Online Image Detection System Of The Defect Of Hatching Egg Embryo

This paper researches mainly the research and development of online image detection system of the defect of hatching egg embryo. The system uses the industrial high-speed CCD camera to photo the defective hatching egg embryo within the image, obtains images and processes these images through the DSP board, then combines with the sorting mechanism, and has achieved the detection of automatic online images of the hatching egg embryos (live embryo, weak embryo, dead embryo, pollution embryo).In this paper, the extraction of feature information in the defective image of hatching egg embryo is completed through using the related knowledge of image preprocessing, threshold segment, RGB and HSI color model, and so on. By using the pattern recognition knowledge, the information of the egg embryo blood vessel and pollution block region pixel's R, G, B, H, S, I and gradation is taken as the characteristic parameter which in the pattern recognition inputs, the appropriate sorter is designed to achieve to the defect determination, the distinction of the defect of hatching egg embryo has been realized, and the detection accuracy is enhanced correspondingly.Hardware design of the system mainly is divided into platform design and sorting mechanism design. The platform design is based primarily on the specifications of the equipment used and number, as well as to complete the function. The sorting mechanism design is based primarily on the theory of the vacuum adsorption circuit. The vacuum cup is as the implementing agency, the commutation solenoid valve is controlled the detection signal which sends by the DSP, and the moving of work table in the entire pneumatic circuit is achieved by the air cylinder.Image processing of the system, using the VC++ pre-verifies detection algorithm to shorten the development time. Algorithms are realized and each function modulation of the image processing is achieved by Using TI's multimedia processing chip TMS320DM642 as the video processing core processor. Ultimately, the algorithms are constantly debugged in the engineering experiment to achieve optimum detection algorithm.The experimental results show that the system has changed the traditional manual mode of operation, avoided the test result which is different from person to person, reduced the detection classification error, improved the detection speed and accuracy.