Irregular Meat Thickness Compensation And Foreign Body Detection Method Based On X-ray Imaging System

The meat thickness of the compensation on-line real-time detection system was first designed and constructed, which collecting X-ray image of meat and laser image of meat which filtering, enhancement and other image preprocessing.After analyzing the cause of X-ray image distortion and establishing distortion of X-ray image mathematical model, a global correction based on RBF neural network and nonlinear local interpolation correction method will be corrected effective, and X-ray image geometric distortion correction accuracy.A mathematical model of thickness measurement of laser meat was established. In order to overcome the laser triangulation image of meat collected partial occlusion, the laser double laser triangulation method for meat image acquisition was established, and design roundpoint calibration plane by two-step camera calibration method is about a unified field of view to the left under the camera field of view, and grayscale image mosaic based on two laser image synthesis method.The first rough registration by SURF pixel calibration block shape the rules of X-ray and laser images, then three-dimensional ABC plastic's Mickey Mouse was used to accurate register base on mutual information, to realize X-ray meat image and laser meet image was registered accuracy, similarity measure-based test statistics prove that the meat had a higher image registration accuracy.This paper analyzes the meat thickness, density and X-ray attenuation through the strength of the relationship, established under the X-ray detection model irregular shapes of meat and meat based on laser triangulation thickness testing the mathematical model, established to compensate for X-ray images uneven thickness of the thickness of the meat compensation image mathematical model, based on the weight function given in detail the meat thickness compensation least squares optimal solution algorithm, and the establishment of compensation for laser imaging and thickness mapping image.The X-ray image and laser thickness of the meat compensation image fusion, the image of laser thickness compensation to eliminate the uneven thickness of the meat X-ray image, the fusion image segment algorithm using OTSU foreign image, experimental results show that OTSU segmentation algorithm can complete segment meat foreign body images, the accuracy of detection for lmm precision foreign bodies.