基于红外热成像的育成期蛋鸡体温检测方法

Body temperature detection method of laying hens in rearing period based on infrared thermography

针对育成期蛋鸡在散养状态下受环境影响大,易生病,体温检测困难的问题,采用红外热成像技术和多元回归分析的方法,对鸡体体表温度和翼下温度之间的关系进行研究,建立温度预测模型。环境温湿度会影响鸡体体表温度的测量,为使红外热像仪采集的体表温度能够准确的反应出鸡体真实温度,试验对鸡体翼下温度和体表温度进行同步测量,并采集饲养环境中的温湿度,通过分析各变量之间的相关性,找到能够反应鸡体真实温度的变量,建立最优温度预测模型。试验结果表明:1)散养状态下育成期蛋鸡羽毛覆盖区域温度和环境温度存在共线性;2)多元线性温度预测模型的平均相对误差为0.38%;3)多元非线性温度预测模型的平均相对误差为0.17%。表明红外热像仪可以用于蛋鸡的体温检测,而且多元非线性温度预测模型结果更加准确。

Aiming at the problems that laying hens in the rearing period are greatly affected by the environment,easy to get sick,and their body temperatures are difficult to be detected in free-range state,the infrared thermal imaging technology and multiple regression analysis are used to analyze the relationship between the body surface temperature and the under-wing temperature,and to establish a temperature prediction model.The environmental temperature and humidity affect the measurement of the body surface temperature.Therefore,to make sure the body surface temperature collected by the infrared camera can accurately reflect the true temperature of the body,synchronous measurement is carried out to measure the temperatures under the wings and the body surface of laying hens temperature,and the temperature and humidity in the breeding environmentare are also collected.Through analyzing the correlation between the variables,it is found that the variables that can reflect the true temperature of the body,and can be introduced to establish an optimal temperature prediction model.The results show that:1)The temperature of the feather-covered area and the ambient temperature of the laying hens in the free-range state are collinearity;2)The average relative error of the multiple linear temperature prediction model is 0.38%;3)The average of the multiple nonlinear temperature prediction model,and the relative error is 0.17%.Inconclusion,the infrared thermal imaging camera can be used to detect the body temperature of laying hens,and the results of the multivariate nonlinear temperature prediction model are more accurate.