Development of precision livestock oriented high-speed serum vitamin A prediction method to produce high-quality beef with high marbling standard

为了生产高质量的大理石花纹牛肉，日本黑牛养殖户通常将牛的血清维生素A控制在最低所需的低水平。在本研究中，我们对瞳孔图像处理估计血清维生素A进行了实验研究。用2CCD相机拍摄了日本黑牛的瞳孔图像。眼底照相机也被使用。血常规测定血清维生素A水平。结果表明，维生素A水平较低的牛瞳孔图像的蓝色成分比趋于较小。采用5个变量（如红色成分比和强度）的多元分类方法。结果表明，轻度和重度维生素A缺乏症组可分类，分类率较高。为了实际应用，研制并测试了一种能够频繁捕获瞳孔图像的自动成像设备。

To produce high-quality beef with high marbling standard, Japanese Black cattle farmers usually manipulate the serum vitamin A in cattle at a minimum desired low level. In this study, we investigated experimentally on the estimation of the serum vitamin A by processing the pupil image. Pupil images of Japanese Black cattle were taken with a 2CCD camera. A fundus camera was also used. Blood assay was done to measure the serum vitamin A level. The result showed that the blue component ratio of the pupil image of cattle with lower vitamin A level tended to be small. Multivariate classification method using five variables (such as the red component ratio and intensity) was used. The result showed that mild and severe vitamin A deficiency groups could be classified with high classification rate. For the practical use, an automatic imaging equipment that can capture the pupil image frequently was made and tested.