Machine Vision Based Yak Body Size Measuring Points

[摘要] In livestock breeding, it is a research hotspot to conduct non-contact measuring and acquire the body size parameters of animals without stirring stress. And traditional animal body size manual measurement methods involve heavy workload and safety risks. This paper aims to acquire the color image of yak by shooting the traditional body size measurement position of yaks in Sanjiangyuan area. Then, a classification processing based foreground extraction method is used to extract the yak from the image; after the profile curve of the yak is extracted, the temple point and ischium end of the yak are obtained by means of calculating the distance between the image border and the central point. Finally, the body height and body dip length points of the yak are acquired. This paper specially extracts the yak in a yak image shot in Sanjiangyuan area, Qinghai Province, marks points and displays them. The experiment suggests that the dark yak body size point marking algorithm studied by this paper is quite effective to point marking of two-dimensional yak images.

[发布机构] Department of Computer Technology and Applications, University of Qinghai, Xining; 810016, China^1;Interdisciplinary Graduate School of Agriculture and Engineering, University of Miyazaki, Japan^2;Yushu Prefecture Animal Husbandry and Veterinary Station, Yushu Qinghai; 815000, China^3

[效力级别] 无线电电子学 [学科分类] 计算机科学（综合）

[关键词] Foreground extraction;Heavy workloads;Livestock breeding;Manual measurement methods;Marking algorithm;Non-contact measuring;Profile curves;Qinghai Province