38 results about "Linear embedding" patented technology

The invention belongs to the technical field of power system load prediction, and relates to a long-range multivariable load prediction method and system based on time-frequency domain collaboration, and the system carries out the normalization and stabilization of a multivariate load time sequence through a data preprocessing module; the feature embedding module performs linear embedding on the block sequence to construct high-dimensional feature representation; the state space coding module extracts long-range dependency features and generates depth time sequence representation; the decoding prediction module maps the coding features into a preliminary prediction sequence; the time sequence alignment module identifies a leading-lagging relation among multiple variables and aligns a time sequence; the frequency domain optimization module realizes frequency domain component fusion based on adaptive filtering; and the model training optimization module is used for performing training and optimization through a signal attenuation loss function. The method can effectively improve the precision and robustness of long-range multivariable load prediction, and especially has obvious advantages in the aspects of processing complex dependency relationships and dynamic time delay.

The invention belongs to the technical field of multi-body dynamics analysis, and discloses a wind driven generator transmission chain rigid-flexible coupling multi-body dynamics analysis method based on dynamic mode decomposition, and the method comprises the steps: firstly, enabling multi-degree-of-freedom time series data to be non-linearly embedded into a high-dimensional feature space through an encoder neural network; extracting a dominant mode by utilizing intrinsic orthogonal decomposition (POD), and constructing a low-dimensional feature space; parameterized dynamic mode decomposition and radial basis function regression are adopted, a mapping relation between system parameters and Koopman operators is established, and accurate prediction of dynamic characteristics under variable working conditions is achieved; and finally, reconstructing a physical response through a decoder, and optimizing model parameters in combination with an error driving mechanism. The problems that a traditional method is low in calculation efficiency, poor in nonlinear adaptability and difficult in multi-parameter coupling prediction are effectively solved, the efficiency and precision of transmission chain dynamic analysis are remarkably improved, and reliable technical support is provided for state monitoring and service life prediction of the wind turbine generator.

The invention relates to a high-altitude parabolic detection method based on a Swinin-Transform and YOLOv8 fusion detection algorithm, and the method comprises the steps: decoding a video stream into an RGB image sequence, preprocessing the RGB image sequence, segmenting the preprocessed image into image blocks, and converting the image blocks into an initial feature sequence through linear embedding; the method comprises the following steps: firstly, extracting features in a layered manner by using a Transform Block of a Swin-Transform, firstly, capturing local features of a bottom layer by window self-attention, performing down-sampling after fusion, and then capturing cross-window features by moving window attention to obtain advanced features; inputting the advanced features into a YOLOv8 detection network, and obtaining a target frame and confidence through multi-scale feature fusion, anchor frame matching and prediction decoding; and finally, optimizing a detection result through non-maximum suppression and outputting the detection result. Through fusion of Swinin-Transform and YOLOv8, global context and local detail features are considered, the recognition precision of a multi-scale parabolic target is improved, the problems of missing detection, false detection and positioning deviation of a traditional method are solved, and the method can adapt to complex scenes and realize real-time detection.

The invention discloses a transformer abnormity identification method based on voiceprint feature analysis, and belongs to the field of power equipment state monitoring and intelligent diagnosis. The method comprises the following steps: firstly, analyzing an iron core acoustic mechanism based on a magnetostrictive effect, and establishing a three-dimensional model through finite element simulation to obtain vibration and sound field characteristics; in a complex substation environment, a hybrid noise reduction method combining density peak clustering and a CEEMDAN-wavelet threshold is provided, and the signal-to-noise ratio is effectively improved. Then extracting Mel-frequency cepstrum coefficients (MFCC) and spectrum features, and performing local linear embedding (LLE) dimension reduction to form a compact feature set; in the recognition stage, a convolutional neural network framework is designed, specifically, a spectrogram and an energy spectrum are modeled through a two-dimensional CNN, an MFCC tensor obtained after dimensionality reduction is modeled through a three-dimensional CNN, and accurate diagnosis of mechanical faults such as core looseness is achieved. The method has the advantages of being non-contact, anti-noise and high in recognition precision, and real-time diagnosis and early warning of mechanical abnormity of the transformer can be achieved under complex working conditions.

A method and an apparatus for forecasting a future time target variate, a computer device, and a storage medium are disclosed. Specifically, a method for forecasting a future time target variate is disclosed. The method includes: dividing a history target variate and covariates of target variates into a plurality of patches channel-wise, and performing linear embedding, wherein the covariates of target variates comprise a covariate of the history target variate and a covariate of the future time target variate; performing position embedding on a plurality of tokens and a learnable forecasting token; wherein the plurality of tokens are linear embedded from the plurality of patches; inputting the plurality of tokens and the learnable forecasting token into an encoder; and forecasting and outputting the future time target variate based on the encoder and the learnable forecasting token. According to the foregoing manner, in a case of zero-shot training, the future time target variate can be accurately forecast through limited history target variates and covariate information of the target variates, thereby having a large inspiration effect on the field of basic models in time series forecasting.

The invention discloses an automatic driving vehicle dynamics modeling method based on a time-varying parameter neural network, and belongs to the field of vehicle dynamics. Comprising the steps of obtaining operation data of a vehicle under different working conditions, performing preprocessing, and constructing a sample set according to a time sequence slicing mode; constructing a neural network model based on the physical constraints and the time-varying parameters; training and verifying the neural network model by adopting the sample set to obtain a neural network model after parameter optimization for state prediction; the neural network model comprises an input layer, a GRU network, a physical constraint layer and an output layer; input data enters the neural network model through the input layer, and dimension transformation and time sequence packaging are carried out through the linear embedding layer; then time sequence features are extracted through the GRU network, bounded mapping is carried out on output vectors of the GRU network through a physical constraint layer, a time-varying parameter estimation value is obtained, the time-varying parameter estimation value is substituted into a vehicle kinetic equation for calculation, and a physical prediction state is obtained.

The invention discloses a time series prediction method based on a multi-modal enhanced large language model, which comprises the following steps: acquiring historical time series data, constructing a standardized input matrix and setting core task parameters; dividing data blocks through a sliding window mechanism, and converting the data blocks into uniform dimension features through linear embedding; a semantic prototype is generated based on a large language model vocabulary, and time sequence features and semantic features are fused through multi-head cross-attention; alternately splicing the data blocks and the corresponding semantic information, and constructing a self-multi-modal input sequence; designing a three-level structured prompt including context, task target and modal guidance, and fusing the three-level structured prompt with a multi-modal sequence; and training the lightweight model by adopting a frozen training strategy, and outputting a prediction result of a specified time step in the future. According to the method, through single-source data enhancement and prompt guidance, the time sequence reasoning capability of a small-parameter large language model is activated, high-precision prediction is guaranteed, efficient deployment is achieved, and the method adapts to long-term and short-term prediction tasks in the fields of electric power, traffic, meteorology and the like.

The invention discloses a steel pipe weld defect detection method based on twin VMama bimodal feature fusion, which comprises the following steps: constructing an image data set and a steel pipe weld defect detection model, the model comprising twin coding, feature fusion, decoding and a classifier; the twin coding is composed of a parameter-shared double-branch coding part 1 and a coding part 2, and each coding part is composed of a block division and linear embedding module and a structure unit formed by stacking a plurality of scales layer by layer; the feature fusion part is composed of a cross VMamba feature fusion module; the decoding part is composed of a channel sensing VSS module and a block extension; the training set trains the model, and the test set monitors convergence to obtain a trained model; inputting the verification set into the trained model to obtain a defect detection result; the problems that unobvious defects and pseudo defects are difficult to detect in a single ultrasonic image or an X-ray image, the characteristics of slender and micro crack defects are difficult to detect in a CNN (Convolutional Neural Network)-based method during image processing, and the calculation complexity of a Transform-based method is high are solved.

The invention discloses a high-speed train traction system fault detection method based on local linear embedding, and belongs to the technical field of fault diagnosis. The method comprises the steps that firstly, traction system sensor data are collected and preprocessed; secondly, mapping high-dimensional data to a low-dimensional space by utilizing a local linear embedding algorithm, and reserving a local geometric structure of the data; thirdly, reconstructing original data based on the data matrix of the low-dimensional space, calculating residual signals, and extracting fault features; and finally, judging whether the high-speed train traction system breaks down or not by comparing the evaluation function with a threshold value. According to the method, the limitation of processing high-dimensional and nonlinear data by a traditional method is effectively solved, the accuracy of fault detection is improved, fault early warning of the high-speed train traction system can be realized, and the method has a relatively high engineering application value.

This invention discloses a hierarchical image classification method and system based on CA-SwinT, belonging to the field of artificial intelligence technology. The method includes: S1, acquiring the original image and preprocessing it; S2, extracting features from the image using a SwinT module; S3, fully fusing the extracted image features and classification label information using a double cross-attention mechanism to construct a fine mapping between the image and hierarchical labels; S4, generating corresponding labels based on the fine mapping in S3 using a label generation module and updating the classification label information; S2 includes: slice segmentation processing, linear embedding processing, two-level SwinT module processing, block merging processing, two-level SwinT processing, and linear embedding processing. This invention can effectively extract rich feature information from images in different dimensional spaces, and more efficiently identify multi-level labels in images.

The invention relates to a river flow prediction method, which comprises the following steps of: 1, dividing a daily runoff data set, and performing multi-stage decomposition on a training set by using Haar wavelets to obtain de-noising parameters; 2, multiplexing the verification set and the test set to obtain data; step 3, carrying out first-order differential transformation, and zooming to an interval of [0, 1] by using MinMaxScaler to obtain a normalized value of a differential sequence; 4, mapping the data to a high-dimensional space through a linear embedding layer, and then carrying out position coding to obtain data; step 5, inputting the data obtained in the step 4 into a DTransformer encoder, so that a catastrophe point obtains a higher attention weight so as to strengthen difference characteristics of adjacent time steps and obtain encoder output; 6, inputting the data obtained in the step 5 into an LSTM decoder, and performing feature recombination to obtain encoder output; 7, mapping the data obtained in the step 6 to 7-dimensional output through a full connection layer; and 8, performing inverse normalization and inverse difference transformation on the data obtained in the step 7 to obtain final output.

The application belongs to the technical field of image processing, and discloses a chest and abdominal cavity bleeding automatic segmentation method and device based on deep learning, which comprises the following steps: obtaining and normalizing chest cavity or abdominal cavity CT volume data, and then performing linear embedding to generate initial features; inputting the initial features into an encoder for multi-stage encoding, in which the encoding features of different levels are respectively enhanced in the frequency domain according to different semantic levels to generate level-aware frequency domain enhanced features; the input features are modeled by a deformable mixed window multi-head self-attention mechanism in at least one stage of the encoder, and the attention calculation results are weighted and fused by using the frequency domain guide weight generated based on the frequency domain enhancement results; the encoding features are input into a decoder for multi-stage decoding and fusion with the corresponding level features of the encoder to gradually restore the spatial resolution; and the bleeding area segmentation results corresponding to the CT volume data are generated according to the final output features of the decoder. The application can efficiently and accurately segment the bleeding area and assist clinical decision-making.

The invention discloses a dual-path Transform-based sky and ground wave identification method in a Rowland system. The method comprises the following steps: performing learnable discrete wavelet decomposition on a preprocessed signal received by a Rowland receiver to obtain a low-frequency component and a high-frequency component; projecting the components to a potential embedding space through independent linear embedding layers, and obtaining an embedding vector with position information in combination with position coding; inputting the embedded vector into a double-path Transform encoder, and respectively extracting low-frequency and high-frequency characteristics; the encoded features are reconstructed and fused through learnable inverse wavelet transform, and unified global representation is obtained; inputting the global representation into two layers of stacked Encoder for further optimization to obtain deep features; and finally, inputting the deep features into a multi-layer perceptron to realize sky and ground wave recognition. According to the invention, by introducing learnable wavelet decomposition and reconstruction, a dual-path encoder structure and a sparse attention mechanism, the sky and ground wave signals are effectively separated and fused, so that the success rate of ground wave recognition is remarkably improved.

The present disclosure provides a training method of an image extraction model and an image extraction method, which can be applied to the fields of image processing and pattern recognition. The image extraction model comprises a linear embedding network, an encoding network and a decoding network. The training method comprises: processing a sample image obtained by using the linear embedding network to obtain an embedding feature map; processing the embedding feature map by using the encoding network to obtain i encoding feature maps; processing the i encoding feature maps by using the decoding network to obtain a weight fusion feature map, wherein the decoding sub-network is constructed based on a local-global attention layer and a weight feature fusion layer; performing segmentation head mapping processing on the weight fusion feature map to obtain an image segmentation result, wherein the image segmentation result represents a geological change attribute of a target geographical environment region; and training the image extraction model according to the image segmentation result and label data corresponding to the image segmentation result to obtain a trained image extraction model.

The present invention relates to the technical field of artificial intelligence, and disclosed are a CA-SwinT-based hierarchical image classification method and system. The method comprises: S1, acquiring an original image, and preprocessing the original image; S2, performing feature extraction on the image by means of a SwinT module; S3, using a dual cross-attention mechanism to fully fuse extracted image features and classification label information so as to construct fine mapping between the image and a hierarchical label; and S4, on the basis of the fine mapping in S3, generating a corresponding label by means of a label generation module, and updating the classification label information. S2 comprises: patch-based processing, linear embedding processing, two-stage Swin module processing, patch merging processing, two-stage Swin processing, and linear embedding processing. In the present invention, rich feature information of an image can be effectively extracted from spaces of different dimensions, thereby identifying a multi-level label of the image more efficiently.

The invention provides a chromosome classification method based on sequence cross-correlation. The method comprises the following steps: acquiring all chromosome images in a whole chromosome cell map; the image is filled to a multiple of 8 and then is cut into 8 * 8 image expansion data blocks, and position information is generated; splicing image extended data blocks into a sequence, and inputting the sequence into a feature encoder containing global and local self-attention modules after linear embedding and position information embedding; output features are pooled and then classified through a classification head, 24 types of probability distributions are obtained through softmax, and the highest probability index is taken as the category. According to the method, chromosome size information is reserved, the overall relation between single chromosome details and whole-graph chromosomes is considered, and generalization is better under the scenes of chromosome crossing, adhesion, slide production difference and the like.

The invention discloses a three-dimensional human body posture estimation method and system based on space-time semantic adaptive learning, computer equipment and a storage medium, and the method comprises the steps: S1, inputting a video frame sequence, obtaining two-dimensional coordinates of human body key points in each frame of a 2D video through a two-dimensional posture extractor, and generating a 2D skeleton; s2, performing linear embedding, spatial embedding and time embedding on the extracted 2D attitude through an LST module; s3, through a double attention mechanism embedded in the DAS-TF module, further capturing a spatial dependency relationship and a time dynamic change in the 2D attitude data, obtaining an enhanced spatial-temporal feature representation, and outputting three-dimensional coordinates of key points of a human body; and S4, after the first round of parameter updating is completed, performing a second round of parameter optimization through a BV-Loss module, and outputting an optimized three-dimensional human body posture estimation result. According to the method, the precision and robustness of three-dimensional human body posture estimation are improved, and the method belongs to the field of vision and the field of deep learning.

A maskless Transform framework-based intelligent identification method for operation events of a small nuclear power system comprises the following steps of: 1, acquiring multi-source heterogeneous time sequence data related to an operation state by using a sensor system or a simulation platform carried by the small nuclear power system to form an original data set; 2, preprocessing the original data set; 3, creating a linear embedded layer network, and mapping a multi-dimensional parameter vector in the sample sequence to a high-dimensional feature space by using the network to form an event data embedded vector; 4, performing position coding on the event data embedding vector, and adding time sequence information in the sample sequence into the event data embedding vector; (5) a maskless Transform encoder network formed by stacking N identical encoder layers is built, and N is a positive integer greater than or equal to 1, and N is a positive integer greater than or equal to 1; 6, the event data embedding vector is input into a maskless Transform encoder network, and a feature sequence is output; and 7, calculating the probability that the sample sequence belongs to each operation event category through a Softmax function, and judging the category with the maximum probability as a final recognition result. The method is high in recognition accuracy and strong in time sequence characteristic capturing capability.

The application belongs to the technical field of multi-body dynamics analysis, and discloses a wind driven generator transmission chain rigid-flexible coupling multi-body dynamics analysis method based on dynamic mode decomposition, which firstly embeds multi-degree-of-freedom time series data into a high-dimensional feature space through an encoder neural network; extracts dominant modes by using proper orthogonal decomposition (POD), and constructs a low-dimensional feature space; adopts parameterized dynamic mode decomposition and radial basis function regression to establish the mapping relationship between system parameters and Koopman operators, and realizes accurate prediction of dynamic characteristics under variable working conditions; and finally reconstructs physical responses through a decoder, and optimizes model parameters in combination with an error driving mechanism. The application effectively solves the problems of low calculation efficiency, poor nonlinear adaptability and difficult multi-parameter coupling prediction of traditional methods, significantly improves the efficiency and accuracy of transmission chain dynamics analysis, and provides reliable technical support for wind turbine condition monitoring and life prediction.

This invention provides a method, apparatus, and device for identifying grain origins based on hyperspectral images, relating to the field of image recognition technology. The method includes: acquiring hyperspectral sample images of grains from different origins using a hyperspectral imaging device; constructing a pseudo-RGB image of each hyperspectral sample image by selecting three representative bands; performing grain instance-level segmentation on each pseudo-RGB image to obtain a corresponding segmentation mask; extracting the average spectral information of each grain using the segmentation mask and constructing a spectral information matrix; sequentially performing principal component analysis, local linear embedding, and sparse subspace clustering on the spectral information matrix to obtain a dimensionality reduction matrix; training a fully connected neural network model based on the dimensionality reduction matrix to obtain a grain origin traceability model, and then performing grain origin traceability based on the grain origin traceability model. This invention can significantly improve the accuracy and processing efficiency of grain origin identification.

The invention discloses an intelligent camera image recognition method based on Transform, and the method comprises the following steps: obtaining image data collected by an intelligent camera, and carrying out the preprocessing of the image data; image blocks are divided, linear embedding is carried out, and position codes are added; inputting a multi-axis self-attention module of the improved MaxViT, and carrying out local feature and global sparse feature modeling; inputting a feature pyramid structure to perform multi-scale feature fusion, interaction and grading loss application; parameter pruning, weight quantization and sparse attention calculation are carried out; the input channel reweighting module is used for carrying out channel weight distribution and carrying out feature mapping activation through a lightweight multilayer perceptron and a SiLU function; label information is generated and stored in the intelligent camera. According to the method, improved MaxViT and lightweight optimization are adopted, intelligent camera image recognition is achieved, and the method has the advantages of being high in precision, low in delay and low in energy consumption.

The application discloses a kind of based on improved attention mechanism's fitting classification detection method, improved Swin Transformer architecture is selected as main network;Introduce the framework module based on the convolution of self-attention, utilize shallow convolution network and mobile window multi-head attention mechanism, the region containing rich fitting feature information is extracted to improve the accuracy of target classification;Among them, the improved Swin Transformer network includes two independent regions, i.e. the convolution based on self-attention and Swin Transformer.The convolution based on self-attention utilizes shallow convolution network and mobile window multi-head attention mechanism, and extracts the region containing rich fitting specific information.Subsequently, the features obtained based on the convolution of self-attention are spliced with the original image processed by linear embedding in Swin Transformer, and then input into Swin Transformer to identify the type of fitting, to more accurately identify fitting.

The invention discloses a rail transit engineering video image-oriented airspace self-supervised learning method and system. The method comprises the steps of completing video image preprocessing; generating a space reduction window set SRW according to a preset step length and a preset window size after the image is partitioned, mixing the image blocks in each window, and recording mask positions; extracting blocks with a preset proportion from the sampled blocks, and randomly adding degradation features; a multi-scale sliding converter comprising a block segmentation module, a linear embedding module, a multi-scale feature fusion module and the like is adopted as an encoder, and a multi-scale feature extraction problem of a model is converted into a channel feature self-attention fusion problem at different depths through a multi-scale feature fusion module to realize multi-scale feature extraction. Resolution is recovered and an image is reconstructed through pixel recombination and an MAE decoder, and training is completed by using a stochastic gradient descent optimizer in combination with reconstruction loss and cross-block loss. And the model can learn potential safety hazard characteristics without additional supervision.

The invention discloses a partial discharge fault diagnosis method based on multi-scale convolution, and aims to improve the diagnosis precision of a partial discharge fault of power equipment. The method comprises the following steps: reading original binary partial discharge time sequence PRPS signal data, and carrying out preprocessing; the method comprises the following steps: extracting local features based on multi-scale one-dimensional convolution, generating sequence input by combining multi-scale Patch division and linear embedding, carrying out sequence modeling by utilizing a multi-layer Transform encoder, and realizing fault classification through attention pooling and a multi-layer full-connection classifier; for a single normalized input signal, convolution feature extraction, multi-scale embedding, Transform coding, attention aggregation and classification prediction, and fault category and corresponding probability output are performed in sequence. According to the scheme, the accuracy and the automation level of partial discharge fault diagnosis can be improved, and the method is suitable for scenes such as state evaluation and fault early warning of power equipment.

This application discloses a modulation identification and symbol rate estimation method, apparatus, storage medium, and device, belonging to the field of communication technology. It converts the radio frequency signal transmitted through a wireless channel, separates the real and imaginary parts of the baseband modulation signal, and concatenates the real and imaginary parts into a two-dimensional matrix; it obtains a multi-task model, which includes a convolution module, a pooling module, a DRA module, an SST module, and a fully connected module. The DRA module consists of multiple DA units and a jump-connected structure with ECA units; the SST module consists of Patch Partition units, linear embedding units, and Swing TransformerBlock units; the multi-task model is used to process the two-dimensional matrix to obtain the modulation scheme and symbol rate. This application can simultaneously perform modulation identification and symbol rate estimation, exhibiting good performance, low complexity and parameter quantity, low overhead, fast derivation speed, and high generalization ability and robustness.

The application belongs to the field of magnetic resonance sounding signal noise filtering, and is a kind of magnetic resonance sounding signal extraction method based on intelligent optimization manifold learning, the parameter group of the local linear embedding manifold learning method is initialized, the genetic algorithm in the intelligent optimization algorithm is used, the signal-to-noise ratio is taken as the fitness function, and the parameter group in the local linear embedding manifold learning method is optimized, the local linear embedding manifold learning method uses the optimized parameter group to sequentially perform first processing and second processing on the magnetic resonance sounding signal, removes random noise, and obtains the final denoised magnetic resonance sounding signal. The application effectively retains the signal characteristics through the nonlinear dimension reduction of manifold learning, avoids the information loss caused by frequency band selection, maintains the local relationship between data points in the dimension reduction process, ensures that adjacent points in m-dimensional space remain adjacent in d-dimensional space, only compresses and filters noise, and realizes signal extraction.

The application discloses a kind of space self-supervised learning method and system for rail transit engineering video image, method includes: completing video image preprocessing;After image block is generated with preset step and window size to generate spatial reduction window set SRW, the mask position is recorded by mixing and adopting the block in each window;From the block after sampling, extract preset proportion block and randomly add degradation feature;Using the multiscale sliding transformer of block segmentation, linear embedding, multiscale feature fusion and other modules as encoder, the multiscale feature extraction problem of model is converted into channel feature self-attention fusion problem at different depths by multiscale feature fusion module to realize multiscale feature extraction, restore resolution and reconstruct image by pixel reorganization and MAE decoder, combined with reconstruction loss and cross-block loss, complete training with stochastic gradient descent optimizer.Model learns potential safety hazard features without additional supervision.

The present application relates to the technical field of power grid super-harmonic, and specifically discloses a power grid super-harmonic online monitoring and identification method and system, which utilizes a preset learning type measurement matrix to perform non-uniform compression sampling on original signals to be measured of a power grid, so as to obtain a low-dimensional observation vector sequence; the sequence is subjected to linear embedding processing and superposition position coding, so as to generate a feature vector containing time sequence information; the feature vector is input into a pre-trained Transformer reconstruction model, a high-dimensional time domain waveform after restoration is output through a decoding layer of the Transformer reconstruction model, so as to realize reconstruction of a super-harmonic signal; the reconstructed waveform is subjected to feature analysis, and monitoring parameters of the power grid super-harmonic or a harmonic source identification result is output. The present application can obtain a high-frequency signal under the condition of lower sampling rate of hardware, can reduce data transmission amount, and thus effectively reduces system cost; compared with a reconstruction algorithm of traditional compression sensing which depends on iterative solution, the present application only needs one-time forward calculation to complete signal reconstruction, and response speed is improved.

The invention provides a grain producing area identification method, device and equipment based on a hyperspectral image, and relates to the technical field of image identification. The method comprises the following steps: acquiring hyperspectral sample images of grains from different producing areas by using hyperspectral imaging equipment; for each hyperspectral sample image, three representative wavebands are selected to construct a pseudo RGB image of the hyperspectral sample image; performing grain instance level segmentation on each pseudo RGB image to obtain a corresponding segmentation mask; extracting average spectral information of each grain by using the segmentation mask and constructing a spectral information matrix; sequentially performing principal component analysis, local linear embedding and sparse subspace clustering on the spectral information matrix to obtain a dimension reduction matrix; and training a full-connection neural network model based on the dimensionality reduction matrix to obtain a grain production place traceability model, and performing grain production place traceability based on the grain production place traceability model. According to the invention, the accuracy and processing efficiency of grain production area identification can be significantly improved.