Finished tea type and grade identification method based on image color and texture features

Abstract

The invention belongs to the technical field of digital image processing, and relates to a finished tea type and grade recognition method based on image color and texture features, and the method comprises the steps: collecting a finished tea image in a self-designed photographing device through employing an industrial camera and a lens; after the image is preprocessed, extracting color features of the image by using a quantized color histogram method, and extracting the texture features of the image by using a gray level co-occurrence matrix method after parameter optimization; and combiningthe two types of features, and importing the two types of features into a support vector machine classifier optimized based on a grid search method for classification. According to the method, the types of the finished tea can be accurately distinguished, the method has excellent performance in classification of the same type of finished tea with different grades, and a new technical route is provided for efficient and automatic classification of the types and grades of the tea.

Description

technical field [0001] The invention belongs to the technical direction of digital image processing, and relates to a method for identifying the type and grade of finished tea based on image color and texture features. Background technique [0002] China is the largest country of origin, production and consumption of tea. There are many kinds of tea in different shapes. At the same time, a kind of tea often has different grades and qualities. The quality requirements of different grades of tea are very different, and the price varies greatly. Merchants take advantage of loopholes and cheat consumers with shoddy goods. Therefore, if the types and grades of tea can be identified accurately and efficiently, it will be of great significance to the production, sales and market regulation of tea. [0003] Traditionally, the classification of finished tea types and grades relies on human experience. The closer the tea leaves are, the more experienced experts are needed, and the ac...

AI Technical Summary

Problems solved by technology

[0004] At present, some scholars use digital image processing, computer vision and other technologies to classify finished tea, and import image features into classifiers for classification. The accuracy rate is high or low, and the stability is poor.

However, current research often stays at the level of application algorithms, and does not optimize parameters for specific needs. There is still room for improvement in classification accuracy.

In particular, the current research on the classification of different grades of the same kind of tea is still blank. If the classification of the same kind of tea grades can be realized, it can provide a new application scenario for the automation of tea classification

Images