Optical fiber spectrum classification detection device and method for fertilized eggs in duck eggs before hatching

Abstract

The invention discloses an optical fiber spectrum grading detection device and method for fertilized eggs in pre-hatch duck eggs. Duck eggs; there is a condenser lens under the perforated conveyor belt, and the condenser lens is connected to a halogen light source through a quartz optical fiber; a collimator lens and a photoelectric sensor are arranged in the optical chamber, and the collimator lens is set correspondingly to the condenser lens above and below, and the collimator lens A fiber optic spectrometer is connected through an optical fiber, and the fiber optic spectrometer is connected to a computer; the photoelectric sensor is connected to a PLC controller, and the PLC controller is connected to the computer; one end of the perforated conveyor belt is connected to a motor, and the other end of the perforated conveyor belt is connected to an egg picking device. . The invention can realize rapid and accurate non-destructive detection on fertilized eggs in breeding duck eggs before hatching.

Description

technical field [0001] The invention relates to the technical field of on-line grading detection of agricultural products and hatching of livestock and poultry breeding, in particular to an optical fiber spectrum grading detection device and method for fertilized eggs in duck eggs before hatching. Background technique [0002] According to the data provided by the China Animal Husbandry Association and the World Food and Agriculture Organization (FAO) in 2018, the world's meat ducks are about 4.74 billion, and the laying ducks are about 187 million. my country's duck breeding accounts for 74.3% of the world's duck breeding. This shows that my country's duck egg hatching industry needs to hatch billions of ducklings every year to meet domestic breeding needs. However, at present, domestic duck breeding methods are based on free-range breeding, resulting in a natural fertilization rate of duck eggs lower than 80%. At present, in my country's poultry egg hatching industry, the...

AI Technical Summary

Problems solved by technology

Zhu Zhihui et al. used hyperspectral technology to collect spectral and image information of eggs before hatching, and established a discriminative model for fertilized eggs and nonfertilized eggs by using correlation vector machines. The identification accuracy rate can reach 96%, but hyperspectral equipment is expensive and difficult to detect. The efficiency is low, and it is difficult to deploy to actual production; use near-infrared diffuse reflectance spectroscopy (Zhu Zhihui, 2012, Spectroscopy and Spectral Analysis) to collect spectral information of pre-hatch eggs, after relevant preprocessing, use Mahalanobis distance to establish discrimination The model recognition accuracy rate is 91.67%, but the diffuse reflectance spectrum technology also has low efficiency; Qin Wuchang et al. used the visible-near-infrared transmission spectrum technology to collect the spectral information of breeding eggs after 3 days of incubation, and established the method after principal component analysis. Fisher discriminant model, the accuracy rate is as high as 96.58%

However, at present, most researchers use spectral technology to statically detect whether the eggs are fertilized before hatching. The models established by static detection have poor generalization ability and often can only stay in the laboratory stage, and the traditional machine learning method is unstable in modeling , it is difficult to apply it to actual production; in addition, the above studies are all aimed at breeding eggs, and it is more difficult to detect breeding duck eggs compared with chicken eggs. There are many stains on the surface, and the color of the eggshells of other kinds of duck eggs is quite different. There are also green shells and white shells in the same species, which makes scholars seldom study duck eggs. fertilized eggs

Images