Tea verification method based on exclusive twin network model

Abstract

The invention discloses a tea verification method based on an exclusive twin network model. Each type of tea is provided with one exclusive twin network model to automatically extract unique implicitcharacteristics of target tea for verification. Wherein the input of the model is the combination of the target tea and any tea sample, the feature vectors of the target tea and any tea sample are extracted by taking the VGG16 as a reference structure, finally, the L1 distance of the dimension reduction vectors of the two channels is subjected to logistic regression, if the types of the tea are consistent, the logistic regression value is 1, otherwise, the logistic regression value is 0. Moreover, in the test stage, the tea leaf picture to be verified is converted into an HSV space, and imagecolor calibration is carried out on the tea leaf picture to be verified by using a histogram specification method, so that the brightness and hue of the tea leaf picture to be verified are close to those of a sampled picture during training, and the problem of low verification accuracy caused by equipment difference is solved. Compared with the mode of directly using a full classification model, the method is more efficient and reliable, and can accurately judge or verify whether the tea purchased by the user is true or not.

Description

technical field [0001] The invention relates to a tea authenticity verification method based on a twin network model, which belongs to the fields of image, artificial intelligence and neural network. Background technique [0002] Tea culture is an important part of Chinese traditional culture. Nowadays, it is difficult to distinguish the authenticity of tea leaves on the market. To help consumers verify the authenticity of the tea they buy, such as Figure 7 As shown, it is particularly important to develop a reasonable and efficient tea authenticity method. This problem is similar to the verification of human identity, but tea does not have a fixed benchmark image like a human face, and there is a certain degree of blindness in comparison using traditional feature extraction methods, often a certain feature is only applicable to the verification of a specific tea variety , and cannot be extended to all tea varieties. In addition, the number of tea categories is also uncert...

AI Technical Summary

Problems solved by technology

[0003] The technical problem to be solved by the present invention is: In view of the blindness in the extraction and selection of features in the traditional verification method, and the uncertain number of tea categories, the present invention proposes to equip each type of tea with an exclusive twin network model authenticity method

At the same time, in order to eliminate the problem of authenticity errors caused by differences in sampling equipment, the present invention proposes to prescribe the HSV histogram of the picture before testing, so that it is close to the sampling equipment during training in terms of brightness and color balance.

Images