Key Technologies For Aluminum Foil Packaging Duck Gizzard Foreign Body Detection

Food foreign bodies harm the human health and hold back the food industry development. To detect food by computer vision and image processing technology, the examination has the advantages of non-destructive, fast, accurate, and reliable etc. Taking aluminum foil packing duck liver as object, this article makes research on the radiographic image of duck liver image by X-ray in order to achieve the recognition and extraction from duck liver of foreign bodies based on the computer vision and the image processing technology.This article has studied X-ray digital image radiographic optimization and emendation method,And the character of the radiographic image of duck liver is analyzed .X-ray coefficient of attenuation picture element value mathematical model is developed to estimate blank image intensity,duck liver and foreign bodies at different voltages and currents . The method of value is higher than 0.995;In order to guarantee the good effect of the foreign bodies Segment, and to enhance the foreign bodies inspection speed, a method of setting X-ray parameter (tube voltage, tube electric current) is proposed in this paper.The characteristics of the noise from Image Intensifier and the CCD camera of the X-ray radiographic imaging system are analyzed. In order to improve the image acquisition quality and reduce the noise disturbing, multi-frame means filter processing, reducing in the image the stochastic noise, is proposed. In order to ensure the foreign bodies of duck liver image specific clearly, 5×5 Gauss filter method is employed in the duck liver image for denoising.This thesis has studied in the foreign bodies of X-ray radiographic duck liver image automatic judgment method to separate the duck liver from the background by using OTSU method;Under the limitation to the size scope and the great gradient changes to foreign bodies of duck liver image, the foreign bodies are successfully separated from duck liver by Active Contour Model method. This approach is of such high inspection precision that it can accurately ascertain foreign bodies such as stone, bone and metal wire above 0.2mm in diameter.On the basis of researches, through the shape and image grayscale of duck liver's foreign bodies'characteristics, an eigenvector that factually reflected foreign bodies'essential properties and their computation is presented. On solution of foreign bodies'classification, the self-organized and self-adaptive back propagation (BP) neural network algorithm is used to intelligently recognize the image of duck liver's foreign bodies. The above-mentioned solution to foreign bodies'recognition of duck liver in X-ray radiographic image passed simulation and emulation with Limited sample simulation, and the result has obtained a perfect effect.The X-ray automatic inspection software system for duck liver's foreign bodies is developed under the Visual c++.net platform. This software system can elementarily achieve the recognition of the image of duck liver in X-ray, and fulfill the basic function of the image processing and analyzing as well.