Click on “Download PDF” for the PDF version or on the title for the HTML version.

If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.

A real-time sow behavior analysis system to predict an optimal timing for insemination

Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan

Citation:  10th International Livestock Environment Symposium (ILES X)  .(doi:10.13031/iles.18-023)
Authors:   Jacquelin Labrecque, Joël Rivest
Keywords:   Animal behavior, Animal breeding, Behavior, Behavior analysis, Data analysis, Insemination management, Machine learning, Pig behavior, Pig breeding, Pig farming, Precision breeding, Precision livestock farming.

Abstract. Ovulation in sows cannot be observed directly – at the commercial farm level – with the currently available methods. Therefore, estrus is taken as the indicator of the best timing for insemination. Ovulation generally occurs at the two thirds through the standing heat period (Martinat-Botté, 1997). To maximize fertility, pig producers generally inseminate once every 24 hours during that period. Heat detection is a continuous challenge for swine production systems. It is time consuming and requires skilled workers. We propose a novel approach to determine an optimal timing for insemination in sows based on behavior analysis. Our system - called PigWatch - continuously collects behavior data from sensors mounted on stalls. Real-time analysis allows to find the statistically optimal timing for insemination based on behavior pattern recognition. Our system was used for 21 months in 7 commercial farms for 20,485 weaned sows. Our approach, when used as a complement to a daily heat detection, allowed to easily manage 99.78% of the sows. Without the daily heat detection, the algorithm could have precisely managed 79.46% of the sows. The developed system should be considered as a tool to use in synergy with the workers to achieve precision breeding. Commercial farm using our approach made an average of 1.32 inseminations per estrus and had an average conception rate of 89.52%. Precision breeding, as proposed here, has the potential to decrease skilled labor dependency, improve reproduction results, optimize the use of the best boars, and ultimately accelerate genetic improvement.

(Download PDF)    (Export to EndNotes)