A kind of tea green classification method

Abstract

The invention discloses a tea-green classification method combining ghost attention capsule network and knowledge distillation, including tea-green image acquisition; data enhancement and data set establishment; construction of ghost attention bottleneck layer and tea-green classification model; Training and weight acquisition; training the tea-green classification model by means of growing knowledge distillation; performance verification of the tea-green classification model. The method of the invention reduces the demand for the number of training samples, and enhances the model's ability to process small-scale data sets; the model can adaptively adjust the weight coefficients to emphasize important feature channels while weakening irrelevant feature channels; while transmitting the huge parameters of the teacher model The matrix is ​​used to enhance the comprehensive performance of the student model, and at the same time achieve the purpose of adaptively reducing the dependence of the student model as the distillation progresses.

Description

technical field [0001] The invention relates to the technical field of tea green classification, in particular to a tea green classification method combining ghost attention capsule network and knowledge distillation. Background technique [0002] When picking tea leaves, different grades of tea greens will be mixed together. The traditional method of sorting only based on the experience of tea makers will cause the classification results to be greatly affected by human subjective factors, which will seriously restrict the production of tea production enterprises. Efficiency, increased time and production costs, are not conducive to large-scale batch customized production and the transformation to the famous tea industry. [0003] Although deep learning technology has achieved excellent performance in classification problems in many fields, the research on tea classification based on this technology is still in its infancy due to the limitations of regional factors in tea cu...

AI Technical Summary

Problems solved by technology

[0005] (1) The traditional model based on the convolutional layer will require a large number of samples in order to train the weight matrix. In practical applications, the collection of brown and green images faces many challenges, and the lack of training samples will restrict the classification performance of the model.

[0006] (2) The existing model cannot measure the importance and degree of correlation of different channels in the matrix features of dark brown images, and adaptively adjust the weight parameters of each channel to achieve the purpose of strengthening effective features and suppressing irrelevant features

[0007] (3) The existing technology mainly uses a single model for training, and the collaboration between multiple models has not yet been realized, and it is difficult to further enhance the performance of the dark green classification model with the help of rich weights of complex models

Images