Meat online non-destructive testing method and apparatus based on fusion of image and spectrum information

Abstract

The invention discloses a meat online non-destructive testing method and apparatus based on fusion of image and spectrum information, including a conveyor belt, a spectrum collection room, an image collection room, a camera mounted beam, an imaging source, a camera, a Y type optical fiber, a spectrograph, a data processing PC, two photosensors, a motor and a conveyor belt bracket. Through online collection of meat image and spectrum information, collecting feature information of characteristic meat quality, establishing a fused quantification prediction model based on image characteristics, spectrum characteristics and characteristics layer data of image and spectrum, using the obtained quantification prediction model to combine with quality evaluation system to establish a decision layer data fusion model based on image and spectrum for testing meat online. The invention uses complementarity and redundancy of the image information and the spectrum information to implement classification of materials and quality test during meat producing process according to optical feature of meat, which improving production efficiency.

Description

technical field [0001] The invention relates to an online detection method and device for agricultural and livestock products, in particular to an online non-destructive detection method and device for meat quality based on fusion of image and spectral information. Background technique [0002] Meat and meat products are one of the main food sources for human beings. With the improvement of consumption level, the quality of meat products has attracted widespread attention from consumers; in order to ensure the quality of products and gain a strong competitiveness in the fierce market competition, meat processing enterprises have carried out research on raw meat and processed semi-finished products Quality monitoring and control, pricing products according to their quality, in order to maximize the economic value of products. [0003] The traditional meat quality testing uses artificial sensory evaluation methods and chemical analysis methods. Artificial sensory evaluation ...

AI Technical Summary

Problems solved by technology

[0005] This kind of non-destructive testing technology based on the optical characteristics of agricultural and livestock products is fast, accurate, and free from subjective human influence. Researchers at home and abroad have carried out a lot of research, but mainly use image information or spectral information to detect meat products with a single index. To carry out grading, it is not possible to test multiple quality indicators inside and outside the meat products

Images