Publication: An Embedded System Device to Monitor Farrowing
Issued Date
2018-11-20
Resource Type
Other identifier(s)
2-s2.0-85059959661
Rights
Mahidol University
Rights Holder(s)
SCOPUS
Bibliographic Citation
ICAICTA 2018 - 5th International Conference on Advanced Informatics: Concepts Theory and Applications. (2018), 208-213
Suggested Citation
Piyanuch Silapachote, Ananta Srisuphab, Warot Banchongthanakit An Embedded System Device to Monitor Farrowing. ICAICTA 2018 - 5th International Conference on Advanced Informatics: Concepts Theory and Applications. (2018), 208-213. doi:10.1109/ICAICTA.2018.8541287 Retrieved from: https://repository.li.mahidol.ac.th/handle/20.500.14594/45542
Research Projects
Organizational Units
Authors
Journal Issue
Thesis
Title
An Embedded System Device to Monitor Farrowing
Other Contributor(s)
Abstract
© 2018 IEEE. Crucial for successful farrowing in swine production is the monitoring of health and welfare of pigs from insemination to farrowing and lactation. Although most are naturally farrowed, a significant number of piglets are lost due to fetal hypoxia. Common causes are ruptured umbilical cords, getting stuck in the birth canal, and exhaustion of sow. Prompt assistance offered by farmers can save lives of many piglets. Time is critical. Minutes of delay could mean another loss. Around-the-clock monitoring, at the same time, is not only labor-intensive but also a financial burden for farm management. One of the keys toward better assessment of the right timing for farmworkers to attend to a farrow with minimal wait-around time - first and foremost - is to collect and analyze the detailed timing of the farrowing process. Bringing digital technology of embedded systems to farming, we developed a monitoring device capable of continuously recording a video of a sow. Farrowing videos were collected for seven weeks. Graphical visualization and statistical analysis of the data were evaluated. Employing computer vision and machine intelligence, we proposed a methodology for extracting features and training a classifier to automatically detect the firstborn piglet. Preliminary image processing results are presented.