Duck eggs' freshness detection is one of the most important steps in the process of producing, selling and processing. The morphological characteristics of egg's air chamber and yolk are closely related to its freshness. In this paper, duck eggs' color images were collected by machine vision technology. Then, pretreatments were conducted on images to remove background. Gradient method was employed to track edge. The position of air chamber was judged firstly. Then, the boundary of duck egg's air chamber was extracted by using Hough transformation to detect straight line. Thus, the air chamber area was obtained, and the air chamber area to the whole area ratio was calculated. At the same time, the ratio of egg yolk area to the whole area and the mean value of images' R, G, I components were calculated. This five values were used as parameters. All of the samples were divided into training set and prediction set according to the ratio of two to one. The least squares support vector machine was chose to establish model for duck eggs' freshness classification. The results showed that the accuracy rate of training set and prediction set were 96.92% and 93.85% respectively. This showed that this machine vision method was feasible to the non-destructive detection and classification of duck eggs' freshness. © 2017, Editorial Office of Journal of CIFST. All right reserved.
