Neural network-based silkworm fresh cocoon defect structure model and establishment method thereof

The invention discloses a defect structure model of fresh silkworm cocoons based on a neural network and an establishment method thereof, and the method comprises the steps: extracting the outer contour features of the fresh cocoons through a fresh cocoon image, and constructing a structure model for the fresh silkworm cocoons and the defect states of the fresh silkworm cocoons. The model construction method is mainly divided into two parts, the first part is to acquire and process images of silkworm cocoons and defect states thereof, and the second part is to construct a silkworm cocoon contour mathematical model. The objective of the invention is to establish a reasonable and effective theoretical model of the fresh silkworm cocoons and the model metamorphosis condition of the defect state of the fresh silkworm cocoons, and provide a construction framework for a calculation algorithm during the subsequent large-scale intelligent screening of the fresh silkworm cocoons by using machine vision. 本发明公开了一种基于神经网络