Egg sex nondestructive identification method based on low-field nuclear magnetic resonance

Abstract

The invention discloses an egg sex nondestructive identification method based on low-field nuclear magnetic resonance, and belongs to the technical field of egg detection, and the method comprises the following steps: S1, adopting low-field nuclear magnetic resonance equipment to carry out echo signal acquisition on eggs in a hatching period; s2, performing spectrogram inversion on the acquired echo signals by utilizing a non-negative least square optimization model and combining with a Tikhonov regular term, and reconstructing a T2 relaxation time spectrum of moisture in the poultry eggs related to tissue structure distribution; s3, extracting multi-dimensional spectroscopic features associated with the gender information of the poultry eggs from the T2 relaxation time spectrum to form high-dimensional feature vectors representing embryonic tissue microstructure differences; and S4, inputting the high-dimensional feature vector into an integrated classification model, and outputting an egg sex identification result. The method can be integrated on an automatic conveying line of an incubation factory, multi-egg-position parallel detection, rapid non-contact recognition and classified output are achieved, and high adaptability and engineering practicability are achieved.

Description

Technical Field

[0001] The present invention belongs to the technical field of poultry egg detection, and in particular relates to a non-destructive method for identifying the sex of poultry eggs based on low-field nuclear magnetic resonance. Background Art

[0002] In the duck industry, the value of male and female individuals differs significantly. This is especially true in breeding and laying ducks. Accurately identifying the sex of embryos in the early stages of incubation would significantly improve resource allocation efficiency, reduce breeding costs, and increase hatching-to-output ratios. However, traditional sex determination methods, such as DNA testing, hormone labeling, and visual observation, suffer from issues such as shell breakage, low efficiency, or a lack of industrial scalability. In recent years, researchers have explored various advanced detection technologies to achieve non-destructive sex determination of poultry eggs. These include near-infrared spectroscopy and laser scattering to analyze light signals transmitted through the eggshell; ultrasonic imaging and reflectometry to assess embryonic tissue structure or sound velocity differences; and multimodal sensing technologies such as hyperspectral imaging and microwave radioscopy to capture internal optical or electromagnetic signatures through the eggshell. While these methods have achieved some progress, they still suffer from limited sensitivity, high noise levels, significant impacts of eggshell thickness and position, and bulky equipment, making them difficult to implement reliably and on a large scale in industrial incubation environments. Summary of the Invention

[0003] In response to the above-mentioned deficiencies in the prior art, the present invention provides a method for non-destructive sex identification of poultry eggs based on low-field nuclear magnetic resonance, which solves the problems of low sensitivity, great influence of shell thickness, and difficulty in industrialization of existing non-destructive testing methods such as optical and ultrasonic methods, and provides a method for non-destructive sex identification of poultry eggs that can achieve high accuracy, non-contact, and assembly-line-level processing in the early stages of incubation.

[0004] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is: a method for non-destructive identification of the sex of poultry eggs based on low-field nuclear magnetic resonance, comprising the following steps:

[0005] S1. Using low-field nuclear magnetic resonance equipment to collect echo signals from eggs in the chick stage;

[0006] S2. Using a non-negative least squares optimization model combined with the Tikhonov regularization term to perform spectral inversion on the collected echo signals, reconstructing the T2 relaxation time spectrum related to the distribution of water and tissue structure in the egg;

[0007] S3, extracting multidimensional spectral features associated with egg sex information from the T2 relaxation time spectrum to form a high-dimensional feature vector representing the microstructural differences of embryonic tissue;

[0008] S4. Input the high-dimensional feature vector into the integrated classification model and output the egg sex identification result.

[0009] Furthermore, in step S1, the signal-to-noise ratio and sampling time of the echo signal are evaluated in real time according to the quality of the sampled echo signal, and the optimal echo interval in the echo signal acquisition process is determined by optimizing the objective function. and echo number ;

[0010] The optimization objective function for:

[0011]

[0012] Where, and Respectively represent the weights for measuring echo signal quality and sampling efficiency, It represents the ratio of the peak value of the echo signal to the standard deviation of the background noise.

[0013] Furthermore, in step S2, the objective function of performing spectrum inversion on the collected echo signal using the non-negative least squares optimization model combined with the Tikhonov regularization term is:

[0014]

[0015] Where, represents the collected echo signal vector, represents the system response kernel, represents the T2 relaxation time spectrum coefficient, Represents the parameter used to control the smoothness of the spectrum, represents the first-order gradient operator.

[0016] Furthermore, in step S3, multidimensional spectral features associated with egg sex information are extracted from the T2 relaxation time spectrum, including main peak position, spectral width, energy density ratio, spectral change rate, spectral skewness, spectral entropy and peak shape sharpness;

[0017] The main peak position is the relaxation time corresponding to the maximum point of the relaxation spectrum in the T2 relaxation time spectrum; the main peak position of the male embryo is shifted to the left as a whole compared to the female embryo, and the main peak position of the female embryo remains in the short relaxation region;

[0018] The spectrum width is the range of the relaxation time distribution in the T2 relaxation time spectrum; the spectrum width of male embryos is greater than that of female embryos;

[0019] The energy density ratio is the ratio of the intensity of the fast relaxation segment to the intensity of the slow relaxation segment; the energy density ratio of the male embryo is greater than that of the female embryo;

[0020] The spectrum change rate is the absolute value of the slope of the T2 relaxation time spectrum in logarithmic coordinates; the spectrum change rate of male embryos is greater than that of female embryos;

[0021] The spectral skewness is the normalized value of the third-order central moment, which describes the left-right asymmetry of the T2 relaxation time spectrum. The spectral skewness of male embryos drifts in the negative direction, while the spectral skewness of female embryos is close to zero or slightly positive.

[0022] The spectral entropy is a measure of the uniformity of the energy distribution of the T2 relaxation time spectrum; the spectral entropy of male embryos is smaller than that of female embryos;

[0023] The peak sharpness describes the sharpness and sharpness of the main peak of the T2 relaxation time spectrum; the peak sharpness of male embryos is smaller than that of female embryos.

[0024] Furthermore, the calculation formula for the main peak position is:

[0025]

[0026] Where, Indicates the spectrum intensity Get the index of the spectrum point with the maximum value, represents the T2 relaxation time corresponding to the maximum spectral intensity, i.e. the main peak position;

[0027] The calculation formula of the spectrum width is:

[0028]

[0029] Where, represents the spectrum width, and represent the maximum and minimum relaxation times in the T2 relaxation time spectrum;

[0030] The calculation formula of the energy density ratio is:

[0031]

[0032] Where, It represents the ratio of the intensity of the fast relaxation segment to the slow relaxation segment. represents the fast / slow relaxation segmentation threshold, represents the energy of lipid-bound water in the fast relaxation phase, represents the free water energy in the slow relaxation phase, and Respectively represent the T2 relaxation time values ​​corresponding to the i-th and j-th spectral points in the spectrum, and Represent the spectral energy of the i-th and j-th spectral points respectively;

[0033] The calculation formula of the spectrum change rate is:

[0034]

[0035] Where, represents the spectrum change rate of the i-th spectrum point in the T2 relaxation time spectrum, and represent the spectral energy of the i-th and i+1-th spectral points respectively, and Respectively represent the T2 relaxation time values ​​corresponding to the i-th and i+1-th spectral points in the spectrum;

[0036] The calculation formula of the spectral skewness is:

[0037]

[0038] Where, represents the spectral skewness, represents the total number of spectral points, and represent the mean and standard deviation of spectral energy, respectively. represents the spectral energy of the i-th spectral point;

[0039] The calculation formula of the spectral entropy is:

[0040]

[0041]

[0042] Where, represents the spectral entropy, represents the normalized energy, represents the spectral energy of the i-th spectral point, represents the total energy of all spectral points;

[0043] The calculation formula of the peak sharpness is:

[0044]

[0045] Where, Indicates the peak sharpness, n indicates the total number of spectral points, represents the standard deviation of the spectral energy, represents the average spectral energy, represents the spectral energy of the i-th spectral point.

[0046] Furthermore, the step S4 includes the following sub-steps:

[0047] S41, input the high-dimensional feature vector representing the microstructural differences of embryonic tissue into the integrated classification model, and output the corresponding probability distribution P m (y|x); where y∈{male, female, uncertain} represents the gender label and x represents the high-dimensional feature vector;

[0048] The integrated classification model includes a support vector machine model, a random forest model and a shallow neural network;

[0049] S42. Use a weighted voting fusion strategy to aggregate the probability distributions output by the support vector machine model, random forest model, and shallow neural network to obtain the final comprehensive predicted gender label. ;

[0050] S43. Comprehensive prediction of gender labels A confidence evaluation is performed, and the comprehensive predicted sex labels that meet the confidence evaluation standards are output as the sex identification results of poultry eggs, which serve as a reference for the control of automated sex sorting of poultry eggs.

[0051] Furthermore, the step S3 further includes:

[0052] A disturbance compensation mechanism based on multi-sensor environmental detection is introduced to compensate for the drift of the extracted multi-dimensional spectral features.

[0053] The formula for compensation calculation is:

[0054]

[0055] Where, represents the corrected stable multidimensional spectral characteristics, represents the characteristics before correction, 、 and It represents the empirically calibrated temperature interference sensitivity coefficient, humidity interference sensitivity coefficient and electromagnetic disturbance interference sensitivity coefficient. 、 and They represent temperature drift value, humidity drift value and electromagnetic disturbance drift value respectively.

[0056] Compared with existing non-shell-breaking sex detection methods such as spectroscopy and ultrasound, the present invention has the following significant technical advantages and effects:

[0057] (1) Ability to discriminate based on internal tissue components: Most existing methods are based on surface transmitted light, reflected sound, or imaging features. However, the present invention uses low-field nuclear magnetic resonance technology to directly respond to the relaxation characteristics of proton structures such as water, fat, and protein in the embryo, reflecting gender differences from the microstructure and metabolic level, and possessing a "deep physiological mechanism-level identification capability" that is difficult to achieve with other methods.

[0058] (2) Higher recognition accuracy: Using the transverse relaxation spectrum T2 characteristics of nuclear magnetic resonance, it can deeply reflect the water-fat status and tissue differences inside the embryo;

[0059] (3) Stronger stability and environmental adaptability: Introducing an environmental disturbance compensation mechanism to ensure that the system can maintain discrimination accuracy under different incubation scenarios such as temperature and humidity, electromagnetic interference, etc., and avoid the sensitivity of existing spectral methods to eggshell thickness, light source interference, etc.;

[0060] (4) Non-destructive and non-contact detection: The entire detection process does not require breaking the shell or attaching external sensors, which preserves the integrity of the embryo and effectively improves the hatching rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 This is a flow chart of the method for non-destructive egg sex identification based on low-field nuclear magnetic resonance provided by the present invention. DETAILED DESCRIPTION

[0062] The specific embodiments of the present invention are described below to facilitate understanding of the present invention by those skilled in the art. However, it should be clear that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the appended claims, these changes are obvious, and all inventions and creations utilizing the concepts of the present invention are protected.

[0063] An embodiment of the present invention provides a method for non-destructive identification of the sex of poultry eggs based on low-field nuclear magnetic resonance. The relaxation response of the internal tissue of the poultry eggs under a magnetic field is obtained through low-field nuclear magnetic resonance technology, and then a gender-sensitive protein is constructed based on the parameter changes of the transverse relaxation spectrum. Finally, the gender prediction output is achieved with the help of a machine learning model.

[0064] The non-destructive sex identification method of poultry eggs in the embodiment of the present invention is as follows: Figure 1 As shown, the following steps are included:

[0065] S1. Using low-field nuclear magnetic resonance equipment to collect echo signals from eggs in the chick stage;

[0066] S2. Using a non-negative least squares optimization model combined with the Tikhonov regularization term to perform spectral inversion on the collected echo signals, reconstructing the T2 relaxation time spectrum related to the distribution of water and tissue structure in the egg;

[0067] S3, extracting multidimensional spectral features associated with egg sex information from the T2 relaxation time spectrum to form a high-dimensional feature vector representing the microstructural differences of embryonic tissue;

[0068] S4. Input the high-dimensional feature vector into the integrated classification model and output the egg sex identification result.

[0069] In step S1 of the embodiment of the present invention, low-field nuclear magnetic resonance equipment is used to collect echo signals from poultry eggs in the chick stage, and an adaptively adjusted CPMG pulse sequence is used to dynamically set the echo interval and the number of echoes to improve the signal-to-noise ratio and detection efficiency.

[0070] In this embodiment, the signal-to-noise ratio and sampling time of the echo signal are evaluated in real time according to the quality of the sampled echo signal, and the optimal echo interval in the echo signal acquisition process is determined by optimizing the objective function. and echo number ;

[0071] Among them, the optimization objective function for:

[0072]

[0073] Where, and Respectively represent the weights for measuring echo signal quality and sampling efficiency, It represents the ratio of the peak value of the echo signal to the standard deviation of the background noise.

[0074] Specifically, in this embodiment, the poultry eggs to be tested are placed in the detection area of ​​a low-field total resonance device (0.2-1T main magnetic field), the sample posture is kept consistent, and the CPMG pulse sequence is started to collect the echo signal S(t), which reflects the transverse magnetization decay trajectory of the water content inside the egg in the magnetic field over time. In order to ensure that the signal has sufficient discrimination resolution, the echo interval is determined by the above-mentioned optimization objective function. and echo number After parameter tuning, the system finally locked in the sampling scheme and completed signal acquisition. This echo sequence serves as the primary input for subsequent spectral inversion and gender determination, and its quality directly determines the final discrimination accuracy.

[0075] In step S2 of the present embodiment, the collected echo signals are merely raw physical quantities. To make these data interpretable, spectral inversion is immediately performed to convert the time-domain information into a T2 spectrum distribution that reflects differences in tissue structure. This spectrum, in other words, is the intensity spectrum of the magnetic response components at different relaxation time intervals. This spectrum reflects various internal structural information, including tissue water binding status, viscosity, and uniformity.

[0076] In this embodiment, the objective function for performing spectrum inversion on the collected echo signal using the non-negative least squares optimization model combined with the Tikhonov regularization term is:

[0077]

[0078] Where, represents the collected echo signal vector, represents the system response kernel, represents the T2 relaxation time spectrum coefficient, Represents the parameter used to control the smoothness of the spectrum,

[0079] In step S3 of the embodiment of the present invention, after the complete T2 relaxation time spectrum is inverted using the above-mentioned method, it still needs to be further abstracted into numerical features before being input into the subsequent integrated classification model for learning. In the T2 relaxation time spectrum obtained by low-field nuclear magnetic resonance, any subtle changes in the dynamic balance of the embryo's "water-lipid-protein" will leave quantifiable spectral traces. Through experimental comparison of the same batch of poultry eggs, the present invention proposes seven spectral features that can stably and complementary map the physiological differences between males and females.

[0080] In this embodiment, the multidimensional spectral features associated with egg sex information extracted from the T2 relaxation time spectrum include main peak position, spectrum width, energy density ratio, spectrum change rate, spectrum skewness, spectrum entropy and peak shape sharpness;

[0081] Main peak location This is the relaxation time corresponding to the maximum point in the T2 relaxation time spectrum. It can be considered the average correlation time of free water molecules, reflecting the degree of water freedom. In male embryos, increased expression of fatty acid synthase and acetyl-CoA carboxylase results in a relative "liberation" of free water within lipid pores, leading to a prolonged rotational correlation time. Consequently, the main peak position in male embryos shifts to the left compared to female embryos, remaining in the short relaxation region in female embryos.

[0082] The main peak position becomes the "peak position" indicator for distinguishing between males and females. The calculation formula for the main peak position is:

[0083]

[0084] Where, Indicates the spectrum intensity Get the index of the spectrum point with the maximum value, represents the T2 relaxation time corresponding to the maximum spectral intensity, i.e. the main peak position;

[0085] The spectral width W is the extreme difference in the relaxation time distribution in the T2 relaxation time spectrum. It measures the degree of discreteness of the relaxation components in the tissue and reflects the complexity of the composition. In male embryos, the number of lipid vacuoles and water-lipid interfaces increases, and the microscopic heterogeneity intensifies, which widens the spectral width. In females, because water is orderly bound by proteins, the component distribution converges and the spectral width value is narrower. Therefore, the spectral width of male embryos is greater than that of female embryos.

[0086] The spectrum width and the main peak together describe the coordinated changes of "position-scale". The calculation formula is:

[0087]

[0088] Where, represents the spectrum width, and Respectively represent the maximum and minimum relaxation times in the T2 relaxation time spectrum.

[0089] The energy density ratio E is the ratio of the intensities of the fast relaxation segment to the slow relaxation segment; specifically, the energy density ratio E divides the T2 relaxation time spectrum at T2=8ms (after acquisition and local extreme value correction) into a "fast zone" and a "slow zone", and calculates the energy ratio of the two zones. The fast zone signal mainly comes from lipid-bound water, so fat deposition in male embryos causes the energy density ratio E to steadily increase, while that of females remains at a lower level. Therefore, the energy density ratio of male embryos is greater than that of female embryos.

[0090] The energy density ratio E directly reflects the lipid ratio and is the core of “energy type” judgment. Its calculation formula is:

[0091]

[0092] Where, It represents the ratio of the intensity of the fast relaxation segment to the slow relaxation segment. represents the fast / slow relaxation segmentation threshold, represents the energy of lipid-bound water in the fast relaxation phase, represents the free water energy in the slow relaxation phase, and Respectively represent the T2 relaxation time values ​​corresponding to the i-th and j-th spectral points in the spectrum, and Represent the spectral energy of the i-th and j-th spectral points respectively;

[0093] The spectral rate of change, ΔR, is the absolute value of the slope of the T2 relaxation time spectrum on a logarithmic scale and is extremely sensitive to the water-fat turning point. When the fat peak in males rises rapidly, the slope in this region becomes steeper, and ΔR rises. In females, the peak is more gradual, resulting in a smaller slope. Therefore, the spectral rate of change in male embryos is greater than that in female embryos.

[0094] The spectrum change rate ΔR provides "gradient-type" detail compensation and measures the speed of local signal change. The calculation formula of the spectrum change rate ΔR is:

[0095]

[0096] Where, represents the spectrum change rate of the i-th spectrum point in the T2 relaxation time spectrum, and represent the spectral energy of the i-th and i+1-th spectral points respectively, and Respectively represent the T2 relaxation time values ​​corresponding to the i-th and i+1-th spectral points in the spectrum;

[0097] spectral skewness is the normalized value of the third-order central moment, describing the left-right asymmetry of the T2 relaxation time spectrum; the male free water tail is elongated, making The spectral skewness of male embryos drifts in the negative direction, while that of female embryos is close to zero or slightly positive because of their short and symmetrical tails. Therefore, the spectral skewness of male embryos drifts in the negative direction, while the spectral skewness of female embryos is close to zero or slightly positive.

[0098] spectral skewness The global information of “morphology” is given, which can complement ΔR to suppress noise and spectral skewness. The calculation formula is:

[0099]

[0100] Where, represents the spectral skewness, represents the total number of spectral points, and represent the mean and standard deviation of spectral energy, respectively. Represents the spectral energy of the i-th spectral point, reflecting the symmetry of the spectral distribution.

[0101] Spectral entropy To measure the uniformity of T2 relaxation time spectrum energy distribution; in male embryos, H is significantly reduced due to the concentrated energy of the fat peak, while in female embryos, H is more evenly distributed. Therefore, the spectral entropy of male embryos is smaller than that of female embryos.

[0102] Introducing spectral entropy After , the model’s robustness to abnormal samples is improved, and its calculation formula is:

[0103]

[0104]

[0105] Where, represents the spectral entropy, represents the normalized energy, represents the spectral energy of the i-th spectral point, Represents the total energy of all spectral points.

[0106] Peak sharpness To describe the sharpness and sharpness of the main peak of the T2 relaxation time spectrum; the free water peak of females is sharper due to the restriction of the protein network, and Kurt is higher than that of males. The peak shape of males is blunted due to lipid filling; therefore, the peak shape of male embryos is less sharp than that of female embryos.

[0107] After taking peak sharpness Kurt as an additional "morphological detail" dimension, the calculation formula for false negative rate and peak sharpness is:

[0108]

[0109] Where, Indicates the peak sharpness, n indicates the total number of spectral points, represents the standard deviation of the spectral energy, represents the spectral energy of the i-th spectral point.

[0110] In this embodiment, the above-mentioned characteristics associated with the multidimensional relaxation spectrum of poultry egg sex respectively characterize the changes in the "water-lipid-protein" balance in the embryo from the perspectives of peak position, scale, energy proportion, local gradient, overall morphology and information entropy, forming a closed and reproducible "feature-mechanism-sex" correspondence chain, which can be effectively used as input data for sex discrimination as input to the sex classification model.

[0111] In the embodiment of the present invention, step S4 includes the following sub-steps:

[0112] S41, input the high-dimensional feature vector representing the microstructural differences of embryonic tissue into the integrated classification model, and output the corresponding probability distribution P m (y|x); where y∈{male, female, uncertain} represents the gender label and x represents the high-dimensional feature vector;

[0113] The integrated classification model includes a support vector machine model, a random forest model, and a shallow neural network. S42 uses a weighted voting fusion strategy to aggregate the probability distributions output by the support vector machine model, the random forest model, and the shallow neural network to obtain the final comprehensive predicted gender label. ;

[0114] Among them, comprehensive prediction of gender label for:

[0115]

[0116] Where M=3 represents the number of models, represents the fusion weight of each model (set or determined by the validation set), ;

[0117] Specifically, in the classification stage, the system outputs the gender label based on the above calculations , the label is one of "male" or "female", indicating that the integrated classification model has completed a clear gender judgment based on the spectral features. The judgment result will serve as the input basis for the subsequent confidence mechanism and sorting control;

[0118] S43. Comprehensive prediction of gender labels Conduct confidence evaluation and output the comprehensive predicted sex labels that meet the confidence evaluation criteria as the egg sex identification results, which serve as a reference for automated sex sorting control of eggs.

[0119] Among them, the confidence evaluation formula is:

[0120]

[0121]

[0122] Where, Represents the probability value of the i-th category of the output, Represents information entropy. If the confidence level is lower than the preset threshold, the sample is judged as "uncertain" and the system marks the sample for manual re-inspection without automatic sorting. The confidence evaluation formula is based on the probability value. Perform confidence evaluation to measure the prediction accuracy of the model for a certain category.

[0123] In step S3 of the embodiment of the present invention, in order to enhance the system's adaptability to non-ideal factors such as temperature, humidity, and electromagnetic disturbance in different production environments, step S3 of the present invention further includes:

[0124] A disturbance compensation mechanism based on multi-sensor environmental detection is introduced to compensate for the drift of the extracted multi-dimensional spectral features.

[0125] The formula for compensation calculation is:

[0126]

[0127] Where, represents the corrected stable multidimensional spectral characteristics, represents the characteristics before correction, 、 and It represents the empirically calibrated temperature interference sensitivity coefficient, humidity interference sensitivity coefficient and electromagnetic disturbance interference sensitivity coefficient. 、 and They represent temperature drift value, humidity drift value and electromagnetic disturbance drift value respectively.

[0128] The above-mentioned disturbance compensation mechanism provided in this embodiment significantly improves the stability and generalization performance of the model under multi-plant, multi-time period, and multi-equipment conditions.

[0129] The egg sexing method provided in the embodiments of this invention is a recommended technical approach. Its key is to convert NMR relaxation information into structural features and input them into a model to achieve intelligent sex determination. Any approach employing "low-field NMR + spectral feature construction + model recognition," or any adjustments based on the method of this invention, such as algorithm substitution, parameter combination, and output mechanism optimization, shall be considered equivalent technical solutions within the scope of this invention.

[0130] Specific embodiments are used in the present invention to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core ideas. At the same time, for those skilled in the art, according to the ideas of the present invention, there may be changes in the specific implementation methods and application scopes. In summary, the contents of this specification should not be understood as limiting the present invention.

[0131] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and it should be understood that the scope of protection of the present invention is not limited to such specific descriptions and embodiments. Those skilled in the art can make various other specific variations and combinations based on the technical teachings disclosed in the present invention without departing from the essence of the present invention, and such variations and combinations are still within the scope of protection of the present invention.

Claims

1. A method for non-destructive sex identification of poultry eggs based on low-field nuclear magnetic resonance, characterized in that: The following steps are involved: S1. Using low-field nuclear magnetic resonance equipment to collect echo signals from eggs in the chick stage; S2. Using a non-negative least squares optimization model combined with the Tikhonov regularization term to perform spectral inversion on the collected echo signals, reconstructing the T2 relaxation time spectrum related to the distribution of water and tissue structure in the egg; S3, extracting multidimensional spectral features associated with egg sex information from the T2 relaxation time spectrum to form a high-dimensional feature vector representing the microstructural differences of embryonic tissue; S4. Input the high-dimensional feature vector into the integrated classification model and output the egg sex identification result.

2. The method for non-destructive sex identification of poultry eggs based on low-field nuclear magnetic resonance according to claim 1, wherein: In step S1, the signal-to-noise ratio and sampling time of the echo signal are evaluated in real time according to the quality of the sampled echo signal, and the optimal echo interval in the echo signal acquisition process is determined by optimizing the objective function. and echo number ; The optimization objective function for: Where, and Respectively represent the weights for measuring echo signal quality and sampling efficiency, It represents the ratio of the peak value of the echo signal to the standard deviation of the background noise.

3. The method for non-destructive sex identification of poultry eggs based on low-field nuclear magnetic resonance according to claim 2, wherein: In step S2, the objective function of performing spectrum inversion on the collected echo signal using the non-negative least squares optimization model combined with the Tikhonov regularization term is: Where, represents the collected echo signal vector, represents the system response kernel, represents the T2 relaxation time spectrum coefficient, Represents the parameter used to control the smoothness of the spectrum, represents the first-order gradient operator.

4. The method for non-destructive sex identification of poultry eggs based on low-field nuclear magnetic resonance according to claim 1, wherein In step S3, multidimensional spectral features associated with egg sex information are extracted from the T2 relaxation time spectrum, including main peak position, spectrum width, energy density ratio, spectrum change rate, spectrum skewness, spectrum entropy and peak shape sharpness; The main peak position is the relaxation time corresponding to the point with the maximum relaxation spectrum amplitude in the T2 relaxation time spectrum; The main peak position of male embryos shifted to the left as a whole compared with that of female embryos, and the main peak position of female embryos remained in the short relaxation region; The spectrum width is the range of the relaxation time distribution in the T2 relaxation time spectrum; the spectrum width of male embryos is greater than that of female embryos; The energy density ratio is the ratio of the intensity of the fast relaxation segment to the intensity of the slow relaxation segment; the energy density ratio of the male embryo is greater than that of the female embryo; The spectrum change rate is the absolute value of the slope of the T2 relaxation time spectrum in logarithmic coordinates; the spectrum change rate of male embryos is greater than that of female embryos; The spectral skewness is the normalized value of the third-order central moment, which describes the left-right asymmetry of the T2 relaxation time spectrum. The spectral skewness of male embryos drifts in the negative direction, while the spectral skewness of female embryos is close to zero or slightly positive. The spectral entropy is a measure of the uniformity of the energy distribution of the T2 relaxation time spectrum; the spectral entropy of male embryos is smaller than that of female embryos; The peak sharpness describes the sharpness and sharpness of the main peak of the T2 relaxation time spectrum; the peak sharpness of male embryos is smaller than that of female embryos.

5. The method for non-destructive identification of egg sex based on low-field nuclear magnetic resonance according to claim 4, characterized in that: The calculation formula of the main peak position is: Where, Indicates the spectrum intensity Get the index of the spectrum point with the maximum value, Indicates the T2 relaxation time corresponding to the maximum spectral intensity, that is, the main peak position; The calculation formula of the spectrum width is: Where, represents the spectrum width, and represent the maximum and minimum relaxation times in the T2 relaxation time spectrum; The calculation formula of the energy density ratio is: Where, It represents the ratio of the intensity of the fast relaxation segment to the slow relaxation segment. represents the fast / slow relaxation segmentation threshold, represents the energy of lipid-bound water in the fast relaxation phase, represents the free water energy in the slow relaxation phase, and Respectively represent the T2 relaxation time values ​​corresponding to the i-th and j-th spectral points in the spectrum, and Represent the spectral energy of the i-th and j-th spectral points respectively; The calculation formula of the spectrum change rate is: Where, represents the spectrum change rate of the i-th spectrum point in the T2 relaxation time spectrum, and represent the spectral energy of the i-th and i+1-th spectral points respectively, and Respectively represent the T2 relaxation time values ​​corresponding to the i-th and i+1-th spectral points in the spectrum; The calculation formula of the spectral skewness is: Where, represents the spectral skewness, represents the total number of spectral points, and represent the mean and standard deviation of spectral energy, respectively. represents the spectral energy of the i-th spectral point; The calculation formula of the spectral entropy is: Where, represents the spectral entropy, represents the normalized energy, represents the spectral energy of the i-th spectral point, represents the total energy of all spectral points; The calculation formula of the peak sharpness is: Where, Indicates the peak sharpness, n indicates the total number of spectral points, represents the standard deviation of the spectral energy, represents the average spectral energy, represents the spectral energy of the i-th spectral point.

6. The method for non-destructive sex identification of poultry eggs based on low-field nuclear magnetic resonance according to claim 4, characterized in that: The step S4 comprises the following sub-steps: S41, input the high-dimensional feature vector representing the microstructural differences of embryonic tissue into the integrated classification model, and output the corresponding probability distribution P m (y|x); where y∈{male, female, uncertain} represents the gender label and x represents the high-dimensional feature vector; The integrated classification model includes a support vector machine model, a random forest model and a shallow neural network; S42. Use a weighted voting fusion strategy to aggregate the probability distributions output by the support vector machine model, random forest model, and shallow neural network to obtain the final comprehensive predicted gender label. ; S43. Comprehensive prediction of gender labels A confidence evaluation is performed, and the comprehensive predicted sex labels that meet the confidence evaluation standards are output as the sex identification results of poultry eggs, which serve as a reference for the control of automated sex sorting of poultry eggs.

7. The method for non-destructive sex identification of poultry eggs based on low-field nuclear magnetic resonance according to claim 1, wherein: The step S3 further includes: A disturbance compensation mechanism based on multi-sensor environmental detection is introduced to compensate for the drift of the extracted multi-dimensional spectral features. The formula for compensation calculation is: Where, represents the corrected stable multidimensional spectral characteristics, represents the characteristics before correction, 、 and It represents the empirically calibrated temperature interference sensitivity coefficient, humidity interference sensitivity coefficient and electromagnetic disturbance interference sensitivity coefficient. 、 and They represent temperature drift value, humidity drift value and electromagnetic disturbance drift value respectively.

Images