Bach A, Iglesias C, Busto I
Institució Catalana de Recerca i Estudis Avançats (ICREA), 08010 Barcelona, Spain.
J Dairy Sci. 2004 Dec;87(12):4207-9. doi: 10.3168/jds.S0022-0302(04)73565-1.
The objective of this study was to develop and validate a computerized system to monitor feeding behavior and feed intake of loose-housed dairy cattle. The system consisted of 28 scales located in front of each self-locking place of a regular feedbunk. All cows had access to all scales indifferently. Each visit to the feedbunk was monitored by a transponder in the ear of each cow that was detected by a proximity reader located at the top right corner of each headlock. The data from the scales and the proximity readers were continuously recorded by a computer with an average scanning time of 3.5 s. The monitoring system was validated using all 28 feeding places and 51 lactating cows in a series of 4-h observations during 5 different d. During the observation periods, for each feeder, 2 observers recorded the cow number and the exact time of the visit. The observed data were then compared with the computer records. To validate the ability of the system to monitor feed consumption, on separate days, the amount of feed consumed by a cow during a visit was also measured manually with an external scale, and the feed that disappeared from each scale in 2 different 24-h periods was compared with the sum of feed consumed in each scale during these 2 periods. The average time spent in a given scale by each cow determined by direct observations was similar to that determined by the computer. The system was accurate and showed a high specificity (98.8%) and sensitivity (99.6%) for cow detections. Feed weights determined by the computer system were similar to those measured manually with an external scale, implying that the system was also accurate in measuring individual intake weights. In conclusion, the system provided a reasonable estimate of the number of visits per animal, length of each visit, amount of feed consumed per visit and animal, the total amount of feed consumed daily by each animal, and the rate at which animals consume feed.
本研究的目的是开发并验证一个用于监测散养奶牛采食行为和采食量的计算机系统。该系统由位于常规饲料槽每个自锁位置前方的28个秤组成。所有奶牛均可无差别地使用所有秤。每头奶牛每次到饲料槽的访问由其耳部的应答器进行监测，该应答器由位于每个头枷右上角的接近读卡器检测到。秤和接近读卡器的数据由一台计算机持续记录，平均扫描时间为3.5秒。在5个不同的日子里，通过对28个采食位置和51头泌乳奶牛进行一系列4小时的观察，对监测系统进行了验证。在观察期内，对于每个采食器，2名观察者记录奶牛编号和访问的确切时间。然后将观察到的数据与计算机记录进行比较。为了验证该系统监测饲料消耗的能力，在不同日子里，还使用外部秤手动测量奶牛一次访问期间消耗饲料的量，并将2个不同连续24小时期间每个秤上消失的饲料量与这两个期间每个秤上消耗的饲料总量进行比较。通过直接观察确定的每头奶牛在给定秤上花费的平均时间与计算机确定的时间相似。该系统准确无误，对奶牛检测显示出高特异性（9８.8% ) 和高灵敏度（99.6% ) 。计算机系统确定的饲料重量与使用外部秤手动测量的重量相似这意味着该系统在测量个体采食量方面也很准确。总之，该系统合理地估计了每头动物的访问次数、每次访问时长、每次访问和每头动物消耗的饲料量、每头动物每天消耗的饲料总量以及动物消耗饲料速率。
