关键词: 南疆红枣, 纹理特征, 自动分级, 特征参数

Abstract: Red jujube grading technology was mainly relied on artificial classification， which caused the low production efficiency and classification precision. With the development of nondestructive testing technology， computer technique was used to realize automatic classification of red jujube， which has become the best realization method. This paper took the external quality characteristics of red jujube as the research object， to explore the characteristic parameters acquisition and classification method for the external quality of red jujube， using the digital image processing technology based on MATLAB. Images were preprocessed and the image feature was extracted by using gray scale. The experimental results showed that this method can realize the collection of red jujube external characteristic parameters， thus realize the texture classification for Xinjiang red jujube， which has practical application value and guiding significance for the future appearance quality comprehensive classification of the red jujube based on machine vision.

Key words: southern Xinjiang red jujube, texture feature, automatic classification, characteristic parameters