The logic behind the study rests on a simple biological fact: how a cow is milked matters enormously. An effective milking routine is characterized by a low-stress environment, adequate stimulation before the milking unit is attached, and consistent pre-milking preparation and post-milking teat disinfection. When these steps are performed properly, farms see fewer intramammary infections, better milk quality, higher milk yields, and faster, more efficient milking. When they are performed sloppily — a cow cleaned with a dirty towel, forestripping skipped, a unit attached too soon after stimulation — the consequences ripple through the bulk tank. Prep-lag time, the interval between the first touch of the udder and the attachment of the milking unit, has been specifically associated with shorter milking times and lower odds of short-term teat tissue changes. Because so much depends on the consistent execution of each step, verifying that the routine is actually followed is one of the most important, and most tedious, tasks in dairy herd management.

Traditionally, that verification has meant parlour audits: a consultant or veterinarian stands in the parlour for hours, stopwatch in hand, watching an operator work through shift after shift. These audits are widely used to monitor compliance and identify deficiencies in milking procedures, but they have well-known weaknesses. Prolonged in-person observation may not always be feasible, and a single observation may not accurately represent routine performance because of the Hawthorne effect — the behavioural changes that occur when individuals know they are being watched. Audits can also miss employees on different shifts, meaning a problem worker might never be observed at all. Remote video auditing has emerged as an alternative when in-person assessments are impractical or cost-prohibitive, and such systems have already been implemented in hospitals and food preparation settings to evaluate worker performance, with evidence of improvements driven by increased task awareness and timely feedback. Computer vision, the branch of artificial intelligence concerned with the automated interpretation of images and video, promises to cut the cost of video-based assessment dramatically — but, as with any AI technology, it must first be validated.

The system put to the test, developed by Cattle Care Inc., incorporates multiple convolutional neural networks trained on video from a multi-parlour setup. Object detection is performed using YOLO — You Only Look Once, version 8.0 — alongside modified detection models to identify the key elements of the scene: humans, cows, and milking equipment. Vision-language models and convolutional neural networks then classify specific milking events. From the raw video, the software captures the key events throughout the milking process and derives performance metrics at the cow, turn, and parlour levels. It can also flag deviations from a farm’s standard milking protocol, such as lack of forestripping or teat-end cleaning, missed pre- or post-milking teat dipping, and the use of one towel on more than one cow, as well as specific operator behaviours including smoking, cellphone usage, and interactions that may compromise cow welfare. When the system flags a critical event with low confidence or identifies an unusual farm-specific event, the case is subject to human verification — though no such verification was needed for the footage in this study.

The validation took place on two commercial dairy farms in Southwestern Ontario, chosen because they offered two very different parlour configurations. Farm A milked 350 cows in a 10-by-2 herringbone parlour with an average daily production of 41 kilograms of milk per cow, while Farm B milked 110 cows in an 8-by-2 parallel parlour averaging 40 kilograms per cow. Both farms milked three times daily with a single operator per shift, and their protocols differed slightly: Farm A worked in groups of five cows with teat pre-dipping, cleaning with a cloth towel, unit attachment, and post-dipping with a dip cup, while Farm B milked groups of four and added forestripping before pre-dipping. The researchers installed a single camera at 2560-by-1920 resolution recording at 20 frames per second, positioned opposite the cows’ entrance to ensure full visibility of the milking units and the operator. Cameras remained in place for two days, and footage was stored on an SD card before being transferred to the software for processing.

The study, reported using the STARD-AI guidelines for diagnostic accuracy studies using artificial intelligence, compared four one-hour segments of uninterrupted milking shifts — two per farm, selected by the software itself — analysed independently by the system and by two trained human observers. The observers were trained on practice footage with review and feedback from a veterinarian experienced in milking routine evaluation, and their inter-observer agreement during training was excellent, with no differences in detecting protocol deviations and timing disagreements of no more than two seconds. Each observer watched the video segments blind to the other’s findings and to the software’s output. The variables of interest were prep-lag time measured per cow, average unit-on time per milking turn, turn time — the time to milk one side of the parlour from gate opening to cows exiting — and the count of events inconsistent with the milking protocol.

The statistical agreement between machine and humans was, by conventional standards, as good as agreement can get. Using Lin’s concordance correlation coefficient, where values above 0.75 are classified as excellent, the system scored 0.999 for mean unit-on time per turn across 44 turns, 0.995 for prep-lag time across 214 cows, and 0.998 for turn time across 25 turns. The human observers themselves agreed with each other at 0.998 for mean unit-on time, confirming a solid reference standard. Bland–Altman plots revealed only small, consistent biases: human observers timed milking per cow about half a second shorter than the software, prep-lag time about 3.8 seconds shorter, and turn time about 1.7 seconds longer, with limits of agreement of just a few seconds in each case. The authors note that the prep-lag discrepancy represents only 3.1 to 4.2 percent of the recommended prep-lag range of 90 to 120 seconds for cows milked three times daily, and that a correction factor would have little practical impact. One outlier in turn time — a 26-second difference — stemmed from a gate-opening delay and cows entering the parlour that the software failed to identify.

The event detection results were even more striking. Human observers flagged 51 events outside the established protocol across the four hours of footage: three instances of cellphone usage during milking, 15 manual detachments, one instance of missed forestripping on Farm B, and 32 towel errors, including multiple cows cleaned with a single towel or a cow missed entirely. The computer-vision system detected every one of those 51 events. It then went further, flagging six additional towel misuse events in which operators flipped used towels and used the clean side on the next cow — a subtle manoeuvre that both human observers had overlooked. When the researchers went back to the recordings to adjudicate, all six events were confirmed to have been correctly identified by the software. That performance yields a sensitivity of 100 percent and a positive predictive value of 100 percent, producing a perfect F-score of 1. Missed pre- or post-milking teat dipping and smoking events did not occur during the observation periods, so the system was not tested on those categories.

The authors argue that the value of this technology may not lie merely in reproducing human observations, but in complementing them. Observational assessments are inherently subject to attention, distraction, fatigue, and differences in awareness, whereas automated systems can continuously monitor predefined events without being affected by these observer-related factors. Continuous monitoring over extended periods offers clear advantages over single-point, in-person audits: consistency in the milking routine has been associated with greater milk yield, while the repetitive nature of milking work can contribute to procedural drift over time. Automated monitoring could therefore catch deviations that would otherwise go unnoticed, supporting improvements in milking consistency with potential downstream benefits for milk quality, udder health, and employee performance. Those longer-term outcomes, however, were beyond the scope of the present validation study and remain to be demonstrated.

The researchers are careful to note the limitations. The study involved only two farms, which restricts the ability to assess generalizability across the broader range of parlour configurations and milking practices, and the limited observation window cannot exclude a Hawthorne effect that may have reduced the frequency of deviations — a factor that could have inflated agreement between methods. Some aspects of milking routine assessment, such as teat-end condition, cleanliness, and post-dipping coverage, require close-up inspection that the camera’s field of view simply cannot provide, though other teams are exploring machine-learning approaches for those tasks. And while the software substantially reduces the need for manual auditing, it cannot replace the trained eye for everything. Still, under the conditions tested, the system demonstrated high agreement with human observations on every metric assessed while catching subtleties the human eye missed — a result that suggests the future of milking parlour oversight may be a quiet camera in the corner, watching every cow, every shift, without ever blinking.

Subject of Research: Validation of a computer-vision system for monitoring milking routines in dairy parlours

Article Title: Validation of a computer-vision milking parlour monitoring system

Article References: Validation of a computer-vision milking parlour monitoring system. (n.d.). https://doi.org/10.3389/fanim.2026.1959684

Image Credits: AI Generated

DOI: 10.3389/fanim.2026.1959684

Keywords: artificial intelligence, computer vision, dairy cattle, milking routine, precision livestock farming, udder health, milk quality, milking parlour, remote video auditing, validation study, protocol compliance, machine learning

The research, published in Food Science of Animal Resources, followed 400 lactating Holstein cows across three commercial dairy farms in Jeju City over 47 months, from January 2021 to November 2024. During this period the farms progressively replaced their conventional concentrate feeds with commercially manufactured methane-mitigating concentrates, allowing the researchers to compare milk records before and after the transition under genuine production conditions. Two products were involved: one containing a yeast-based additive, garlic-derived diallyl disulfide, and condensed tannins, and another built around a commercial essential-oil blend marketed as Agolin. The full composition of the additives is proprietary, a common feature of the emerging methane-mitigation market that itself complicates independent scientific evaluation. All cows received a total mixed ration formulated to Korean feeding standards, delivered twice daily, with free access to water and minerals.

The headline result is straightforward: milk volume survived the transition intact. Across all statistical models, including multifactorial general linear models and linear mixed-effects models that treated individual cows as random effects, methane-mitigating feeding showed no significant effect on daily milk yield. For dairy farmers weighing whether to adopt climate-friendly rations, this is the single most commercially consequential finding, because any drop in milk output would translate directly into lost revenue and would likely doom adoption regardless of environmental benefits. The result also aligns with earlier experimental work showing that additives such as 3-nitrooxypropanol, which can cut enteric methane by 24 to 40 percent, and the red seaweed Asparagopsis, which has suppressed methane by more than 80 percent in controlled trials, generally leave production untouched.

Beneath the calm surface of total yield, however, the milk itself changed. Cows receiving the methane-mitigating concentrate showed significantly lower concentrations of milk protein, solid non-fat, and milk urea nitrogen, along with a modest but statistically consistent increase in somatic cell count, an indicator of udder health and immune activity. Milk urea nitrogen is a window into nitrogen metabolism: when rumen microbes capture dietary nitrogen inefficiently, excess ammonia is converted to urea and appears in blood and milk. Lower MUN in the methane-mitigating group therefore hints that the additives altered rumen fermentation in ways that shifted how the animals handled nitrogen, a plausible consequence of additives that redirect hydrogen away from methanogenic archaea and thereby reshape the entire microbial economy of the rumen.

The somatic cell count finding deserves careful handling. The increase was moderate and does not indicate clinical mastitis, and the authors explicitly caution against causal interpretation. Still, the signal appeared across all three farms, reaching statistical significance everywhere, which lends it weight. One mechanistic explanation involves the gut-immune axis: dietary components that alter rumen fermentation can shift ruminal pH and microbial turnover, promoting the release of lipopolysaccharides, molecular fragments of bacterial membranes that act as inflammatory triggers and can elevate somatic cell counts even in the absence of infection. Previous trials with 3-nitrooxypropanol and Asparagopsis have reported similarly inconsistent somatic cell outcomes, suggesting this may be a recurring, if subtle, trade-off in methane-mitigation feeding.

Perhaps the most striking feature of the study is how context-dependent the responses proved to be. Milk fat fell significantly on Farm A but not on the other two. Milk protein dropped on Farm C alone. Solid non-fat declined on Farms B and C but not on A. Interactions between feed type and farm, parity, and month were significant for nearly every compositional trait, meaning the effect of the same additive family depended on where the cow lived, how many lactations she had completed, and what time of year it was. Farm-level differences in forage quality, hygiene protocols, microclimate, and baseline nutrition likely explain much of this heterogeneity, and the authors argue that any policy or implementation framework must account for it.

Parity emerged as a particularly powerful moderator. Milk yield climbed steadily from the first through the fifth lactation before plateauing and declining, while protein, solid non-fat, and MUN drifted downward with age, consistent with the cumulative metabolic stress and reduced hepatic and mammary efficiency of older cows. Against this backdrop, the methane-mitigating feed produced parity-specific effects: milk yield rose significantly in sixth-parity cows but fell in third and fourth parity animals, solid non-fat declined consistently across parities three through six, and somatic cell counts rose in several mid-life parities. MUN was lower across all parities, though significance faded after the second. These patterns suggest that younger, more metabolically resilient cows absorb dietary perturbation more easily, while older animals, already operating closer to their physiological limits, respond more visibly.

Seasonality added another layer of complexity. Milk protein and solid non-fat dipped during the hot summer months of June through August, a classic heat-stress signature reflecting reduced feed intake and impaired rumen function. Yet the methane-mitigating group’s solid non-fat decline progressed gradually from summer into winter rather than showing a sharp heat-driven drop, and MUN differences persisted across most months with only scattered exceptions. The monthly patterns therefore do not map neatly onto thermal stress alone, implying that individual cow variability, forage quality, and intake dynamics modulate how the feed expresses itself through the calendar. Notably, somatic cell counts remained stable across months, suggesting mammary health in these herds was governed more by parity structure and farm hygiene than by climate.

To distill these tangled effects, the team ran stepwise regression guided by the Schwarz Bayesian Criterion, testing whether methane-mitigating feeding independently explained variation in each trait once farm, parity, and month were considered. The answer was nuanced. Feed group was not retained for milk yield, fat, or protein, indicating those traits are governed by structural factors. But it was retained with strong statistical support for solid non-fat, MUN, and log-transformed somatic cell count, and its inclusion improved model fit, most notably for MUN. The overall explanatory power of the models was modest, with adjusted R-squared values between roughly 0.07 and 0.20, which the authors attribute to the enormous unmeasured variability inherent in field data, from genetics and individual intake to rumen microbial communities.

The study carries an important caveat that tempers its climate message: no methane was actually measured. The feeds were formulated with mitigation in mind, but the research assessed downstream consequences on milk, not upstream effects on emissions. The authors are explicit that the results should not be read as proof of methane reduction. What the study does establish is feasibility and safety at scale: commercially available methane-mitigating concentrates can be deployed across real farms for years without sacrificing milk volume, while producing detectable, context-sensitive shifts in nitrogen-related milk traits and a modest rise in somatic cell counts. The next generation of research, the authors argue, must pair direct methane measurement with rumen fermentation analysis, inflammatory marker monitoring, and cow-level physiological profiling. Until then, the Jeju experiment stands as a rare and valuable bridge between the optimistic world of controlled methane trials and the messy, variable reality of the commercial dairy farm, where the climate solution ultimately has to work.

Subject of Research: Field evaluation of methane-mitigating concentrate feed effects on milk production, composition, and udder health in commercial Holstein dairy cows

Article Title: Impact of methane-mitigating concentrate feed on milk production, composition, and udder health in Holstein cows

Article References: Lee, S., Ko, K., Kim, G., Park, J.-E., & Ryu, Y. (2026). Impact of methane-mitigating concentrate feed on milk production, composition, and udder health in Holstein cows. Food Science of Animal Resources, 46(1), Article 49. https://doi.org/10.1007/s44463-025-00033-w

Image Credits: AI Generated

DOI: 10.1007/s44463-025-00033-w

Keywords: methane mitigation, dairy cows, Holstein, milk composition, milk urea nitrogen, somatic cell count, enteric fermentation, feed additives, greenhouse gas emissions, udder health, commercial dairy farms, sustainable dairy production