基于三维视觉的荷斯坦奶牛中遗传育种个体的最优选择

发布时间：2018-03-16 00:33

本文选题：三维视觉　切入点：荷斯坦奶牛　出处：《科技通报》2017年02期 　论文类型：期刊论文

【摘要】：提出基于三维视觉对荷斯坦奶牛进行体型线性评定,对奶牛个体实现规范化的生产性能测定,择优选取优良奶牛育种。首先,利用摄像分别采集奶牛三个方向的整体图像,运用图像采集卡数字转化技术将图像输入计算机,利用改进的Canny算子图像边缘检测算法对获取的奶牛图像进行去噪处理,通过调整和选取合适的阈值达到抑制噪声的目的;其次,将处理后的图片运用Fisher线性鉴别分析方法,归类不同种类奶牛图片样本,对归类后的图片进行标准图像和特征点模板匹配;最后,在目标特征中识别出体型线状符合优秀标准的荷斯坦奶牛进行育种繁育后代。实验证明,利用图像识别技术对奶牛的体型线性鉴定有助于选育出奶效率高、身体健康及对于机械挤奶适应较好的优质牛群。
[Abstract]:In this paper, the linear evaluation of Holstein cows' body shape based on 3D vision is put forward, the production performance of dairy cows is standardized, and excellent dairy breeding is selected. Firstly, the whole images of the three directions of cows are collected by using the camera. The image is input into the computer with the digital transformation technology of image acquisition card, and the improved Canny operator image edge detection algorithm is used to de-noise the obtained dairy cow image, and the noise suppression is achieved by adjusting and selecting the appropriate threshold value. Secondly, using Fisher linear discriminant analysis method, the processed images are classified into different kinds of dairy cow image samples, and the standard images and feature point template matching are carried out. Finally, The breeding and breeding of Holstein dairy cattle with line shape according to the excellent standard were identified from the target features. The experiment proved that linear identification of the body shape of dairy cattle by image recognition technology was helpful to the breeding efficiency of milk. Good health and good adaptability to mechanical milking.
【作者单位】：河南牧业经济学院自动化学院;
【基金】：河南省高等学校重点科研项目(16A520014);河南省高等学校重点科研项目(15B520009) 河南省科技厅重点科技攻关(152102210320) 河南省教育技术装备与实践教育研究重点项目(GZS026) 河南牧业经济学院科技创新团队(HUAHE2015006)
【分类号】：S823;TP391.41

【相似文献】