PS1-15. Comparing Use of UAV Imagery for Cattle Headcounts to Conventional Headcount Practices in a Commercial Feedyard

Accurate cattle headcounts are critical in commercial feedyards, especially under limit-fed programs where precise numbers ensure animals receive the feed needed to maximize performance. This study evaluated accuracy of unmanned aerial vehicle (UAV)–based cattle headcounts compared to conventional management software records and explored implications of headcount accuracy on feed delivery efficiency under a limit-fed system. UAV images were collected over nine weeks across four lots of commercial beef heifers (ranging 245 kg to 281 kg; 36 to 88 head per pen) near Minneola, Kansas. Animals were limit-fed a growing cattle ration (1.26 Mcal/kg NEg) for the duration of the trial targeting 1.36 kg ADG. UAV images were captured twice weekly on non-consecutive mornings between 0600 and 0800 and analyzed using imaging software to determine a method-specific headcount (UAV method). Data were also recorded from the AMS management platform (Turnkey Computer Systems, Loveland, CO), including method-specific pen headcount (Management Software method), feeding time, feed delivered per pen, and pulls to the hospital pen. For each pen, dry matter intake (DMI; kg/hd) was calculated by dividing total feed delivered by the headcount determined by each method. Feeding consistency was evaluated using variability in DMI across days within each period, expressed as the DMI range (maximum minus minimum DMI over each 3-week period). Data were analyzed using the lmer package in R, with method and period as fixed effects and pen as a random effect. Variables analyzed included headcount, DMI, standard deviation (SD) of DMI, and DMI range. No method × period interactions were observed for any variable (P > 0.05). Headcounts and resulting DMI estimates did not differ between UAV and Management Software methods (P > 0.05), indicating strong agreement between technologies. DMI range differed by period (P < 0.01); period 1 had a greater range than periods 2 and 3 (P < 0.01), with no difference between periods 2 and 3 (P = 0.23). Period effects on standard deviation of DMI matched those of DMI range (P < 0.01). Declining DMI variability over time aligns with typical feedyard management as cattle transition through diet step-up during the first 21 days before reaching consistent intake under limit-feeding. Overall, UAV-based headcounts demonstrated accuracy comparable to software-generated counts and can reliably characterize feeding management and intake dynamics in growing cattle.