The invention relates to a
machine vision technology based method for rapidly detecting egg
internal quality, belonging to non-
destructive testing technologies for agricultural and animal products. According to the method, a
machine vision technology is used for acquiring characteristics on three aspects including a
color space, an
eggshell shape and an egg density from an egg
transmission image so as to comprehensively judge the freshness characteristic of the egg. The
machine vision technology based method specifically comprises the steps: acquiring the
transmission image of an egg under a certain illumination condition, transmitting the
transmission image to a computer through
image acquisition equipment, acquiring the information on the three aspects including the
color space, an egg
yolk area and the egg density from the acquired egg image, fusing and screening the information on the three aspects, and finally obtaining a detection result of the egg freshness by means of a multi-variable analysis model in a standard
library, wherein the information on the three aspects including the
color space, the egg
yolk area and the egg density is in one-to-one correspondence to three phenomena of diluted concentrated
egg white, enlarged egg
yolk area and raised gas
cell height which are expressed under an egg freshness drop condition. In addition, grading can be carried out according to standards so that the egg freshness can be detected more objectively and accurately.