformation of Government through the e-Agriculture principle, storage and maintenance of information systems in the agro-industrial sector, management and processing of required data from existing electronic registers in order to streamline the development and implementation of development policies in agriculture. The Animal Identification and Traceability System (AITS), later referred to as the State Animal Register (SAR), is only a component part of the IAR and has as its primary objective the food security and consumer assurance with products of animal origin. AITS is one of the basic subsystems as an integral part of the traceability process of animal products. The SAR Automated Informational System is a set of elements and procedures that allow the identification and registration of animals and holdings, ensuring compliance of the traceability principle. All animals belonging to the bovines, sheep, goats, swine, horses, donkeys and descendant species obtained by crossing them, all holdings in which these animals are kept or handled, either farm, sheepfold, fair, slaughterhouse, will be identified. [1]

An Automated Animal Classification System: A Transfer Learning Approach

Animal classification from images obtained by various techniques in forest become an important task to carry out focused distribution and abundance estimation. In the following paper a frame work based on Transfer Learning (TL) in a Convolutional Neural Network is proposed for the construction of an automated animal identification system. The framework is used to analyze & identify focal species in the images. A dataset of 6,203 camera trap images of 11 species including Wild pig, Barking deer, Chital, Elephant, Gaur, Hare, Jackal, Jungle cat, Porcupine, Sambhar, Sloth bear was obtained. Superior performance can be achieved by using Transfer learning in Deep Convolutions Neural Network (DCNN) for species classification. The accuracy achieved by the proposed model on the test dataset is 96% in 18 epochs by using batch-size of 32. This, in turn, can speed up research findings, construct more efficient and reliable animal monitoring systems, and consequently, save the time and effort of the Indian scientists. Therefore, having the potential to make significant impacts in the classification and analysis of camera trap images of the site under observation.

190 ASAS-EAAP Exchange Speaker Talk: Beyond stainless steel and milk pumps: precision technology in the milking equipment world

Abstract What do you think of when you hear the words: milking equipment? Do you think of the different brands? Boumatic, DeLaval, GEA or Lely? Do you think of parlours, parallels or rotaries? Do you think of teat cups, liners or rubber hosing? Do you think of teat dips and cleaners? If you think of mostly hardware – I can’t blame you. If I look at DeLaval, the company I work for, it has a 136-year history of offering dairy producers revolutionary innovations. In 1878, Gustaf de Laval patented the cream separator, which was the basis for forming our company in 1883. (That was even before Old MacDonald’s farm.) DeLaval has since made other significant improvements in animal health and welfare, milk production, food safety and labor efficiency with technologies like: the vacuum operated milking machine (1917); a commercial rotary (1930); animal identification systems (1978); variable speed vacuum pumps (1977); milking robots (late 1990s); an on-farm lab called Herd Navigator; Clover-shaped liners (2013); a teat spray robot (2015); a body condition score camera (2015); and our latest robot, the VMSTM V300 (2018). DeLaval has been pushing the Brave New World for many years. Today, we have a software program called DelProTM Farm Manager – a powerful data management tool providing valuable information and analysis. It enables the dairy producer to make efficient daily actions and for advisors to monitor and provide profitable and sustainable management actions. As animal scientists, what is your role in this? One avenue is to consider working with the data to provide relevant summaries on what the information is telling us. You can have a role in influencing management changes for dairy farms for the future. While not all farms have access to this data and information, your research projects can help dairy farmers wanting to make improvements to their operations. Producers need non-bias, third party sources they can reference. The dairy farming world is changing; we all need to keep up. Make sure the research that you decide upon fits into this rapid pace.