Research On Automatic Identification System Of Male And Female Silkworm Chrysalis Based On Computer Vision

As the birthplace of sericulture industry in the world,China has always occupied an important position in the sericulture industry.In the process of silkworm breeding,the accuracy of the classification of male and female identification of silkworm pupa will directly affect the yield and quality of silkworm cocoon and cocoon silk.Aiming at the problem of high cost and low accuracy of artificial identification in a sericulture enterprise in Jiangsu Province,this paper studies the automatic identification technique of male and female in silkworm pupa based on computer vision.The specific work is as follows:(1)Image preprocessing of silkworm pupa.In order to solve the problem of unclear feature information and noise pollution of silkworm pupa image,the original image was enhanced,including image filtering denoising and image grayscale enhancement.In order to realize the accurate segmentation of the target and background in the image,the method of combining threshold segmentation with morphological processing technology is adopted.(2)Feature extraction and analysis of silkworm pupa image.In order to analyze the differences between male pupa and female pupa in physical characteristics,three characteristic parameters of area,circumference and dispersion were extracted and analyzed.Aiming at the difference of texture between the front and back of silkworm pupa,the method of Gray Level Co-occurrence Matrix is used to extract and analyze the texture feature of the image.Aiming at the difference of female pupa and male pupa in the texture of the tail,the improved LBP operator was used to extract the texture feature information.(3)Identification and classification of male and female silkworm pupa based on PCA and SVM.Because the dimension of the obtained feature data is too high,the method of PCA is used to reduce the dimension of the characteristic data of silkworm pupa.Considering the excellent effect of support vector machine SVM on high dimensional and nonlinear problems,the method of support vector machine is used to design the corresponding classifier.(4)Study on the automatic identification and sorting control system of male and female silkworm pupae.Based on the investigation and literature review in the early stage of the project,the overall design scheme of the automatic identification and sorting control system of male and female silkworm pupae was proposed at first.Then,a more specific implementation scheme was proposed for the image collection and recognition and HMI monitoring system.