A method for nondestructive identification of gender of crossbred duck eggs based on megaphone nuclear magnetic image

Abstract

The application discloses a kind of based on trumpet nuclear magnetic image's sex nondestructive identification method of turning duckling eggs, belong to poultry egg detection technical field, this method uses low-field nuclear magnetic resonance technology, first in nuclear magnetic system Positioning egg and determine the optimal imaging section of trumpet, image is collected and preprocessed to obtain high-quality trumpet nuclear magnetic image, then through analysis relaxation characteristics, inversion nuclear magnetic signal identifies trumpet structure and extracts features, finally constructs gender decision model, combined with physical constraint obtains gender discrimination result and is sorted as reference.The application uses the nuclear magnetic characteristics of the gender difference of trumpet structure, realizes nondestructive, early, high-accuracy gender identification through special imaging technology and algorithm, can optimize hatching and breeding strategy in egg duck and meat duck breeding, promote the intelligent development of industry.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of poultry egg detection, and particularly relates to a method for nondestructive identification of the gender of a duckling breeding egg based on a syrinx nuclear magnetic image. BACKGROUND

[0002] In the duck industry, the values of male and female individuals are significantly different, especially in the process of breeding and raising of duck eggs, accurate identification of the gender of the embryo in the egg at an early stage of hatching can greatly improve resource allocation efficiency, reduce feeding costs and improve hatching output. However, traditional gender identification methods such as DNA detection, hormone marking and appearance observation have problems such as shell breaking, low efficiency or inability to achieve industrialization. In recent years, in order to achieve nondestructive gender detection of poultry eggs, researchers have tried to use various advanced detection technologies, such as near-infrared spectroscopy and laser scattering methods for analyzing eggshell transmission light signals; ultrasonic imaging and reflection analysis technology for evaluating embryo tissue structure or sound velocity difference; multi-modal sensing technologies such as hyperspectral imaging and microwave perspective for obtaining internal optical or electromagnetic characteristics through the eggshell. Although these methods have made some progress, they still have limitations such as limited sensitivity, large noise interference, significant influence of eggshell thickness or position, and large equipment size, making it difficult to achieve non-contact identification in an industrial hatching environment. In particular, existing technologies do not fully utilize the important physiological characteristics of the duckling stage, during which the syrinx has fully developed and male and female individuals show obvious differences in syrinx structure. SUMMARY

[0003] In view of the above problems in the prior art, the present application provides a method for nondestructive identification of the gender of a duckling breeding egg based on a syrinx nuclear magnetic image, which is based on the characteristic performance of the syrinx tympanic cavity structure of male duck embryo in the nuclear magnetic image, combined with optimized nuclear magnetic imaging technology, image processing technology and discrimination algorithm, to realize early nondestructive identification of the gender of a duckling breeding egg.

[0004] In order to achieve the above-mentioned application purposes, the technical scheme adopted by the present application is as follows: a method for nondestructive identification of the gender of a duckling breeding egg based on a syrinx nuclear magnetic image, comprising the following steps:

[0005] S1, positioning the duckling breeding egg in a nuclear magnetic resonance system and determining the optimal imaging section of the syrinx structure of the duckling breeding egg;

[0006] S2, based on the optimal imaging section, collecting a plurality of syrinx nuclear magnetic images through a special nuclear magnetic resonance imaging sequence combination of the syrinx structure, and pre-processing the syrinx nuclear magnetic images to select high-quality syrinx nuclear magnetic images;

[0007] In the high-quality syrinx nuclear magnetic image, the right side of the end of the trachea of the male duck embryo has a high-light circular structure;

[0008] S3, identify the syrinx structure in the high-quality syrinx magnetic resonance image by analyzing the relaxation characteristics of the syrinx structure and inverting the nuclear magnetic signal, and quantitatively extract the syrinx structure characteristics;

[0009] S4, construct a syrinx gender decision model, and identify the gender discrimination result corresponding to the syrinx structure characteristics as a reference for sorting the duckling eggs.

[0010] Further, the step S1 is specifically:

[0011] In the nuclear magnetic resonance system, the duckling eggs are positioned at the center of the magnetic field by positioning the center, the best magnetic field uniformity is obtained, and the position of the radio frequency coil is optimized.

[0012] Based on the positioning position of the duckling eggs, a plurality of optimal imaging sections of the syrinx structure are determined by multi-plane trial scanning.

[0013] Further, the step S2 includes the following steps:

[0014] S21, design a special nuclear magnetic resonance imaging sequence combination of the syrinx structure, including T2 weighted sequence, T1 weighted sequence, magnetic susceptibility weighted sequence and fat suppression sequence;

[0015] S22, based on the special nuclear magnetic resonance imaging sequence combination, the embryo position of the duckling eggs is positioned by multi-plane positioning scanning to determine the tracheal bifurcation plane, and the optimal imaging time window is selected according to the embryo development stage, and a plurality of syrinx magnetic resonance images under the optimal imaging section of the syrinx structure are collected;

[0016] S23, the collected syrinx magnetic resonance images are subjected to image enhancement processing to obtain enhanced syrinx magnetic resonance images;

[0017] The image enhancement processing includes contrast enhancement, multi-sequence image fusion, three-dimensional reconstruction and parameter map generation processing;

[0018] S24, the enhanced syrinx magnetic resonance images are subjected to image quality optimization processing to obtain high-quality syrinx magnetic resonance images;

[0019] The image quality optimization processing includes motion correction, artifact removal processing and quality evaluation.

[0020] Further, in the step S24, the quality of each enhanced syrinx magnetic resonance image is evaluated by constructing an image quality evaluation model, and the enhanced syrinx magnetic resonance image with the highest image quality score is selected as the high-quality syrinx magnetic resonance image;

[0021] When the quality evaluation scores of all enhanced syrinx magnetic resonance images are less than the preset threshold, return to step S1 to reposition and collect the syrinx magnetic resonance images;

[0022] The expression of the image quality evaluation model is:

[0023] ;

[0024] In the formula, Q represents the image quality score, SNR represents the signal-to-noise ratio, CNR represents the contrast-to-noise ratio, SD represents the structural definition, AR represents the artifact degree, 、 、 and respectively represent the weight coefficients of 、 、 and .

[0025] Further, the step S3 comprises the following sub-steps:

[0026] S31, analyze the relaxation characteristics of the syrinx structure in the high-quality syrinx nuclear magnetic image, and determine the T1 and T2 relaxation times of the syrinx structure;

[0027] On the basis of the relaxation characteristic analysis, syrinx structure-specific nuclear magnetic signal inversion is performed to determine the magnetic susceptibility;

[0028] S32, locate the syrinx structure in the high-quality syrinx nuclear magnetic image through a trachea tracking algorithm;

[0029] S33, according to the T1 and T2 relaxation times and the magnetic susceptibility of the identified syrinx structure, perform feature quantization analysis on the high-quality syrinx nuclear magnetic image, and extract multi-dimensional features of the syrinx structure, including morphological features, signal features and texture features;

[0030] S34, based on the extracted multi-dimensional features, construct a high-dimensional feature space, and through feature screening, dimension reduction and recursive feature elimination processing, obtain quantitative syrinx structure features.

[0031] Further, in the step S32, the syrinx structure in the high-quality syrinx nuclear magnetic image is located through a trachea tracking algorithm, comprising:

[0032] Initialize the seed point in the trachea region at the top of the turning duck egg;

[0033] For the high-quality syrinx nuclear magnetic image apply morphological operations to obtain a first image ;

[0034] extract the skeleton in the first image For each point on the extracted skeleton , calculate its direction vector ; ;

[0035] Starting from the seed point , trace the trachea along the direction to generate a trachea centerline track ;

[0036] In the terminal region of the trachea centerline track , search for a syrinx structure tympanic candidate region that meets the conditions;

[0037] The conditions include: located on the right side of the trachea terminal, circularity greater than 0.8, signal intensity ratio greater than 1.8, and diameter in the range of 2.2-3.0mm;

[0038] wherein, denotes a morphological closing operation, denotes a morphological opening operation, denotes the size of the opening operation structure element, denotes the size of the closing operation structure element, denotes a skeleton extraction function;

[0039] Further, the step S4 comprises the following sub-steps:

[0040] S41, establish a syrinx gender feature discrimination standard system, which includes the existence and position of the syrinx tympanicum, the diameter threshold of the syrinx tympanicum, the signal intensity ratio threshold, the T2 relaxation time threshold, the magnetic susceptibility threshold and the structural consistency in consecutive layers;

[0041] S42, construct a syrinx gender decision model;

[0042] The syrinx gender decision model includes a gender discrimination layer, a confidence assessment layer and a decision output layer; wherein the gender discrimination layer is used to pre-screen the syrinx tympanic structure and its basic features based on the constructed syrinx gender feature discrimination standard system; then according to the extracted syrinx structure features and their weights, and combined with the physical constraints of the syrinx structure, the quantified gender discrimination result is obtained through a linear discriminant function; the confidence assessment layer is used to evaluate the accuracy of the gender discrimination result; the decision output layer is used to output the final gender discrimination result according to the gender discrimination result and its confidence;

[0043] S43, input the syrinx structure features into the syrinx gender decision model to output the corresponding gender discrimination result.

[0044] Further, in the gender discrimination layer in the step S42, the gender discrimination function is:

[0045] ;

[0046] wherein, the discriminant function value, when the gender discriminant result is male, and when the gender discriminant result is female, represents a bias term, represents the i-th syrinx structural feature corresponding weight coefficient;

[0047] The physical constraint of the gender discriminant function is:

[0048] ;

[0049] ;

[0050] wherein, represents the actual nuclear magnetic signal intensity after considering the influence of magnetic field inhomogeneity, represents the intensity of the nuclear magnetic signal with echo time TE and repetition time TR, represents the gyromagnetic ratio, represents the local magnetic field uncertainty, , represents the magnetic susceptibility difference, represents the main magnetic field degree, represents the geometric factor related to the structural direction, represents the echo time, represents the repetition time, represents the relaxation time, represents the relaxation time, represents the proton density.

[0051] Further, the step S4 further comprises: based on the gender discriminant result, displaying and optimizing the syrinx structure in the high-quality syrinx nuclear magnetic image through a quantitative enhancement algorithm.

[0052] The enhancement algorithm comprises adaptive pseudo-color mapping of the syrinx structure, syrinx structure boundary sharpening enhancement, quantitative superimposed display of the gender discriminant result, and enhancement effect quantitative evaluation.

[0053] Compared with the existing non-destructive gender detection methods such as spectroscopy and ultrasonic method, the present application has the following significant technical advantages and effects:

[0054] 1. Direct identification of syrinx structure difference:

[0055] The application utilizes the syrinx tympanic cavity structure unique to male duck embryos, directly observes anatomical differences through nuclear magnetic imaging, avoids the uncertainty of indirect signal speculation, and realizes the precise identification of "structure visualization".

[0056] 2. Non-destructive deep detection capability:

[0057] In the application, nuclear magnetic resonance technology can be used to non-destructively penetrate the eggshell to directly image the internal structure of the embryo, without the need for contact with external sensors or the breaking of the shell, thereby maintaining the integrity of the embryo and effectively improving the hatching rate.

[0058] 3. Optimal detection timing during the brooding period:

[0059] In the application, detection is performed during the brooding period when the syrinx is fully developed, thereby fully utilizing the physiological characteristics of the most significant gender difference at this stage and ensuring the accuracy and stability of the detection.

[0060] 4. Intelligent airway tracking technology:

[0061] In the application, a dedicated airway tracking algorithm is used to automatically locate the syrinx structure and extract multi-dimensional features, thereby realizing a technical breakthrough from manual identification to intelligent detection.

[0062] 5. Anti-interference imaging advantage:

[0063] In the application, a combination of T2-weighted, T1-weighted, magnetic susceptibility-weighted, and fat suppression sequence special nuclear magnetic resonance imaging sequences is used, which is not affected by external factors such as eggshell thickness and lighting conditions, and has strong environmental adaptability. BRIEF DESCRIPTION OF DRAWINGS

[0064] Figure 1 A flowchart of the non-destructive gender identification method for brooding duck eggs based on syrinx nuclear magnetic images is provided. DETAILED DESCRIPTION

[0065] The specific embodiments of the application are described below to facilitate understanding of the application by those skilled in the art, but it should be clear that the application is not limited to the scope of the specific embodiments, and for those skilled in the art, any changes within the spirit and scope of the application as defined in the appended claims are obvious, and all applications utilizing the concept of the application are within the scope of protection.

[0066] The embodiment of the application provides a non-destructive gender identification method for brooding duck eggs based on syrinx nuclear magnetic images, as shown in Figure 1 The method comprises the following steps:

[0067] S1, positioning the brooding duck eggs in a nuclear magnetic resonance system and determining the optimal imaging section of the syrinx structure of the brooding duck eggs;

[0068] S2, based on the optimal imaging section, a plurality of syrinx nuclear magnetic images are collected through a special nuclear magnetic resonance imaging sequence combination of the syrinx structure, and preprocessed to obtain high-quality syrinx nuclear magnetic images;

[0069] In the high-quality syrinx nuclear magnetic image, the male duck embryo has a high-light circular structure on the right side of the tracheal end;

[0070] S3, the syrinx structure in the high-quality syrinx nuclear magnetic image is identified and the syrinx structure features are quantitatively extracted through the relaxation characteristic analysis and nuclear magnetic signal inversion of the syrinx structure;

[0071] S4, a syrinx gender decision model is constructed, and the gender discrimination result corresponding to the syrinx structure feature is identified as a reference for sorting the duck egg.

[0072] The step S1 of the embodiment of the application is specifically:

[0073] In the nuclear magnetic resonance system, the duck egg is placed in the center of the magnetic field through positioning and centralization, the best magnetic field uniformity is obtained, and the position of the radio frequency coil is optimized.

[0074] Based on the positioning position of the duck egg, a plurality of optimal imaging sections of the syrinx structure are determined through multi-planar trial scanning.

[0075] In order to ensure that the best syrinx structure image is obtained in the embodiment, the placement position and direction of the duck egg in the nuclear magnetic resonance system are optimized in the embodiment to ensure that the best imaging effect is obtained; specifically, the duck egg is placed in a specific direction so that the tracheal long axis and the main magnetic field direction are as parallel as possible, the magnetic susceptibility artifact is minimized, the syrinx region should be located at the center position of the coil sensitive area to maximize the signal, usually the duck egg is placed with the large end up, and a special fixing device is used to ensure stability; such accurate positioning technology is a prerequisite for obtaining high-quality syrinx nuclear magnetic images, which ensures that the nuclear magnetic resonance system can image the syrinx region in the best state.

[0076] Based on the reasonable positioning of the egg body, the best display section of the syrinx structure is determined through multi-planar trial scanning in the embodiment, and multi-planar reconstruction is performed using isotropic voxel data to obtain a display of the section at any angle; specifically, the tracheal long axis vertical plane, located at 0.5-1.5mm before the tracheal bifurcation, is usually the best imaging plane of the syrinx structure, and the curved surface reconstruction technology is used to reconstruct along the tracheal running direction, which can display the entire trachea-syrinx system at one time.

[0077] In this embodiment, through egg positioning and section optimization, a complete set of syrinx region imaging optimization scheme is established, which provides technical support for obtaining high-quality syrinx nuclear magnetic images, and creates favorable conditions for subsequent image analysis and feature extraction.

[0078] In step S2 of the embodiment of the present application, first, based on the anatomical difference research on the syrinx structure of male and female duck embryos, the key anatomical landmarks for gender identification are determined; the male duck has a developed syringeal bulla, which is a gas-filled cavity structure, located on the right side of the tracheal bifurcation, while the syrinx structure of the female duck is relatively simple and has no obvious bulla; this anatomical difference is manifested as obvious signal characteristic difference in nuclear magnetic resonance imaging, which provides a basis for non-destructive gender identification. Specifically, the syringeal bulla of the male duck appears as a high signal circular or spherical bright spot in the nuclear magnetic image, with a diameter usually in the range of 2.2-3.0 mm; while the signal of the female duck tracheal end region is uniform, without characteristic high-brightness structure, and the overall morphology is relatively smooth. Nuclear magnetic resonance accurately captures this anatomical difference by detecting the distribution density and relaxation characteristics of hydrogen protons in different tissues, providing a reliable imaging basis for gender identification.

[0079] Further, the tissue contrast characteristics of nuclear magnetic resonance imaging technology are used, based on the different relaxation characteristics of the gas in the syringeal bulla and the surrounding soft tissue in nuclear magnetic resonance, to form signal contrast and realize clear display of the syrinx structure. By optimizing the nuclear magnetic resonance sequence parameters, the signal difference between the syrinx structure and the surrounding tissue is maximized, and the identification accuracy is improved. Specifically, the gas-filled syringeal bulla appears as a clear high signal area in the T2 weighted image, with clear boundaries and a distinct contrast with the surrounding soft tissue. In particular, the gas-soft tissue interface of the syringeal bulla produces a unique magnetic susceptibility effect in the gradient echo sequence, further enhancing the accuracy of structure recognition.

[0080] In this embodiment, through the above in-depth research on the anatomical characteristics of the syrinx structure and the signal characteristics of nuclear magnetic resonance, the theoretical basis for the subsequent implementation of the optimized imaging technology and parameters for gender identification of cross-foster duck eggs is provided, guiding the design and optimization of specific nuclear magnetic resonance sequences.

[0081] Based on this, step S2 of the embodiment of the present application includes the following sub-steps:

[0082] S21, design a special nuclear magnetic resonance imaging sequence combination for the syrinx structure, including T2 weighted sequence, T1 weighted sequence, magnetic susceptibility weighted sequence and fat suppression sequence;

[0083] S22, based on the combination of the special magnetic resonance imaging sequence, on the basis of the optimal imaging section, the embryo position of the duckling egg is determined by multi-plane positioning scanning, and the tracheal bifurcation plane is determined, and combined with the optimal imaging time window selected according to the embryo development stage, a plurality of tracheal magnetic resonance images under the optimal imaging section of the tracheal structure are collected;

[0084] S23, the collected tracheal magnetic resonance image is subjected to image enhancement processing to obtain an enhanced tracheal magnetic resonance image;

[0085] The image enhancement processing includes contrast enhancement, multi-sequence image fusion, three-dimensional reconstruction and parameter map generation processing;

[0086] S24, the enhanced tracheal magnetic resonance image is subjected to image quality optimization processing to obtain a high-quality tracheal magnetic resonance image;

[0087] The image quality optimization processing includes motion correction, artifact removal processing and quality evaluation.

[0088] In step S21 of the embodiment, based on the understanding of the anatomical characteristics and signal characteristics of the tracheal structure, a special magnetic resonance imaging sequence combination is designed in the embodiment, which is optimized for the special tissue characteristics of the tracheal structure, including T2 weighted sequence, T1 weighted sequence, susceptibility weighted sequence and fat suppression sequence, etc., to obtain the best display effect of the tracheal structure. The combination of these sequences makes full use of the multi-parameter imaging characteristics of nuclear magnetic resonance, and is optimized for the physical characteristics of the tracheal tissue.

[0089] Specifically, the fast spin echo T2 weighted sequence (TR=2500-3500ms, TE=80-100ms, slice thickness 2mm, matrix 256x256, FOV=80mm, NEX=2) is used to highlight the gas-soft tissue contrast; the three-dimensional gradient echo T1 weighted sequence (TR=400-600ms, TE=10-15ms, flip angle=30°, isotropic voxel 0.5mm) is used to finely display the structure boundary; the susceptibility weighted imaging (TR=49ms, TE=40ms, flip angle=15°) enhances the gas-tissue interface contrast; and the fat suppression sequence using STIR technology (TI=150ms) highlights the tracheal structure.

[0090] In the embodiment, the above-mentioned optimized imaging sequence lays a foundation for obtaining high-quality tracheal structure images, and provides reliable raw data for subsequent image processing and analysis.

[0091] In step S22 of the embodiment, based on the optimization of the imaging sequence, a special nuclear magnetic resonance positioning scanning technology is developed, the embryo position is determined by multi-plane positioning, the tracheal bifurcation plane is determined based on the positioning image, the multi-layer thin layer scanning strategy is adopted, and the best imaging time window selected according to the embryo development stage is combined to ensure that a plurality of syrinx nuclear magnetic images under the optimal imaging section of the syrinx structure are collected.

[0092] Specifically, three-plane positioning is performed using a fast gradient echo sequence (TR / TE=10 / 4 ms), and 8-10 layers are continuously scanned from the top of the egg downward (layer thickness 1.5 mm, layer spacing 0.2 mm). The best imaging time window is during the hatching period (after 28 days of incubation), at which time the syrinx structure is mature but the embryo moves less. The best imaging layer is usually the first 4 layers of images from the top of the egg downward, in which the syrinx development is relatively mature, the structural differences are obvious, the signal-to-noise ratio of the images is the highest, and the artifacts are the least.

[0093] In the embodiment, the above-mentioned precise positioning and scanning technology ensures that the syrinx nuclear magnetic image of the best display layer of the syrinx structure can be obtained, which provides an accurate anatomical basis for subsequent image analysis and gender determination.

[0094] In step S23 of the embodiment, after the syrinx nuclear magnetic image is collected, the display of the syrinx structure is optimized through contrast enhancement, multi-sequence image fusion, three-dimensional reconstruction, and parameter map generation processing to improve the visualization effect of the syrinx structure in the nuclear magnetic image.

[0095] Specifically, the adaptive contrast enhancement algorithm is used to automatically adjust the window width and window level according to the signal characteristics of the syrinx region, the T1WI, T2WI, and SWI sequence images are registered and fused, the trachea-syrinx system is three-dimensionally reconstructed based on the isosurface extraction algorithm, and the T1 and T2 relaxation time maps and the susceptibility map are generated to provide quantitative analysis basis for tissue characteristics.

[0096] In the embodiment, the above-mentioned image enhancement and reconstruction technology significantly improves the display quality of the syrinx structure in the nuclear magnetic image, and lays a foundation for the next feature extraction and quantitative analysis. Through the optimized imaging sequence, precise layer positioning, and effective image enhancement, a complete high-quality imaging technology system of the syrinx structure is formed.

[0097] In step S24 of the embodiment, during the image quality optimization process:

[0098] To solve the influence of embryo micro-movement on the quality of nuclear magnetic images, on the basis of obtaining enhanced syrinx nuclear magnetic images, motion compensation and correction techniques are used in the embodiment, including fast acquisition sequence, real-time motion detection, post-processing motion correction, and encoding direction optimization, to effectively solve the interference of embryo micro-movement on the imaging of the syrinx structure.

[0099] Specifically, single-shot fast spin echo (SS-FSE) and parallel imaging technique (SENSE factor = 2) are adopted to shorten the single-layer scanning time to <1 second, the embryo motion is monitored in real time by using the navigator echo technique, the frequency encoding direction is parallel to the expected motion direction, and the signal-to-noise ratio is improved by using NSA = 2-3 times average.

[0100] In the embodiment, the motion correction technology effectively reduces the influence of embryo micro-motion on image quality, ensures the clarity and accuracy of the syrinx structure image, and provides a reliable data basis for subsequent image analysis.

[0101] On the basis of motion correction, the embodiment adopts a series of syrinx region-specific artifact reduction methods, including motion-insensitive imaging, local shimming, gradient correction, and signal averaging, according to the special structural characteristics of the syrinx region, which significantly improves the quality of the syrinx region magnetic resonance image.

[0102] Specifically, PROPELLER technology is used for motion-insensitive imaging, local magnetic field homogeneity optimization is performed on the syrinx region, geometric distortion caused by gradient nonlinearity is compensated, and spatial consistency of multi-layer syrinx images is ensured through inter-layer registration.

[0103] To ensure the controllability and consistency of image quality, an image quality evaluation model is constructed to consider factors such as signal-to-noise ratio, contrast, clarity, and artifact degree to evaluate the quality of each enhanced syrinx magnetic resonance image, and the enhanced syrinx magnetic resonance image with the highest image quality score is selected as the high-quality syrinx magnetic resonance image; wherein, by constructing the image quality evaluation model, the system can automatically evaluate the quality of each syrinx image, select the best image for gender discrimination, or prompt to reacquire.

[0104] When the quality evaluation scores of all enhanced syrinx magnetic resonance images are less than the preset threshold, return to step S1 to reposition and acquire the syrinx magnetic resonance image;

[0105] wherein, the expression of the image quality evaluation model is:

[0106] ;

[0107] In the formula, represents the image quality score, represents the signal-to-noise ratio, represents the contrast-to-noise ratio, represents the structural clarity, represents the artifact degree, 、 、 and respectively represent 、 、 and The weighting coefficients.

[0108] In this embodiment, a complete system for optimizing the quality of nasal MRI images was constructed through motion correction, artifact removal, and quality evaluation. This ensured the accuracy and reliability of image analysis and feature extraction, laying a solid data foundation for subsequent gender determination.

[0109] In step S3 of this embodiment of the invention, after acquiring a high-quality nasal MRI image, the precise location and feature extraction of the nasal duct structure are achieved through tracheal tracking and automatic nasal duct structure recognition algorithms, providing a reliable data foundation for gender determination; specifically, step S3 includes the following sub-steps:

[0110] S31. Analyze the relaxation characteristics of the syrinx structure in high-quality syrinx MRI images and determine the T1 and T2 relaxation times of the syrinx structure.

[0111] Based on the relaxation characteristics analysis, the susceptibility was determined by inverting the syrinx structure-specific NMR signal.

[0112] S32. Locate the nasal duct structure in a high-quality nasal duct MRI image using a tracheal tracking algorithm;

[0113] S33. Based on the T1 and T2 relaxation times and magnetic susceptibility of the syrinx structure, perform feature quantification analysis on high-quality syrinx MRI images to extract multidimensional features of the syrinx structure, including morphological features, signal features and texture features.

[0114] S34. Based on the extracted multidimensional features, a high-dimensional feature space is constructed, and quantitative syrinx structure features are obtained through feature filtering, dimensionality reduction and recursive feature elimination.

[0115] In step S31 of this embodiment, a specially designed nuclear magnetic resonance sequence is used to analyze the relaxation characteristics of the syrinx structure, measure the T1 and T2 relaxation times and magnetic susceptibility characteristics of different tissues, and establish a tissue characteristic map to provide a quantitative basis for sex identification.

[0116] Specifically, the T2 relaxation time was measured using a multi-echo spin echo sequence, and the T1 relaxation time was measured using a variable flip angle gradient echo sequence. The CPMG sequence parameters were set to a sampling bandwidth of 333.333 kHz, a sampling point count of 1,333,374, and a waiting time of 2500 ms. In the tympanic region of the male duck's trochanter, the T2 value is typically in the range of 150-200 ms, and the T1 value is in the range of 800-1000 ms. In contrast, the T2 value of the surrounding soft tissue is 60-80 ms, and the T1 value is 400-600 ms.

[0117] Based on the relaxation characteristic analysis, this embodiment adopts a dedicated NMR signal inversion analysis method to extract the characteristic information of the syrinx structure from the complex NMR signal, including multi-component T2 inversion, quantitative mapping of magnetic susceptibility and diffusion-weighted imaging analysis, etc., to improve the level of NMR from qualitative observation to quantitative evaluation.

[0118] Specifically, a multi-component T2 inversion model based on nonnegative least squares (NNLS) was employed, with time ranges set at Tmin = 0.01 ms and Tmax = 10000 ms. Magnetization susceptibility maps were calculated using gradient echo phase information, and the apparent diffusion coefficient (ADC) of the syrinx region was measured to distinguish different tissue components. The magnetic susceptibility value of the tympanic cavity in male ducks is typically ≤ -0.02 ppm, while the corresponding region in female ducks has a magnetic susceptibility value > -0.01 ppm.

[0119] In step S32 of this embodiment, the nasal tract structure in a high-quality nasal tract MRI image is located using a trachea tracking algorithm, including:

[0120] Initialize seed point In the tracheal region at the top of the transitional duck egg;

[0121] High-quality nasal MRI images The first image is obtained by applying morphological operations. ;

[0122] Extract the skeleton from the first image For skeleton extraction Each point on Calculate its direction vector ;

[0123] From seed point Start, along the direction Track the trachea and generate the tracheal centerline trajectory. ;

[0124] Trajectory along the center line of the trachea The terminal region of the tympanic cavity is searched to find candidate regions of the tympanic tube structure that meet the conditions.

[0125] The conditions include: located on the right side of the trachea, with a roundness greater than 0.8, a signal strength ratio greater than 1.8, and a diameter in the range of 2.2~3.0 mm;

[0126] in, Represents the morphological closing operation. Represents the morphological opening operation. The size of the structure element representing the opening operation. The size of the structure element representing the closing operation. This represents the skeleton extraction function;

[0127] In this embodiment, the trachea tracking method achieves a trachea tracking accuracy of greater than 95%, and the positioning error is less than 0.8 mm, thereby providing a reliable technical means for accurate positioning and feature extraction of the syrinx structure.

[0128] In step S33 of this embodiment, on the basis of automatically identifying the syrinx structure, comprehensive feature quantification analysis is performed on the syrinx magnetic resonance image, morphological features, signal features, and texture features are measured, the consistency of the syrinx structure in consecutive slices is analyzed, and a multi-dimensional feature space is established to provide data support for gender discrimination.

[0129] Specifically, the morphological features include area, perimeter, circularity, and diameter, the signal features include signal intensity ratio, signal uniformity, and signal-to-noise ratio, and the texture features include gray level co-occurrence matrix features (energy, contrast, correlation, and entropy), local binary pattern, and multi-scale texture features of wavelet decomposition.

[0130] In this embodiment, the quantitative analysis of the features establishes a complete feature description of the syrinx structure, thereby providing comprehensive data support for subsequent gender discrimination.

[0131] In step S34 of this embodiment, based on the comprehensive feature quantification analysis, a high-dimensional feature space of the syrinx features is constructed, and through feature screening, dimension reduction, and recursive feature elimination processing, quantitative syrinx structure features are obtained, thereby improving the accuracy and efficiency of gender discrimination.

[0132] Specifically, the features are screened through feature importance calculation based on information gain and Fisher score, the features are reduced in dimension through PCA and LDA methods, the features are eliminated by determining feature weights and removing the features with the smallest weights until the target number of features is reached, and finally the final quantitative syrinx structure features are formed.

[0133] In this embodiment, through the construction and dimension reduction of the feature space, efficient expression and extraction of the syrinx features are achieved, thereby providing high-quality feature input for the subsequent gender discrimination algorithm and ensuring the accuracy and efficiency of the discrimination process.

[0134] Step S4 of the embodiment of the present application includes the following sub-steps:

[0135] S41, a syrinx gender feature discrimination standard system is established, and the discrimination standard includes the existence and position of the syrinx tympanic cavity, a syrinx tympanic cavity diameter threshold, a signal intensity ratio threshold, a T2 relaxation time threshold, a magnetic susceptibility threshold, and structure consistency in consecutive slices;

[0136] S42, a syrinx gender decision model is constructed;

[0137] The syrinx gender decision model comprises a gender discrimination layer, a confidence evaluation layer and a decision output layer; wherein the gender discrimination layer is used for pre-screening the syrinx tympanic structure and its basic characteristics based on the constructed syrinx gender characteristic discrimination standard system; then the quantized gender discrimination result is obtained through a linear discrimination function according to the extracted syrinx structure characteristics and their weights, and in combination with the physical constraints of the syrinx structure; the confidence evaluation layer is used for evaluating the accuracy of the gender discrimination result; and the decision output layer is used for outputting the final gender discrimination result according to the gender discrimination result and its confidence.

[0138] S43, input the syrinx structure characteristics into the syrinx gender decision model, and output the corresponding gender discrimination result.

[0139] In step S41 of the embodiment, based on the above syrinx structure characteristic analysis, a syrinx gender characteristic discrimination standard system is constructed, the morphological characteristics, signal characteristics and quantitative parameters are comprehensively considered, a multi-dimensional discrimination space is constructed, and accurate gender discrimination is realized; specifically, the discrimination standard includes the existence and position of the syrinx tympanic cavity, the syrinx tympanic cavity diameter threshold, the signal intensity ratio threshold, the T2 relaxation time threshold, the magnetic susceptibility threshold and the structural consistency in continuous layers; this discrimination standard system is an important link for converting the above-mentioned characteristic analysis results into actual discrimination basis, and provides a theoretical basis for the subsequent automatic discrimination algorithm.

[0140] In step S42 of the embodiment, in the gender discrimination layer, the gender discrimination function is:

[0141] ;

[0142] In the formula, represents the discrimination function value, when , the gender discrimination result is male, when , the gender discrimination result is female, represents a bias term, represents the i-th syrinx structure characteristic corresponding weight coefficient;

[0143] Wherein, the physical constraint of the gender discrimination function is:

[0144] ;

[0145] ;

[0146] In the formula, represents the actual nuclear magnetic signal intensity considering the influence of magnetic field inhomogeneity, represents the intensity of the nuclear magnetic signal with echo time TE and repetition time TR, represents the gyromagnetic ratio, denotes local magnetic field uncertainty, , denotes magnetic susceptibility difference, denotes main magnetic field strength, denotes geometric factor related to structure direction, denotes echo time, denotes repetition time, denotes relaxation time, denotes relaxation time, denotes proton density.

[0147] In the embodiment, the reliability of the gender discrimination result is evaluated based on feature consistency and image quality through the confidence evaluation layer. When multiple features point to different genders, the confidence is reduced; when the image quality is poor, it is marked as "re-detection required". The decision output layer outputs the final gender discrimination result according to the comprehensive discrimination index and the confidence threshold: when it is a certain result, "male" or "female" and the confidence percentage are output; when it is an uncertain result, "uncertain" and the suggestion of re-detection are output; when it is an abnormal result, "detection failure" and the failure reason are output.

[0148] In step S4 of the embodiment of the application, the syrinx structure in the high-quality syrinx nuclear magnetic image is displayed and optimized through a quantitative enhancement algorithm based on the gender discrimination result, thereby providing an operator with a reliable discrimination result verification means.

[0149] The enhancement algorithm includes adaptive pseudo-color mapping of the syrinx structure, syrinx structure boundary sharpening enhancement, quantitative superimposed display of the gender discrimination result, and enhancement effect quantitative evaluation.

[0150] Specifically, an adaptive pseudo-color mapping model is established based on the T2 relaxation time of the syrinx structure region:

[0151] ;

[0152] In the formula, denotes the color mapping output value of the pixel point, denotes the T2 relaxation time value of the pixel point, denotes the T2 relaxation time value of the pixel point, denotes the male-female boundary threshold value, which is determined by statistical analysis of the T2 relaxation time distribution of male and female samples, and the boundary value that gives the highest classification accuracy is selected, which is usually in the range of 100-150 ms, denotes the mapping sensitivity parameter, denotes the hyperbolic tangent function for realizing smooth color transition, denotes the color gain coefficient, Color shift coefficient, which is determined according to the display device characteristics and human visual characteristics, the male syrinx tympanic cavity area (T2 value is higher than the threshold value) is mapped to warm color display, and the female area remains original gray scale display.

[0153] An anisotropic diffusion filter based on gradient direction is used to enhance the boundary of the syrinx and the surrounding tissue structure, and the calculation formula is as follows:

[0154] ;

[0155] In the formula, , which represents the signal intensity distribution function of each pixel point in the nuclear magnetic image, , which represents the diffusion iteration time parameter, , which represents the rate of change of signal intensity with time, , which represents the spatial gradient vector of signal intensity, , which represents the amplitude of the gradient vector, , which represents the divergence operator, , which represents the diffusion coefficient function based on the gradient amplitude, wherein the diffusion coefficient function is defined as:

[0156] ;

[0157] In the formula, , which represents the gradient amplitude, , which represents the corresponding diffusion coefficient, controlling the anisotropy degree of diffusion, , which represents the diffusion control parameter, which is adaptively determined according to the gradient contrast of the syrinx structure and the surrounding tissue, the calculation method is that the ratio of the average gradient of the syrinx area to the average gradient of the background area is multiplied by an adjustment factor, and the iteration number is determined according to the image quality and the enhancement effect requirement.

[0158] The process of realizing the quantitative superimposed display of the gender discrimination result includes: automatically outlining the syrinx tympanic cavity contour, numerically displaying and labeling the syrinx diameter, signal intensity ratio and discrimination confidence, and color coding according to the confidence, including labeling with green for high confidence, labeling with yellow for medium confidence, and labeling with red for low confidence, and the threshold is determined according to the actual application requirement and the misjudgment cost.

[0159] The quantitative evaluation of the enhancement effect is realized by constructing an objective evaluation model, and the objective evaluation model is:

[0160] ;

[0161] In the formula, , which represents the comprehensive enhancement effect evaluation score, , which represents the improvement amount of the contrast noise after enhancement relative to the original image, , which represents the edge sharpening degree index calculated by the edge detection operator, represents the amount of visual contrast improvement, 、 and represents the corresponding weight coefficient, determined through user evaluation test and cross-validation, so that the evaluation score is maximally correlated with artificial visual evaluation result, and satisfies the normalization constraint of .

[0162] In the embodiment, through the quantitative enhancement algorithm described above, the visual recognition effect of the syrinx structure is significantly improved, a clear and intuitive discrimination result verification interface is provided for the operator, and the reliability and operability of the entire gender identification system are ensured. Form a complete technical chain with the foregoing syrinx detection, feature extraction, and gender discrimination algorithm, and realize the whole process optimization from image acquisition to result presentation.

[0163] In the embodiment of the application, after the technical development of the syrinx nuclear magnetic image analysis and gender discrimination is completed, it is integrated with the automatic sorting system to build a complete gender identification and sorting pipeline for duck eggs. The system includes a nuclear magnetic detection unit, an image analysis unit, a sorting control unit, and a mechanical sorting unit, realizing the whole process automation from detection to sorting. Among them, the nuclear magnetic detection unit adopts a low-field nuclear magnetic resonance system, equipped with a special imaging sequence and parameter preset, and based on the foregoing imaging technology to obtain high-quality syrinx structure images; the image analysis unit performs image processing, feature extraction, and gender discrimination, and automatically identifies the gender of duck eggs based on the foregoing discrimination algorithm; the sorting control unit receives the discrimination result and controls the action of the sorting mechanism; and the mechanical sorting unit divides the duck eggs into different collection devices according to the gender discrimination result.

[0164] The principles and implementation manners of the present application are described in the specific embodiments in the present application, and the above embodiment descriptions are only used to help understand the method of the present application and its core idea; at the same time, for those skilled in the art, according to the idea of the present application, the specific implementation manners and application ranges will be changed, and the above description should not be understood as a limitation of the present application.

[0165] Those skilled in the art will realize that the embodiments described herein are for the purpose of helping the reader to understand the principles of the present application and should be understood as not limiting the protection scope of the present application to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations according to the technical inspirations disclosed in the present application without departing from the essence of the present application, and these modifications and combinations are still within the protection scope of the present application.

Claims

1. A non-destructive method for sex determination of transgenic duck eggs based on syrinx nuclear magnetic resonance imaging, characterized in that, Includes the following steps: S1. Locate the transgenic duck egg in the nuclear magnetic resonance system and determine the optimal imaging section of the transgenic duck egg's syrinx structure; S2. Based on the optimal imaging section, several nasal MRI images are acquired by combining a dedicated MRI sequence for the nasal duct structure, and these images are preprocessed and screened to obtain high-quality nasal MRI images. In the high-quality syrinx MRI image, a bright circular structure is visible on the right side of the trachea of ​​the male duck embryo; S3. By analyzing the relaxation characteristics of the syrinx structure and inverting the NMR signal, the syrinx structure in high-quality syrinx NMR images is identified, and the structural features of the syrinx are quantitatively extracted. S4. Construct a syrinx sex decision model and, in conjunction with the physical constraints of the syrinx structure, identify the sex determination results corresponding to the syrinx structure characteristics, which will serve as a reference for sorting transitional duck eggs.

2. The non-destructive method for sex determination of transgenic duck eggs based on syrinx nuclear magnetic resonance imaging according to claim 1, characterized in that, Step S1 specifically involves: In the nuclear magnetic resonance system, the transgenic duck egg is placed at the center of the magnetic field by positioning and centering to obtain the best magnetic field uniformity and optimize the position of the radio frequency coil; Based on the location of the transitional duck eggs, several optimal imaging sections of the syrinx structure were determined through multi-planar trial scanning.

3. The non-destructive method for sex determination of transgenic duck eggs based on syrinx nuclear magnetic resonance imaging according to claim 1, characterized in that, Step S2 includes the following sub-steps: S21. A dedicated combination of magnetic resonance imaging sequences for designing syrinx structures, including T2-weighted sequences, T1-weighted sequences, magnetic susceptibility-weighted sequences, and fat suppression sequences; S22. Based on a dedicated combination of nuclear magnetic resonance imaging sequences, and on the basis of the optimal imaging section, the embryo position of the duck eggs is determined by multi-planar positioning scanning to locate the tracheal bifurcation plane. Combined with the optimal imaging time window selected at the embryonic development stage, several syrinx nuclear magnetic resonance images under the optimal imaging section of the syrinx structure are acquired. S23. Perform image enhancement processing on the acquired nasal MRI images to obtain enhanced nasal MRI images; The image enhancement process includes contrast enhancement, multi-sequence image fusion, 3D reconstruction, and parametric map generation. S24. Perform image quality optimization processing on the enhanced nasal MRI images and select high-quality nasal MRI images; The image quality optimization process includes motion correction, artifact removal, and quality evaluation.

4. The non-destructive method for sex identification of transgenic duck eggs based on syrinx nuclear magnetic resonance imaging according to claim 3, characterized in that, In step S24, an image quality evaluation model is constructed to evaluate the quality of each enhanced syrinx MRI image, and the enhanced syrinx MRI image with the highest image quality score is selected as a high-quality syrinx MRI image. When the quality evaluation score of all enhanced tympanic MRI images is less than the preset threshold, return to step S1, reposition and acquire tympanic MRI images; The expression for the image quality assessment model is: ； In the formula, Indicates the image quality score. Indicates the signal-to-noise ratio. Indicates the contrast-to-noise ratio. Indicates structural clarity. Indicates the degree of artifacts, , , and They represent , , and The weighting coefficients.

5. The non-destructive method for sex identification of transgenic duck eggs based on syrinx nuclear magnetic resonance imaging according to claim 1, characterized in that, Step S3 includes the following sub-steps: S31. Analyze the relaxation characteristics of the syrinx structure in high-quality syrinx MRI images and determine the T1 and T2 relaxation times of the syrinx structure. Based on the relaxation characteristics analysis, the susceptibility was determined by inverting the syrinx structure-specific NMR signal. S32. Locate the nasal duct structure in a high-quality nasal duct MRI image using a tracheal tracking algorithm; S33. Based on the T1 and T2 relaxation times and magnetic susceptibility of the syrinx structure, perform feature quantification analysis on high-quality syrinx MRI images to extract multidimensional features of the syrinx structure, including morphological features, signal features and texture features. S34. Based on the extracted multidimensional features, a high-dimensional feature space is constructed, and quantitative syrinx structure features are obtained through feature filtering, dimensionality reduction and recursive feature elimination.

6. The non-destructive method for sex determination of transgenic duck eggs based on syrinx nuclear magnetic resonance imaging according to claim 5, characterized in that, In step S32, the nasal tract structure in a high-quality nasal tract MRI image is located using a tracheal tracking algorithm, including: Initialize seed point In the trachea region at the top of the transitional duck egg; High-quality nasal MRI images The first image is obtained by applying morphological operations. ; Extract the skeleton from the first image For skeleton extraction Each point on Calculate its direction vector ; From seed point Begin, along the direction Track the trachea and generate the tracheal centerline trajectory. ; Trajectory along the center line of the trachea The terminal region of the tympanic cavity is searched to find candidate regions of the tympanic tube structure that meet the conditions. The conditions include: located on the right side of the end of the trachea, with a roundness greater than 0.8, a signal strength ratio greater than 1.8, and a diameter in the range of 2.2~3.0 mm; in, Represents the morphological closing operation. Represents the morphological opening operation. The size of the structure element representing the opening operation. The size of the structure element representing the closing operation. This represents the skeleton extraction function.

7. The non-destructive method for sex identification of transgenic duck eggs based on syrinx nuclear magnetic resonance imaging according to claim 1, characterized in that, Step S4 includes the following sub-steps: S41. Establish a standard system for sex identification of the tympanic tube, wherein the identification standard includes the presence and location of the tympanic tube tympanic cavity, the tympanic tube tympanic cavity diameter threshold, the signal intensity ratio threshold, the T2 relaxation time threshold, the magnetic susceptibility threshold, and the structural consistency in continuous layers. S42. Construct a syrinx gender decision-making model; The tympanic tract gender decision-making model includes a gender discrimination layer, a confidence assessment layer, and a decision output layer. The gender discrimination layer is used to pre-screen the tympanic tract and its basic features based on a constructed tympanic tract gender feature discrimination standard system. Then, based on the extracted tympanic tract structural features and their weights, combined with the physical constraints of the tympanic tract structure, a quantified gender discrimination result is obtained through a linear discriminant function. The confidence assessment layer is used to evaluate the accuracy of the gender discrimination result. The decision output layer is used to output the final gender discrimination result based on the gender discrimination result and its confidence level. S43. Input the structural features of the syrinx into the syrinx gender decision model and output the corresponding gender discrimination result.

8. The non-destructive method for sex identification of transgenic duck eggs based on syrinx nuclear magnetic resonance imaging according to claim 7, characterized in that, In step S42, the gender discrimination function in the gender discrimination layer is: ； In the formula, This represents the value of the discriminant function, when When the sex determination result is male, when At that time, the sex determination result was female. Indicates the bias term. Represents the structural features of the i-th syrinx The corresponding weighting coefficients; The physical constraints of the gender discrimination function are: ； ； In the formula, This represents the actual NMR signal intensity after considering the influence of magnetic field inhomogeneity. This indicates the intensity of an NMR signal with an echo time of TE and a repetition time of TR. Indicates the gyromagnetic ratio, This indicates local magnetic field uncertainty. , Indicates the difference in magnetic susceptibility. Indicates the degree of the main magnetic field. This represents a geometric factor related to the structural orientation. Indicates echo time. Indicates the repetition time. express Relaxation time, express Relaxation time, This represents the proton density.

9. The non-destructive method for sex determination of transgenic duck eggs based on syrinx nuclear magnetic resonance imaging according to claim 1, characterized in that, Step S4 further includes: based on the gender discrimination result, optimizing the display of the nasal structure in the high-quality nasal MRI image using a quantitative enhancement algorithm; The enhancement algorithm includes adaptive pseudo-color mapping of the syrinx structure, sharpening and enhancement of the syrinx structure boundary, quantitative overlay display of gender discrimination results, and quantitative evaluation of the enhancement effect.

Images