Machine vision-based real-time detection and grading method and machine vision-based real-time detection and grading device for pork appearance quality

Abstract

The invention discloses a machine vision-based real-time detection and grading method and a machine vision-based real-time detection and grading device for pork appearance quality. The device comprises a hardware part for pork image acquisition and a software part for grading, wherein the hardware part for image acquisition consists of a camera, a lens, a polarizer, a light source, a computer, and the like. The method comprises the following steps: acquiring a pork image; carrying out real-time image processing and characteristic extraction through grading software; acquiring characteristic information reflecting the pork appearance quality; and evaluating the pork quality through a forecasting model, completing the grade estimation of color, marble vein, the tenderness and the comprehensive quality of underback, and completing the grade estimation of muscle color, the fat and the comprehensive quality of marbled meat. The method and the device can make the pork quality detection have objectivity, accuracy and high efficiency, and are suitable for the pork quality evaluation carried out by supermarkets, local food inspection authorities and import and export food inspection departments so as to standardize the pork market and consequently truly realize the pricing by the quality.

Description

technical field [0001] The invention relates to a machine vision-based real-time detection and grading method and device, in particular to a machine vision-based real-time detection and grading method and device for pork appearance quality. Background technique [0002] Meat is one of the common foods for people. According to statistics, the total output of pork in my country showed a gradual upward trend from 2000 to 2006, from 40.31 million tons in 2000 to 51.97 million tons in 2006, with an average annual growth rate of 5.8%, accounting for 10% of livestock meat. More than 60% of the total output. It can be seen from this that pork has a very large share in the domestic market, and people's demand for pork is increasing. [0003] Although my country is a big meat producer, it is not a big meat industry. In recent years, the export volume of meat has been decreasing year by year. The main reason for this phenomenon is that the meat products in our country are seriously mi...

AI Technical Summary

Problems solved by technology

The image processing algorithms and detection devices introduced in the above-mentioned documents and patents are all aimed at the quality evaluation of beef. However, due to the large differences in color and texture between pork and beef, it is mainly manifested in the fact that the fresh pork muscle is light red and different. The color of meat from different parts is quite different, and the texture of pork is finer than that of beef. At the same time, pork is prone to water seepage when placed in the open air, which makes it more difficult to evaluate the quality of pork than beef. Therefore, early scholars proposed various images for beef detection and grading. Processing algorithms and setups no longer suitable for pork research

Images