Research On Detection And Identification Sorting System Of Vaccine Egg Embryo Based On Machine Vision

As vaccines are the major means to control infectious diseases,the immune effect is determined by the quality of vaccines.Furthermore,the vaccines’ quality is dependent on the production testing means.Because egg-embryos are used as cultivation medium in the process of vaccine production,egg-embryo activity testing is the core of vaccine preparation testing.How to improve the efficiency of egg-embryo activity testing and guarantee the quality of vaccines has become a key problem to be solved.However,it is a great challenge to conduct high-throughput online testing for egg-embryo activity.Combining the industry requirements of vaccine production quality testing in China,this thesis,with egg-embryos’ activity characterisitics as research object,conducts fundamental and applied researches on egg-embryo imaging theories and method,quantitative analysis means of egg-embryo biological characteristics,egg-embryo activity identification model construction,egg-embryo negative pressure sorting system control,and egg-embryo activity online identification and sorting equipments.The research has established a series of methods for egg-embryo activity testing and developed a set of high-efficient identification and sorting equipments,providing a scientific testing means for safe vaccine production.The main contents are as follows:(1)The research has studied imaging features of different egg-embryos and analyzed unstable factors in online imaging dark-box environment and the image acquisition system.The V-component image automatic sorting method is put forward to process the images in diversified ways.Since the background of egg-embryo image and egg-embryo countour have the similar gray level,an edge segmentation method is proposed by which adaptive filter and double peak threshold are used to segment egg-embryo component image.Based on the different developmental characteristics of egg-embryos,the research explores the distribution of egg embryos blood vessels and analyzes the influence of egg posture on the blood vessels and air cell distribution in egg-embryo images.A method of obtaining egg-embryo images has been formulated and an egg-embryo image acquisition system has been developed.(2)The research analyzes the image features of egg-embryos with different biological activity.To solve the problem of the similar gray level between egg-embryo countour and blood vessels,the thesis proposes a segmentation method based on the minimum intra-class exponential variance threshold method and multi-level image logical operations to obtain the blood vessel binary form.Through analyzing blood vessel features of egg-embryo component images,a method of segmenting egg-embryo G-component image to obtain egg-embryo contour is proposed.Based on the features of the egg-embryo contour and blood vessel binary image,the method in which multi-level image logic operations remove egg-embryo contour to obtain blood vessel texture binary form is put forward.The method can obtain the evidence of identifying egg-embryo activity and analyze the distribution of air chamber and blood vessel in egg-embryos.(3)By investigating texture features of dead embryos with discontinous blood vessels,the pattern of different proportions between embryo body and air chamber affected by the position of the egg-embryo is identified.The method of identifying blood vessel texture and the proportion of embryo body is proposed,providing a means to tell the discontinuity of blood vessels.The blood vessel texture features of lumpy dead embryos are analyzed to study the relation between blood vessel texture of lumpy dead embryos and activity of egg-embryos.The method of identifying geometric features in blood vessel maximum connection areas is proposed.Through the analysis of sample testing data,the feature parameters of dead embryos in Bayesian classifier are optimized and the activity judgment model based on egg-embryos’ blood vessel features is constructed.(4)The research studies the changing pattern of airflow field of the vacuum cup holding egg-embryos,proposes the critical negative pressure method for holding egg-embryos and predicates the vacuum conditions for reliable holding egg-embryos.The features of holding egg-embryos are analyzed to examine axial force and radial force of holding egg-embryos based on the requiements of moving suction devices.Pleated suction cup and connecting rod absorbers are designed to reduce the speed of critical airflow.The research constructs equivalent analytical model of subatmospheric pressure for holding egg-embryos and analyzes the association characteristics of air-pipe length,vacuum generator’s performance,geometric parameters of suction cups and subatmospheric pressure system.The thesis proposes one-point subatmospheric pressure independent control method by supplying pressure before coupling,which can realize negative pressure balance in pressure system.(5)The research investigates the anti-pressure ability of egg-embryos in different postures.Considering the anti-pressure ability of egg-embryos and spatial constraint of egg vehicles,a loading and unloading system composed of lifting mechanism and scissor-type combination mechanism is devised.In the scissor-type combination mechanism,the rigid transmission mechanism is used to grasp egg-trays and the compliant transmission mechanism is used to transmit egg-trays.Based on the previous study on egg-embryos’ blood vessel texture extraction,activity identification,automatic laoding and unloading,and sorting methods,a set of equipments for high-efficiently testing,identifying and sorting vaccine egg-embryos are developed,which are composed of egg-embryo loading and unloading mechanical arms,activity identification system and sorting devices.