Insect medium-fine-grained picture classification method based on attention mechanism

The invention discloses a method for classifying fine-grained pictures in insects based on an attention mechanism, and the method comprises the steps: collecting different forms of fine-grained pictures of different types of insects, and deleting repeated, fuzzy and overexposed low-quality pictures in a manual screening manner, so that the remaining pictures meet the characteristics of large intra-class difference and small inter-class difference. Then insect subject enhancement processing is carried out, and a data set with category labels is established; then, a classification network used for insect classification and based on an attention mechanism is built and trained, and finally, classification detection is performed on to-be-detected fine-grained insect pictures through a trained neural network model, so that categories of insects in the pictures are directly output. 本发明公开了一种基于注意力机制的昆中细粒度图片分类方法，采集不同种类昆虫的不同形态的细粒度图片，经过人工筛选方式删除重复、模糊、过曝的低质量图片，使剩余的图片满足类内差异大、类间差异小的特点，然后进行昆虫主体增强处理，并建立带有类别标签的数据集；接着搭建用于昆虫分类的基于