Detection and classification on surface smudges of duck eggs based on machine vision technology

With the fast development of modern society, it is necessary to detect the surface smudges of duck eggs and then classify them so as to satisfy people’s pursuit of high quality life. This article has realized detecting the surface smudges and classifying them through analysing the surface characters, such as feather, smudges, blood spots and so on. The methods in this research are dividing the group dynamic images into singles using MATLAB and pretreating them firstly, and then getting character parameters of colour and smudge area proportion, finding the parameter range of classifying clean eggs and dirty eggs, building mathematical model, and reaching the aim of detecting surface smudges and classifying them automatically. Also, the accuracy of classifying duck eggs reaches 90%. This article has laid an academic foundation for detecting eggs’ smudges and classifying them.