23-06-2023 дата публикации
Номер: CN116310391A
Принадлежит:
The invention relates to a method for identifying tea diseases, which comprises the following steps: step 1, data preprocessing: collecting tea disease image data, cutting each picture containing a plurality of diseases and background noise, and cutting the shot picture into a single disease leaf as an image; step 2, constructing an optimal small sample graph network model and training the model: firstly, embedding the tea disease image into feature vectors, then taking each feature vector as a double-domain node initialization graph of the tea disease image, and performing graph updating optimization according to the constructed double-domain node initialization graph, the optimal small sample graph network model comprises three parts of bottom-to-top reasoning, top-to-bottom reasoning and jump connection; and step 3, carrying out tea disease image identification on the image. According to the invention, manpower and material resources consumed by artificial disease identification are ...
Подробнее