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.

Assessment and control of environment and thus animal comfort in production confinement is typically based on predetermined ambient temperature levels. This traditional approach often falls short in meeting the animals true thermal need because it does not integrate the effects of other contributing factors, such as drafts, humidity (particularly in hot conditions), radiation (in poorly insulated barns), floor type and/or condition (dry vs. wet), nutritional plane and health status of the animal. Singular and certain combined effects of physical and nutritional factors on swine have been researched and documented quite extensively over the years (Boon, 1981, 1982; Brody, 1945; Bruce and Clark, 1979; Close, 1981; DeShazer and Overhults, 1982; Hahn et al., 1987; Le Dividich, 1982; Mount et al., 1968, 1975; Riskowski et al., 1990; Sallvik, 1984; Ye and Xin, 2000; Xin and DeShazer, 1991; Xin et al., 1999, 2000). Yet it is formidable to measure all influencing factors to produce a comprehensive thermal comfort index for assessment and control. The best indicator of the environment adequacy and thus animal comfort is animals themselves that integrate both external and internal factors, which in turn lead to distinctive resting behaviors. Huddling, resting next to one another, and spreading are the stereotypical postural patterns of animals that experience cold, comfortable, and warm/hot sensation, respectively. Dedicated animal caretakers often use such behavioral patterns to fine-tune the ideal air temperature settings. However, it is laborious and impractical for the caretakers to perform such manual adjustments on a continual and consistent basis.

Computer vision offers a potential alternative to replace human observation of the animals and adjustment of control set-point. Previous work has examined feasibility and technical aspects of such an approach for assessment of thermal comfort based on image analysis of resting behavior of group-housed pigs (Geer et al., 1991; Hu and Xin, 2000; Shao, 1997; Shao et al., 1997, 1998). Though positive and promising, the results and data processing methods were limited to static conditions. For such an approach to be ultimately applicable to production conditions, it must be implemented on a real-time basis. Therefore, the objective of this endeavor was to explore a realtime computer vision system that allows continuous assessment and control of thermal comfort of group-housed pigs based on their resting patterns.