Skip to main content
ARS Home » Plains Area » Clay Center, Nebraska » U.S. Meat Animal Research Center » Nutrition, Growth and Physiology » Research » Publications at this Location » Publication #340910
Title: Evaluation of using a depth sensor to estimate the weight of finishing pigs

Author
item CONDOTTA, ISABELLA - Brazil University
item Brown-Brandl, Tami
item SILVA-MIRANDA, K - Brazil University

Submitted to: European Conference on Precision Agriculture Proceedings
Publication Type: Proceedings
Publication Acceptance Date: 7/2/2017
Publication Date: 9/1/2017
Citation: Condotta, I., Brown-Brandl, T.M., Silva-Miranda, K.O. 2017. Evaluation of using a depth sensor to estimate the weight of finishing pigs. Proceedings of the 8th European Conference on Precision Livestock Farming, September 12-14, 2017, Nantes, France. 495-502.

Interpretive Summary:

Technical Abstract: A method of continuously monitoring weight would aid producers by ensuring all pigs are healthy (gaining weight) and increasing precision of marketing. Therefore, the objective was to develop an electronic method of obtaining pig weights through depth images. Seven hundred and seventy-two images and weights were acquired from 4 different ages (8, 12, 16, and 21 weeks) of finishing pigs (a mix of gilts and barrows) of three sire-lines (Landrace, Duroc and Yorkshire). Weights ranged from 10.8 – 125.7 kg. The images were analyzed using MATLAB image processing toolbox and summing the columns to calculate the volume. Sixty percent of the data was used for equation development, and 40% was used for testing. Individual equations for weight predictions by volume were developed for gilts and barrows and for the three sire-lines. A global equation using the combined data was developed and then compared with individual equations using the Efroymson’s algorithm. The results showed that there was no significant difference between the global equation and the individual equations for barrows and gilts (p<0.05), and the global equation was also not different than individual equations for each of the sire-lines (p<0.05). In addition, the results from the global equation indicate that volume accounted for 99.05% of the variation in weight. Using the test data set, the global equation predicted weights using volume calculated with an average error of 4.6% or 2.2 kg. Therefore, the results of this study show that the depth sensor would be a reasonable approach to continuously monitor weights.