Study On Image Thickness-Compensation Algorithm Of Bone Fragment Detection In The Meat Or The Meat Products

Small bones embedded in the meat or the meat products harm the human health. In this paper, chicken breasts are taked as sample and small bones are detected through X-ray imaging system.This article makes research on thickness-compensation algorithm in order to remove deviations caused by meat because of different shape and uneven thickness.The result have good effect on the extraction and identification of small bones in meat.The mathematical model are established in different X-ray energy level based on X-ray images characteristics of meat products, which includes the image contrast and signal to noise ratio changes. A reasonable X-ray energy is selected in order to improve image resolution through the above test.Laser scanning methods are used on thickness measurement of the meat and meat products and a model of laser thickness measurement device is builded up.Then leaser images and X-ray images registration are resolved through affine transformation, which can eliminate the impact of different resolution, translation and rotation on the position and scale to the image. The dates of laser images are fitted based on the least squares method and related thickness-compensation formulas are seted up. The mapping of laser images to compensation images can be realized by Matlab, then the X-ray images and compensation images can be fused. OTSU algorithm as the threshold segmentation method are used to fusion images. First of all, the image is divided into ten sub-images and the optimal threshold are selected as the overall internal threshold in image segmentation. The outcome of images can achieve good identification of bones in meat.A man-machine interface is designed based on the algorithm in this paper. Bones in different shape are tested and tests show that thickness-compensation algorithm can effectively eliminate the impact of meat shape and thickness, also have a high recognition rate.