Intelligent grading method for pork in pigs based on deep learning

Abstract

The present application belongs to the technical field of grading detection of pig cut meat, and discloses a kind of intelligent grading method of pig middle side meat based on deep learning. First, a ruler is designed; the improved UNet model is used to segment the middle side meat image, the triangular meat and pigskin area are extracted, the intersection point of the triangular meat tip point and its normal direction and the upper edge of pigskin is located, the pixel distance between the two points is calculated and converted to actual distance through the ruler scale, and the fat thickness measurement value is obtained; according to the threshold value set by the standard and enterprise measured data of NY / T 3380-2018 pork grading, the fat thickness grading standard is established; at the same time, 24 color block detection and color correction are carried out on the image, the muscle area is segmented and converted to Lab color space to extract the chroma value, the chroma threshold value is set according to the standard of NY / T 821-2019 pork quality grading, and the meat color grade is divided; finally, the fat thickness grade and muscle chroma grade are weighted and fused according to the weight coefficient, a double-index comprehensive grading model is constructed, and the intelligent and objective grading of middle side meat is realized.

Images