Research On The Intelligent Testing System Of Egg Embryo Based On Multi-information Fusion

Egg embryo as the culture medium of biological vaccine production, it has two disadvantages such as long time cultivation and low yield. Therefore, it is very important for the pharmaceutical industry to detect the egg embryo promptly and accurately. However, the current detection method mainly using the relationship between the egg embryo’s one characteristic quantity and the egg embryo’s type to establish related model, and then detect it. As a kind of organism, egg embryo itself structural complexity makes the models not the same as what we expect. Accordingly, this detection method has neither accuracy rate nor strong stability. In order to overcome these shortcomings and improve the accuracy of the egg embryo detection, we need not only to make full use of the detection information, but also to investigate the mapping relationship between detection information and egg embryo types.This will be systematic and comprehensive investigation to the process of feature extraction, modeling and egg embryo’s image, temperature, and light intensity information. This paper proposed a method of information fusion based on Neural Networks. This will improve the accuracy of detection for egg embryo if it applied to actual production.This article researches on the intelligent testing system of egg embryo based on multi-information fusion for the detection of egg embryo. The system uses the mixed programming combined Visual C++and MATLAB, designs and constructs multi-sensor data acquisition platform, selects the appropriate light source and sensor. Then, the system acquires parameters that related with egg embryo characteristic such as image, temperature, and light intensity information. Making use of the strong nonlinear mapping function of neural networks, we would also build a BP neural network, which is expected to help build up the mapping relationship between the types of egg embryo and the feature signals. Then make the distinguish of the type of egg embryo based on the output of the fusion center.The experimental verification process of the pattern recognition system of egg embryo has been done in combination with the hardware platform, the results of test showed that:the system can capture images, temperature and light intensity signal of egg embryo accurately, using BP neural network technology to extract characteristic parameters of each signal and its fusion, the result can be used to detect the type of egg embryo.