Year: 2024
Type: conference-abstract
Abstract: New technologies such as automatic in-pen weighing of cattle offer great opportunities to monitor, improve and control livestock production. The large amount of raw BW data recorded must be processed to estimate BW minimizing the impact of erroneous records and the large variability due to various factors such as fill of the gastrointestinal tract. The objective of the present study was to compare two data processing methods of remotely collected body weight (BW) in feedlot steers against traditional weighing measured in central handling facilities. The BW data were collected from a serial slaughter trial with 26 steers of each Angus, Brahman, and Charolais slaughtered at 50, 100, 150, and 200 DOF (initial BW = 345 ± 29.9 kg). Body weight was measured every 50 d in the central handling facilities (stationary BW in a weighing box) and then using an in-pen weighing system setup in front of the water trough to record animal BW linked to RFID while drinking water. The remotely collected BW data were processed with two smoothing: 1) penalized b-splines (BWpbs) and 2) natural cubic splines (BWncs). After fitting the data from each animal to each smoothing algorithm, datapoints greater or less than the estimated BW × 1.5 residual error were deleted. The BW estimated from remotely collected data with each method was regressed against traditional BW. The former algorithm resulted in a linear regression of the form BWpbs = -13.1 (± 1.46) + 1.04 (± 0.003) BW (R2 = 0.996; RMSW = 6.46 kg). The natural cubic spline algorithm resulted in a linear regression of the form BWncs = -11.0 (± 3.02) + 1.04 (± 0.007) BW (R2 = 0.985; RMSW = 12.88 kg). Both algorithms had acceptable accuracy and precision although the penalized b-spline algorithm was more sensitive because it better detected sudden changes in BW. The intercept showed that both algorithms estimated heavier BW compared with traditional BW, likely because of BW loss due to defecation and urination during yarding. In conclusion, penalized b-splines smoothing algorithm is recommended over natural cubic splines to process remotely collected BW because of decreased error and greater precision.
Source: Journal of Animal Science
Institution: The University of Sydney
Language: English
Topic: Smart Agriculture and AI
Subfield: Plant Science
Domain: Life Sciences
SDG: Zero hunger
Open Access status: bronze
OpenAlex ID: W4402533461