A vision-based instrument for measuring milk somatic cell count

The method of rapidly measuring the somatic cell count of raw milk is important for detecting the quality of milk and monitoring the health of dairy cows. This paper investigates a vision-based measurement algorithm and develops an instrument for measuring the somatic cell count. In detail, first, microscopic images of glass slides are automatically captured. Second, the Hough transform is used to calibrate the microscopic image so as to compute the sample volume. Third, a discrete optimization model is set up to determine the image segmentation threshold. Fourth, the features of milk somatic cell are expressed in area, ratio of length to width and circularity. The least-square circle method is employed to detect the circularity of the milk somatic cell. Then, an automatic milk somatic cell count is achieved via a recursion algorithm. Finally, a milk somatic cell counting instrument is developed based on the proposed algorithm, which is verified to be applicable via comparison with the FossMatic 5000, made in Denmark. (paper)