Research On Silkworm Cocoon Gender Recognition System Based On Machine Vision

Sericulture industry as the traditional advantage of Chinese,it occupies an important proportion in the agricultural economy.In the main producing areas of silkworm breeding, it has important significance to the employment, the export of related products and economic construction. etc. In silkworm breeding and breeding process, the classification of male and female silkworm chrysalis have direct effect to silkworm breeding and silk products processing, it related to the quality of silkworm eggs, cocoon and silk quality, so classification of the male and female silkworm chrysalis is producing excellent critical steps.In view of the current classification of male and female silkworm are mostly artificial, low efficiency, high intensity, influenced by subjective factors, based on the present working situation, and the differences in body shape and texture structure which in the front of the silkworm moth emergence period, with the method of image processing to research the identification of silkworm chrysalis gender. In the course of the study, through the image acquisition, image pre-processing, feature extraction, and the use of Mahalanobis distance analysis, establish the PCA-BP model training and simulation experiment based on principal component analysis and artificial neural network, recognition accuracy rate is up to 98%, conformed to the standard of ?The quality standards and inspection procedures for silkworm generation hybrids?.This study provide theoretical support and a new way for silkworm classification to it and has important application value for the actual production. Generally speaking, the main research contents involved are as follows:1.Describes the background and significance of silkworm classification, discusses the domestic and foreign related research progress and trends.2.To explore the feasibility of machine vision recognition approaches, and introduces its application scope and trend..3.The silkworm chrysalis image acquisition device design and the selection of the relevant components, build the platform of machine vision, do preparatory work for image acquisition.4. After the image acquisition, do a series of image pre-precessing work, such as the gray scale, shape, quality etc. and according to the shape of the male and female gender, different texture, extract the shape and texture feature parameters.5.On the basis of image processing,and based on the purpose of improve the automation,analyze the extracted feature data,and applied the distance discriminant analysis and get its relevant principal component;To explore and improve the network structure, establish the PCA-BP network structure and program suit for this research object. Through the use of silkworm chrysalis sample training, simulation and validation of the classifier. The results show that the method of silkworm chrysalis gender discrimination accuracy rate reached 98%, conformed to the industry standard.