Rice saline-alkaline tolerance prediction method and system based on big data analysis

Abstract

The invention relates to the technical field of crop saline-alkaline tolerance analysis, in particular to a rice saline-alkaline tolerance prediction method and system based on big data analysis, and the method comprises the following steps: collecting soil environment data, rice physiological phenotype data and historical stress response data, and constructing a multi-dimensional data cube by using a space-time alignment engine; quantization feature analysis is adopted to carry out quantum probability modeling on Na < + > / K < + > ion transmembrane transport and PSII photosynthetic exciton transfer features, and a quantum probability cloud chart of rice physiological response is generated. Based on a geodesic line domain adaptation model of geodesic line mapping, a soil conductivity thermodynamic diagram and a quantum probability cloud chart are fused, a rice saline-alkaline tolerance adaptability prediction model is established, finally, a three-dimensional saline-alkaline tolerance performance prediction map is output, and the adaptability of rice varieties in specific saline-alkaline land is visually displayed. According to the method, the regional generalization ability of saline-alkali tolerant varieties is effectively improved, the limitation of a traditional statistical model is broken through, and a scientific basis is provided for saline-alkali tolerant rice breeding and cultivation and saline-alkali soil agricultural optimization management.

Description

Technical Field

[0001] The present invention relates to the technical field of crop salt and alkali tolerance analysis, and particularly relates to a method and system for predicting the salt and alkali tolerance performance of rice based on big data analysis. Background Art

[0002] Rice is one of the food crops, and its yield and quality are significantly affected by the soil saline-alkali environment. Under saline-alkali stress conditions, rice plants are prone to problems such as osmotic imbalance, ion toxicity, and decreased photosynthesis efficiency, resulting in growth inhibition and ultimately affecting yield and quality.

[0003] The current methods for evaluating the salt and alkali tolerance performance of rice mainly have the following technical problems: Traditional methods mainly rely on single-point measurements and cannot effectively integrate unmanned aerial vehicle (UAV) remote sensing, groundwater monitoring, and phenomics data, resulting in large spatio-temporal alignment errors and affecting the model accuracy. Existing methods mostly use statistical analysis or simple regression models and are difficult to accurately analyze physiological processes such as Na+ / K+ ion transmembrane transport and PSII photochemical efficiency, which limits the in-depth understanding of the salt and alkali tolerance mechanism. Moreover, traditional variety evaluations are mostly based on experimental station tests and are difficult to be extended to different ecological regions. Summary of the Invention

[0004] The present invention provides a method and system for predicting the salt and alkali tolerance performance of rice based on big data analysis, which realizes accurate prediction of the salt and alkali tolerance of rice and regional adaptability evaluation through spatio-temporally consistent data fusion, quantization of physiological mechanism modeling, and geodesic mapping domain adaptation model.

[0005] A method for predicting the salt and alkali tolerance performance of rice based on big data analysis includes the following steps: S1, Multi-source heterogeneous data collection: Obtain multi-source heterogeneous data through a multi-source heterogeneous data collection module. The multi-source heterogeneous data includes dynamic monitoring data of saline-alkali land environment, rice phenomics data, and historical stress response data; S2, Spatio-temporal alignment and feature matrix construction: Use a spatio-temporal alignment engine to perform frequency-domain - time-domain joint calibration on the obtained multi-source heterogeneous data, and eliminate the millisecond-level time series deviation between UAV hyperspectral sampling and groundwater level monitoring through the dynamic time warping algorithm to generate a multi-dimensional feature matrix with a unified spatio-temporal coordinate system; S3, Physiological response feature extraction: Construct a dynamic coupling model of saline-alkali gradient - physiological response, input the generated multi-dimensional feature matrix into a bidirectional LSTM network with an attention mechanism based on a gated recurrent unit, extract the correlation features of Na+ / K+ ion transmembrane transport and exciton transfer in the PSII reaction center of chloroplasts, and output a spatio-temporal coupling feature vector set; S4. Quantum Feature Analysis: The output spatio-temporal coupling eigenvector set is subjected to Hilbert space reconstruction by the quantum feature analysis module, the tunneling probability distribution of the proton pump dynamics of root system cells is calculated, and a quantum probability cloud map containing the ion channel state transition path is generated. S5. Regional Adaptability Prediction: Based on the regional adaptability prediction module of transfer learning, the generated quantum probability cloud map and the soil conductivity heat map of the target area are subjected to multi-scale convolution operation, and a geodesic domain adaptation model of the saline-alkali stress feature space and regional soil parameters is constructed through the geodesic mapping algorithm, and a three-dimensional saline-alkali tolerance performance prediction map containing the variety adaptability score is output.

[0006] Optionally, S1 includes: S11. Collection of Dynamic Monitoring Data of Saline-Alkali Land Environment: The dynamic monitoring data of the saline-alkali land environment are collected through the multi-source heterogeneous data collection module, and the dynamic monitoring data of the saline-alkali land environment include the soil surface reflection spectrum sequence and the groundwater depth-salinity time series data. S12. Collection of Rice Phenomics Data: The rice phenomics data are collected through the multi-source heterogeneous data collection module, and the rice phenomics data include the canopy chlorophyll content matrix, the dynamic change sequence of plant height, the stomatal conductance time series data, and the root ion flow rate matrix. S13. Collection of Historical Stress Response Data: The historical stress response data are collected through the multi-source heterogeneous data collection module, and the historical stress response data include the saline-alkali stress event record table, the ion transport efficiency mapping table, and the PSII photochemical efficiency grade table. S14. Spatio-Temporal Benchmark Alignment of Multi-Source Data: The collected dynamic monitoring data of the saline-alkali land environment, rice phenomics data, and historical stress response data are subjected to spatio-temporal benchmark alignment to generate a spatio-temporally aligned data stream. S15. Standardization of Heterogeneous Data and Generation of Structured Cube: The spatio-temporally aligned data stream is converted into a standardized multi-dimensional data cube to generate a structured data cube, and the structured data cube includes the following independent data channels: Soil Parameter Layer: It includes the soil surface salt concentration data and the groundwater salinity data. Phenotype Parameter Layer: It includes the temperature-compensated stomatal conductance data and the standardized root ion flow rate data. Historical Parameter Layer: It includes the PSII stress level data.

[0007] Optionally, S2 includes: S21. Joint Frequency Domain-Time Domain Calibration of Multi-Source Data: The structured data cube is subjected to joint frequency domain-time domain calibration through the spatio-temporal alignment engine to generate a synchronously calibrated data stream, and the synchronously calibrated data stream includes a synchronously time domain data stream and a frequency domain coupling eigenvector. S22, Optimization of the dynamic time warping algorithm: Apply the dynamic time warping algorithm to the synchronous time-domain data stream in the synchronous calibration data stream to generate a dynamically time-warped optimized data stream; S23, Spatiotemporal coupling coding generation: Based on the generated dynamically time-warped optimized data stream, combine with the frequency-domain coupling eigenvector to generate a spatiotemporal coupling coefficient matrix.

[0008] Optionally, S2 further includes: S24, Multidimensional feature matrix generation: Integrate the spatiotemporal coupling coefficient matrix generated in S23 with the phenotypic parameter layer and historical parameter layer in S15 to generate a multidimensional feature matrix; S25, Data quality verification: Perform integrity verification on the generated multidimensional feature matrix to generate a verified feature matrix.

[0009] Optionally, S3 includes: S31, Construction of the salinity gradient - physiological response dynamic coupling model: Based on the output verified feature matrix, construct a salinity gradient - physiological response dynamic coupling model to generate a dynamic coupling model architecture; S32, Spatiotemporal feature joint training and optimization: Perform spatiotemporal feature joint training on the dynamic coupling model to generate a trained coupling model; S33, Na+ / K+ ion transmembrane transport feature extraction: Extract the Na+ / K+ ion transmembrane transport features through the trained coupling model to generate an ion transport correlation eigenvector.

[0010] Optionally, S3 further includes: S34, Chloroplast PSII exciton transfer correlation feature extraction: Extract the exciton transfer correlation features based on the PSII stress level data to generate a PSII exciton transfer eigenvector; S35, Spatiotemporal coupling eigenvector set generation: Integrate the ion transport correlation eigenvector and the PSII exciton transfer eigenvector to generate a spatiotemporal coupling eigenvector set.

[0011] Optionally, S4 includes: S41, Hilbert space reconstruction: Perform Hilbert space reconstruction on the spatiotemporal coupling eigenvector set through the quantization feature analysis module to generate a quantum state vector space; S42, Root cell proton pump dynamics modeling: Calculate the root cell proton pump dynamics parameters based on the quantum state vector space to generate a proton pump tunneling probability model; S43, Ion channel state transition path analysis: Combine with the generated ion transport correlation eigenvector to analyze the ion channel state transition path to generate a state transition Markov chain; S44, Quantum Probability Cloud Map Generation: Integrate the proton pump tunneling probability model and the state transition Markov chain to generate a quantum probability cloud map; S45, Quantum Analysis Result Encapsulation: Standardize and encapsulate the quantum probability cloud map to generate an ion channel state transition quantum probability cloud map.

[0012] Optionally, the S5 includes: S51, Target Area Soil Conductivity Heat Map Generation: Based on the soil surface salt concentration data and the spatio-temporal coupling coefficient matrix, generate a target area soil conductivity heat map, and perform Gaussian pyramid decomposition on the target area soil conductivity heat map to generate a multi-scale soil conductivity feature tensor; S52, Quantum-Soil Multi-Scale Convolution Operation: Integrate the generated ion channel state transition quantum probability cloud map (in HDF5 format) with the multi-scale soil conductivity feature tensor to generate a cross-domain fusion feature map; S53, Geodesic Domain Adaptation Model Construction: Based on the cross-domain fusion feature map, construct a domain adaptation model for saline-alkali stress feature space and regional soil parameters to generate a geodesic domain adaptation model.

[0013] Optionally, the S5 further includes: S54, Rice Variety Adaptability Score Calculation: Predict the saline-alkali tolerance performance of rice varieties through the geodesic domain adaptation model to generate an adaptability score matrix; S55, Three-Dimensional Saline-Alkali Tolerance Performance Prediction Atlas Generation: Encapsulate the adaptability score matrix into a visualization atlas to generate a three-dimensional saline-alkali tolerance performance prediction atlas.

[0014] A rice saline-alkali tolerance performance prediction system based on big data analysis, used to implement the above-mentioned rice saline-alkali tolerance performance prediction method based on big data analysis, includes the following modules: Multi-Source Heterogeneous Data Acquisition Module: Used to collect multi-source heterogeneous data, including saline-alkali land environment dynamic monitoring data, rice phenomics data, and historical stress response data, and record the spatio-temporal coordinates of the data through the GNSS positioning unit to achieve data positioning and alignment; Spatio-Temporal Alignment and Feature Matrix Construction Module: Used to perform frequency-domain and time-domain joint calibration on the collected multi-source heterogeneous data, eliminate the time deviation of data collection through the dynamic time warping algorithm, and construct a multi-dimensional feature matrix with a unified spatio-temporal coordinate system; Physiological Response Feature Extraction Module: Used to establish a saline-alkali gradient-physiological response dynamic coupling model, extract rice saline-alkali tolerance physiological response features, including Na+ / K+ ion transmembrane transport features and PSII exciton transfer efficiency, and generate a spatio-temporal coupling feature vector set; Quantization Feature Analysis Module: It is used to perform Hilbert space reconstruction on the generated spatio-temporal coupling feature vector set, calculate the proton pump kinetic parameters of root system cells, and generate a quantum probability cloud map by combining the ion channel state transition path analysis; Regional Adaptability Prediction Module: It is used to fuse the generated quantum probability cloud map and the collected soil surface salt concentration data, analyze the matching degree between the saline-alkali stress feature space and the regional soil parameters through the geodesic domain adaptation model, calculate the adaptability score of rice varieties, and finally generate a three-dimensional saline-alkali tolerance performance prediction map; Data Storage and Visualization Module: It is used to store the generated three-dimensional saline-alkali tolerance performance prediction map and support the visual display of data.

[0015] Advantages of the present invention: In the present invention, through means such as unmanned aerial vehicle multispectral remote sensing, ground hyperspectral imaging, groundwater monitoring, and plant physiological measurement, saline-alkali land environmental data, rice phenotype data, and historical stress response data are collected, and the dynamic time warping algorithm and geodesic mapping technology are used to achieve millimeter-level spatio-temporal alignment and construct a standardized multi-dimensional data cube. Compared with traditional methods, this method reduces the spatio-temporal error between different data sources, ensures the spatial consistency and temporal comparability of data, and provides high-quality data support for subsequent prediction modeling.

[0016] In the present invention, Hilbert space reconstruction is used to perform quantization analysis on spatio-temporal coupling features, construct a tunneling model of proton pumps in root system cells, calculate the ion transmembrane transport probability, and combine the ion channel state transition Markov chain to generate a quantum probability cloud map. This method can more accurately analyze the dynamic changes of Na+ / K+ ion flow rate and PSII photochemical efficiency, reveal the cell-level adaptation mechanism of rice under saline-alkali stress, and can better simulate the nonlinear behavior of biological systems compared with traditional statistical models, improving the prediction accuracy of rice physiological characteristics.

[0017] In the present invention, through cross-domain fusion feature extraction, multi-scale convolution of the quantum probability cloud map and the soil conductivity heat map is performed, and a regional generalization prediction system for rice saline-alkali tolerance performance is established using the geodesic mapping domain adaptation model (based on Riemannian manifold projection and gradient reversal adversarial training). By optimizing the domain adaptation loss, this model can transfer the saline-alkali tolerance characteristics of existing varieties to new planting areas, improving the accuracy of variety adaptability prediction. The finally generated three-dimensional saline-alkali tolerance performance prediction map visually displays the adaptability of different rice varieties in specific saline-alkali lands, guiding precise seed selection and optimizing cultivation management strategies, and improving the popularization efficiency of saline-alkali tolerant rice varieties. Description of the Drawings

[0018] To more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only those of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0019] Figure 1 Schematic diagram of the method flow of the embodiment of the present invention; Figure 2 Schematic diagram of the system flow of the embodiment of the present invention. Detailed implementation manners

[0020] The present invention will be described in detail below in conjunction with the drawings and specific embodiments. At the same time, it should be noted here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments. For some well-known technologies, those skilled in the art can also adopt other alternative methods for implementation; moreover, the drawing part is only for more specific description of the embodiments, and is not intended to specifically limit the present invention.

[0021] It should be pointed out that in the specification, terms such as "one embodiment", "embodiment", "exemplary embodiment", "some embodiments", etc. indicate that the described embodiments may include specific features, structures or characteristics, but not necessarily every embodiment includes the specific feature, structure or characteristic. In addition, when combining embodiments to describe specific features, structures or characteristics, implementing such features, structures or characteristics in combination with other embodiments (whether explicitly described or not) should be within the knowledge of those skilled in the relevant art.

[0022] Generally, terms can be understood at least in part from their use in the context. For example, at least in part depending on the context, the term "one or more" used herein can be used to describe any feature, structure or characteristic in a singular sense, or can be used to describe a combination of features, structures or characteristics in a plural sense. In addition, the term "based on" can be understood as not necessarily intended to convey a set of exclusive factors, but instead, at least in part depending on the context, allowing for the existence of other factors that may not be explicitly described.

[0023] As Figure 1 shown, a method for predicting the saline-alkali tolerance performance of rice based on big data analysis includes the following steps: S1, Multi-source heterogeneous data collection: Obtain multi-source heterogeneous data through a multi-source heterogeneous data collection module. The multi-source heterogeneous data includes dynamic monitoring data of saline-alkali land environment, rice phenomics data, and historical stress response data; S2, Spatiotemporal Alignment and Feature Matrix Construction: Use a spatiotemporal alignment engine to perform frequency-domain and time-domain joint calibration on the acquired multi-source heterogeneous data. Eliminate the millisecond-level timing deviation between UAV hyperspectral sampling and groundwater level monitoring through the dynamic time warping algorithm, and generate a multi-dimensional feature matrix with a unified spatiotemporal coordinate system; S3, Physiological Response Feature Extraction: Construct a dynamic coupling model of salinity gradient - physiological response. Input the generated multi-dimensional feature matrix into a bidirectional LSTM network with an attention mechanism based on gated recurrent units to extract the correlation features of Na+ / K+ ion transmembrane transport and exciton transfer in the PSII reaction center of chloroplasts, and output a set of spatiotemporal coupling feature vectors; S4, Quantum Feature Analysis: Reconstruct the output set of spatiotemporal coupling feature vectors in Hilbert space through a quantum feature analysis module, calculate the tunneling probability distribution of root cell proton pump dynamics, and generate a quantum probability cloud map containing the ion channel state transition path; S5, Regional Adaptability Prediction: Based on a regional adaptability prediction module using transfer learning, perform multi-scale convolution operations on the generated quantum probability cloud map and the soil conductivity heat map of the target area. Construct a geodesic domain adaptation model between the saline-alkali stress feature space and regional soil parameters through the geodesic mapping algorithm, and output a three-dimensional saline-alkali tolerance performance prediction map containing variety adaptability scores.

[0024] S1 includes: S11, Collection of Dynamic Monitoring Data of Saline-Alkali Land Environment: Collect dynamic monitoring data of the saline-alkali land environment through a multi-source heterogeneous data collection module. The dynamic monitoring data of the saline-alkali land environment includes soil surface reflection spectral sequences and groundwater depth - salinity time series data. The specific collection methods are as follows: Generation of soil surface reflection spectral sequences: Use an airborne multi-spectral sensor array on a UAV (wavelength range 400 - 2500nm, resolution 0.5m) to conduct aerial photography scans of the target plot at a sampling frequency of 10 frames per second, and synchronously trigger the GNSS positioning unit (accuracy ±2cm) to record spatial coordinates, generating soil surface reflection spectral sequences (data format: three-dimensional tensor [wavelength × spatial coordinates × timestamp]); Generation of groundwater depth - salinity time series data: Deploy a groundwater level sensor network (sampling frequency 10Hz) in monitoring wells. Use a capacitive pressure sensor (accuracy ±0.1m) to measure the groundwater depth, and at the same time obtain groundwater salinity data through a four-electrode conductivity sensor (range 0 - 50dS / m, resolution 0.1dS / m), generating groundwater depth - salinity time series data (data format: timestamp - depth - salinity triple); S12, Rice phenomics data collection: Collect rice phenomics data through a multi-source heterogeneous data collection module. The rice phenomics data includes a canopy chlorophyll content matrix, a dynamic plant height change sequence, stomatal conductance time-series data, and a root ion flux rate matrix. The specific collection methods are as follows: Generation of canopy chlorophyll content matrix: Use a ground mobile phenotyping platform equipped with a hyperspectral imager (256 spectral channels) to scan the rice canopy in a vertical incidence manner. Invert the chlorophyll content (SPAD value) based on the NDVI705 narrowband vegetation index to generate a canopy chlorophyll content matrix (spatial resolution 1 cm², data format: two-dimensional grid [row × column]); Generation of dynamic plant height change sequence: Reconstruct the three-dimensional point cloud of rice plants through lidar (scanning accuracy ±1 mm), and use the voxelization difference algorithm to calculate the daily incremental plant height change to generate a dynamic plant height change sequence (data format: timestamp-plant height key-value pair); Generation of stomatal conductance time-series data: Use an infrared gas analyzer (IRGA, accuracy ±0.1 mmol / m² / s) to measure the stomatal conductance of leaves, and synchronously record the leaf surface temperature (±0.5 °C) to generate stomatal conductance time-series data (data format: timestamp-stomatal conductance-temperature triple); Generation of root ion flux rate matrix: implant a non-invasive micro-test technology ion microelectrode array (detection limit 0.1 μmol / L) in the root zone, and continuously measure the transmembrane flux of Na+ and K+ ions at 10-second intervals to generate a root ion flux rate matrix (data format: timestamp-ion type-flux three-dimensional tensor); S13, Historical stress response data collection: Collect historical stress response data through a multi-source heterogeneous data collection module. The historical stress response data includes a saline-alkali stress event record table, an ion transport efficiency mapping table, and a PSII photochemical efficiency grade table. The specific collection methods are as follows: Generation of saline-alkali stress event record table: Extract the saline-alkali stress event records of the target area in the past 10 years from the agricultural meteorological database, including the soil electrical conductivity (number of consecutive days with EC value > 4 dS / m) and the corresponding rice variety numbers, to generate a saline-alkali stress event record table (data format: timestamp-EC value-variety number triple); Generation of ion transport efficiency mapping table: Retrieve the historical data of Na+ / K+ ion transport efficiency in rice roots (unit: μmol / cm² / h) from the plant physiology database, and associate the soil pH value (range 6.0 - 9.5) and salt gradient (0 - 15 g / kg) conditions to generate an ion transport efficiency mapping table (data format: pH-salt gradient-efficiency key-value mapping); Generation of PSII photochemical efficiency level table: Based on the historical measurement results of the chlorophyll fluorescence imager, the maximum photochemical efficiency (Fv / Fm ratio) of the PSII reaction center is extracted and discretized into five stress levels according to the thresholds (0.8, 0.7, 0.6, 0.5, 0.4) to generate the PSII photochemical efficiency level table (data format: timestamp - stress level key - value pair); S14, Spatiotemporal benchmark alignment of multi - source data: Spatiotemporal benchmark alignment is performed on the collected dynamic monitoring data of saline - alkali land environment, rice phenomics data, and historical stress response data to generate a spatiotemporally aligned data stream. The specific method is as follows: Environmental data alignment: The GNSS coordinates of the soil surface reflectance spectrum sequence and the monitoring well coordinates of the groundwater depth - salinity time - series data are converted to a unified UTM coordinate system (grid size 1m×1m) through Gauss - Krüger projection, and the timestamps are aligned with the PPS pulse signal (synchronization error < 1ms); Phenotype data alignment: The canopy chlorophyll content matrix and the lidar point cloud coordinates of the plant height dynamic change sequence are rigidly registered with the GNSS reference station (residual < 0.5cm), and the timestamps of the stomatal conductance time - series data are aligned to the UTC standard time; Historical data mapping: The discrete meteorological station data in the saline - alkali stress event record table is mapped to the UTM grid of the target plot using Kriging interpolation, and the time axis is resampled according to the UNIX timestamp (millisecond level); S15, Standardization of heterogeneous data and generation of structured cube: The spatiotemporally aligned data stream is converted into a standardized multi - dimensional data cube to generate a structured data cube. The specific method is as follows: Generation of soil surface salt concentration data: The soil surface reflectance spectrum sequence is used to invert the soil surface salt concentration using the partial least squares regression model (PLSR, R² > 0.85) to generate soil surface salt concentration data (data format: UTM grid × timestamp × salt concentration); Generation of stomatal conductance data: The stomatal conductance time - series data is compensated for leaf temperature (corrected based on the Stefan - Boltzmann law) to generate temperature - compensated stomatal conductance data, data format: (UTM grid × timestamp × stomatal conductance); Generation of groundwater salinity data: The salinity values in the groundwater depth - salinity time - series data are interpolated according to the UTM grid to generate groundwater salinity data (data format: UTM grid × timestamp × salinity); Generation of root ion flow rate data: The root ion flow rate matrix is normalized to the standard temperature of 25°C to generate standardized root ion flow rate data (data format: UTM grid × timestamp × ion flow rate); PSII stress level data generation: Map the stress levels in the PSII photochemical efficiency level table to the UTM grid to generate PSII stress level data (data format: UTM grid × timestamp × stress level); The final output structured data cube includes the following independent data channels: Soil parameter layer: Includes soil surface salt concentration data and groundwater salinity data; Phenotype parameter layer: Includes stomatal conductance data after temperature compensation and standardized root ion flux rate data; Historical parameter layer: Includes PSII stress level data; The data cube format is: UTM grid × timestamp × data channels (salt concentration, salinity, stomatal conductance, ion flux rate, stress level).

[0025] S2 includes: S21, Multi-source data frequency-domain and time-domain joint calibration: Perform frequency-domain and time-domain joint calibration on the structured data cube through a spatio-temporal alignment engine to generate a synchronized calibration data stream. The synchronized calibration data stream includes a synchronized time-domain data stream and a frequency-domain coupling eigenvector. The specific method is as follows: Time-domain synchronization calibration: Based on the UNIX timestamp (millisecond level) of the data cube, perform time series alignment on the soil surface salt concentration data in the soil parameter layer and the stomatal conductance data after temperature compensation in the phenotype parameter layer. Use the cubic spline interpolation algorithm (interpolation window ±50ms) to eliminate the sensor trigger delay and generate a synchronized time-domain data stream (data format: UTM grid × synchronized timestamp × salt concentration-stomatal conductance dual channel); Frequency-domain feature alignment: Perform short-time Fourier transform (STFT, Hanning window length 1s, overlap rate 75%) on the soil surface salt concentration data, extract the amplitude spectrum of the salt osmotic pressure fluctuation in the 0-5Hz frequency band, and perform cross-spectrum analysis (coherence threshold > 0.6) with the frequency-domain phase spectrum (frequency band 0-2Hz) of the stomatal conductance data after temperature compensation to generate a frequency-domain coupling eigenvector (data format: UTM grid × frequency band × coherence coefficient); S22, Dynamic time warping algorithm optimization: Apply the dynamic time warping algorithm to the synchronized time-domain data stream in the synchronized calibration data stream to generate a dynamically time-warped optimized data stream. The specific steps include: Time series segmentation: Segment the soil surface reflectance spectrum sequence (generated by S11, time step 100ms) and the groundwater depth sequence (generated by S11, time step 100ms) into local subsequences (length 10 seconds) according to the UTM grid (1m × 1m); Local path optimization: Using the Manhattan distance as the cost function, constraining the bending window width to 15% of the subsequence length, calculating the cumulative distance matrix between adjacent grids, and retaining the slope constraint (±45°) of the optimal bending path; Global timing correction: Non-linearly scale the timestamps of the soil surface reflection spectrum sequence according to the optimal bending path, so that the timing deviation between the spectral sampling moment and the peak of the groundwater salinity change is <5 ms, and generate a dynamic time warping optimization data stream (data format: UTM grid × corrected timestamp × spectrum - burial depth alignment data); S23, Spatiotemporal coupling coding generation: Based on the generated dynamic time warping optimization data stream, combined with the frequency-domain coupling eigenvector, generate a spatiotemporal coupling coefficient matrix, and the specific method is as follows: Salt osmotic pressure gradient calculation: Calculate the spatial gradients (∂C / ∂x, ∂C / ∂y) of the soil surface salt concentration data (generated by S15) according to the UTM grid, and combine the groundwater burial depth data (generated by S11) to construct a Darcy's law-driven osmotic pressure diffusion model (the hydraulic conductivity K ranges from 10^-6 to 10^-4 m / s); Stomatal conductance dynamic analysis: Perform a first-order time difference (ΔG / Δt) on the temperature-compensated stomatal conductance data (generated by S15) to extract the stomatal response lag characteristics; Spatiotemporal coupling coefficient generation: Perform a Hadamard product operation (⊙) on the salt osmotic pressure gradient matrix and the stomatal conductance difference matrix, and weighted fusion with the frequency-domain coupling eigenvector (weight coefficient 0.7:0.3) to generate a spatiotemporal coupling coefficient matrix (dimension: UTM grid × timestamp × feature channel, coefficient range 0 - 1).

[0026] S2 also includes: S24, Multidimensional feature matrix generation: Integrate the spatiotemporal coupling coefficient matrix generated by S23 with the phenotypic parameter layer and historical parameter layer of S15 to generate a multidimensional feature matrix, specifically including: Phenotypic feature injection: Perform inverse distance weighted interpolation (power parameter p = 2) on the standardized root ion flow rate data (generated by S15) according to the UTM grid, and add it as an independent feature channel to the matrix; Historical feature mapping: Map the stress level of the PSII stress level data (generated by S15) to the corresponding UTM grid and associate the current salt gradient value; Matrix structure definition: The finally output multidimensional feature matrix includes a spatial dimension (UTM grid), a time dimension (UNIX timestamp), feature channels (salt osmotic pressure gradient, stomatal conductance difference, ion flow rate, stress level), and stores the normalized weights of the spatiotemporal coupling coefficient matrix; S25, Data quality verification: Perform integrity verification on the generated multidimensional feature matrix to generate a verified feature matrix, and the specific method is as follows: Space-time continuity check: Verify that the data coverage rate of each UTM grid on the time axis is > 98%, and the missing data is filled in using space-time Kalman filtering (the state equation is based on the osmotic pressure diffusion model); Coupling effectiveness check: Check the significance of the space-time coupling coefficient matrix (p-value < 0.05, chi-square test). If not up to standard, trigger the space-time benchmark realignment process of S14; Output interface encapsulation: Convert the verified multi-dimensional feature matrix into the FlatBuffers binary format (the field ID corresponds strictly to the input dimension of the gated recurrent unit in step S3), generate the verified feature matrix, and use it as the input data for step S3.

[0027] S3 includes: S31, Construction of the salinity gradient - physiological response dynamic coupling model: Based on the output verified feature matrix, construct the salinity gradient - physiological response dynamic coupling model, and generate the dynamic coupling model architecture, specifically including: Definition of the input layer: Map the channel dimensions (salt osmotic pressure gradient, stomatal conductance difference, ion flow rate, stress level) of the verified feature matrix to the model input nodes, and the input dimension corresponds strictly to the FlatBuffers binary field ID; Model structure design: Adopt a cascaded structure of a gated recurrent unit (GRU) and a bidirectional LSTM network with an attention mechanism, where: GRU layer configuration: The number of hidden layer units is 128, the time step is consistent with the UNIX timestamp density of the verified feature matrix (step size 100 ms), and the output gate activation function is Sigmoid; Attention mechanism layer configuration: Calculate the attention weight matrix of the GRU output state (dimension: timestamp × feature channel), and the weight assignment formula is: , where is the query vector, is the key vector, = 64; Bidirectional LSTM layer configuration: There are 64 forward and backward LSTM units each, and the outputs are concatenated into a 128-dimensional feature vector; S32, Space-time feature joint training and optimization: Conduct space-time feature joint training on the dynamic coupling model to generate the trained coupling model. The specific method is as follows: Training data division: Divide the verified feature matrix into a training set (70%), a validation set (15%), and a test set (15%) according to the timestamp, and maintain the space-time continuity of the UTM grid; Definition of the loss function: Adopt the weighted mean square error (WMSE), and assign a weight of 2 to the ion flow rate and PSII stress level channels; Optimizer configuration: Use the Nadam optimizer (learning rate 0.001, momentum β1 = 0.9, β2 = 0.999), early stopping mechanism (patience value = 10 epochs); S33, Na+ / K+ ion transmembrane transport feature extraction: Extract Na+ / K+ ion transmembrane transport features through the trained coupled model to generate ion transport correlation feature vectors, specifically including: Ion flux rate time series analysis: Input the standardized root ion flux rate data (generated by S15) into a bidirectional LSTM layer to extract forward-backward hidden states (dimension: 128 × timestamp); Transmembrane transport probability calculation: Focus on highly correlated time segments through the attention weight matrix to calculate the transmembrane transport probabilities of Na+ and K+ fluxes (output based on Sigmoid, threshold > 0.6); Feature vector encoding: Associate the transport probabilities with timestamps to generate ion transport correlation feature vectors (data format: UTM grid × timestamp × transport probability).

[0028] S3 also includes: S34, Chloroplast PSII exciton transfer correlation feature extraction: Extract exciton transfer correlation features based on PSII stress level data to generate PSII exciton transfer feature vectors, and the specific method is as follows: Stress level time series mapping: Input the PSII stress level data into a GRU layer to output hidden states (dimension: 128 × timestamp); Exciton transfer efficiency calculation: Screen key stress periods through the attention mechanism to calculate the exciton transfer efficiency of the PSII reaction center , calculated as: , where is the maximum fluorescence under light adaptation, is the real-time fluorescence; Feature vector encoding: Tensor concatenate the exciton transfer efficiency with the spatio-temporal coupling coefficient matrix (generated by S23) to generate PSII exciton transfer feature vectors (data format: UTM grid × timestamp × efficiency value); S35, Spatio-temporal coupling feature vector set generation: Fuse ion transport correlation feature vectors and PSII exciton transfer feature vectors to generate a spatio-temporal coupling feature vector set, and the specific method is as follows: Feature concatenation: Concatenate ion transport correlation feature vectors and PSII exciton transfer feature vectors along the feature channel dimension to generate a concatenated feature tensor (dimension: UTM grid × timestamp × 256); Dimensionality reduction processing: Reduce the 256-dimensional features to 32 dimensions through principal component analysis (PCA, retaining variance > 95%); Vector set encapsulation: The features after dimensionality reduction are encapsulated into a spatio-temporal coupled feature vector set according to the UTM grid (data format: grid ID × timestamp × feature vector), which is used as the input of step S4.

[0029] S4 includes: S41, Hilbert space reconstruction: Perform Hilbert space reconstruction on the spatio-temporal coupled feature vector set through the quantization feature analysis module to generate a quantum state vector space. The specific method is as follows: Orthogonalization of feature vectors: Perform Gram-Schmidt orthogonalization on the spatio-temporal coupled feature vector set (data format: grid ID × timestamp × feature vector) to generate an orthogonal basis vector group (dimension: 32×32); Hilbert space mapping: Map the feature vectors of each UTM grid to an infinite-dimensional Hilbert space, and define the quantum state vector as: ; Among them, is the Hermite orthogonal basis function, is the time-varying coefficient; Quantum state vector generation: Convert the spatio-temporal coupled feature vector into a quantum state amplitude through the quantum Fourier transform (QFT) to generate a quantum state vector space (data format: grid ID × timestamp × quantum state amplitude); S42, Root cell proton pump dynamics modeling: Calculate the root cell proton pump dynamics parameters based on the quantum state vector space to generate a proton pump tunneling probability model, which specifically includes: Proton concentration gradient extraction: Extract the soil surface salt concentration data and groundwater salinity data from the soil parameter layer generated in S15, and calculate the proton concentration difference inside and outside the root cell membrane ; Tunneling barrier construction: According to construct a proton pump transmembrane barrier model (barrier height , width , and define the barrier function as: ; Tunneling probability calculation: Calculate the proton tunneling probability by solving the one-dimensional stationary Schrödinger equation : , where is the transmembrane barrier function, describing the potential energy distribution of protons inside and outside the cell membrane, is the proton mass, with a value of kg, is the proton kinetic energy, is the reduced Planck constant, with a value of J.s, is the spatial coordinate of the transmembrane barrier; S43. Ion channel state transition path analysis: Combining the generated ion transport correlation feature vectors (grid ID × timestamp × transport probability), analyze the ion channel state transition path and generate a state transition Markov chain. The specific method is as follows: Definition of state space: Discretize the ion channel state into open ( ), closed ( ), and inactivated ( ) three states (state set ; Construction of transition probability matrix: Calculate the state transition probability based on the ion transport probability (threshold > 0.6) : ; Among them, is the number of transitions from state to j at time t, is the total number of times of state i, is the state transition probability at time t; Path optimization: Extract the maximum probability transition path through the Viterbi algorithm to generate a state transition Markov chain (data format: grid ID × timestamp × transition path probability); S44. Generation of quantum probability cloud map: Integrate the proton pump tunneling probability model and the state transition Markov chain to generate a quantum probability cloud map, which specifically includes: Probability tensor fusion: Perform a tensor product operation on the proton tunneling probability and the ion channel transfer probability to generate a joint probability distribution , where is the tensor product operation; Cloud map rendering: Use the Monte Carlo random sampling method (number of samples ) to draw probability density isosurfaces (isosurface threshold > 0.8) on the UTM grid and color them through quantum phase encoding (color mapping: blue → low probability, red → high probability); Path superposition: Superimpose the maximum probability transition path on the cloud map in the form of vector arrows, and the arrow width is positively correlated with the transition probability; S45. Encapsulation of quantization analysis results: Standardize and encapsulate the quantum probability cloud map to generate an ion channel state transition quantum probability cloud map. The specific method is as follows: Spatial encoding: Convert the UTM grid coordinates to the GeoJSON format (EPSG:32649), and encode the timestamp according to the ISO8601 standard; Metadata association: Associate the spatio-temporal coupling coefficient matrix (coefficient range 0 - 1) generated in S23 as the confidence index of the probability cloud map; Output interface definition: Store the quantum probability cloud map in the HDF5 format (the data set includes probability density, transition path, and spatio-temporal coupling coefficient), generate the quantum probability cloud map of ion channel state transition, and use it as the input data for step S5.

[0030] S5 includes: S51, generation of the thermal map of soil conductivity in the target area: Based on the soil surface salt concentration data and the spatio-temporal coupling coefficient matrix, generate the thermal map of soil conductivity in the target area, and perform Gaussian pyramid decomposition on the thermal map of soil conductivity in the target area to generate a multi-scale soil conductivity feature tensor. The specific method is as follows: Conductivity inversion: Convert the salt concentration data into soil conductivity values through the soil salt-conductivity empirical formula (EC = 0.64 × salt concentration + 0.12, R² > 0.89); Thermal map construction: Use the Kriging interpolation method (the semi-variogram is an exponential model) to map the discrete conductivity values to the UTM grid (1m × 1m) to generate the thermal map of soil conductivity in the target area (data format: UTM grid × conductivity value); Multi-scale decomposition: Perform Gaussian pyramid decomposition (scale factor 1:2:4) on the thermal map of soil conductivity in the target area to generate a multi-scale soil conductivity feature tensor (dimension: UTM grid × scale × conductivity gradient); S52, quantum-soil multi-scale convolution operation: Integrate the generated quantum probability cloud map of ion channel state transition (HDF5 format) with the multi-scale soil conductivity feature tensor to generate a cross-domain fusion feature map, which specifically includes: Data alignment: Rigidly register the UTM grid of the quantum probability cloud map of ion channel state transition with the UTM grid of the thermal map of soil conductivity (residual < 0.1m), and align the timestamps to the UNIX millisecond standard; Multi-scale convolution: Use a 3D convolution kernel (size 3×3×3, stride 1) to perform convolution on the quantum probability density channel and the conductivity gradient channel scale by scale, and the activation function is LeakyReLU (α = 0.2) to generate a cross-domain fusion feature map (data format: UTM grid × scale × fusion feature channel); S53, construction of the geodesic domain adaptation model: Based on the cross-domain fusion feature map, construct a domain adaptation model of saline-alkali stress feature space and regional soil parameters to generate a geodesic domain adaptation model. The specific method is as follows: Feature space projection: Project the fusion feature map into a unified feature space through the geodesic mapping algorithm (Riemannian manifold dimension d = 8), and calculate the geodesic distance matrix (based on the Mahalanobis distance); Domain adaptation loss calculation: Define the domain difference loss as: , where, is the characteristic mean of the source domain (quantum probability space), is the characteristic mean of the target domain (soil conductivity space); Model optimization: Using the adversarial training strategy (gradient reversal layer = 0.3), minimize the weighted sum of the classification loss (cross-entropy) and the domain difference loss (weight ratio 1:0.5) to generate a geodesic domain adaptation model.

[0031] S5 also includes: S54, Calculation of rice variety adaptability score: Predict the saline-alkali tolerance performance of rice varieties through the geodesic domain adaptation model to generate an adaptability score matrix, specifically including: Feature embedding: Input the spatio-temporal coupling feature vector set of S35 (dimension: grid ID × timestamp × 32) into the geodesic domain adaptation model to output a variety-specific embedding vector (dimension: 128); Definition of scoring rules: Define the adaptability score according to the ion channel state transition probability (generated by S43) and the PSII exciton transfer efficiency (generated by S34) , expressed as: ; Three-dimensional scoring mapping: Interpolate the score into a three-dimensional space (including elevation data, sourced from the GNSS elevation records of S14) according to the UTM grid to generate an adaptability score matrix (data format: UTM grid × elevation × score value); S55, Generation of three-dimensional saline-alkali tolerance performance prediction map: Package the adaptability score matrix into a visualization map to generate a three-dimensional saline-alkali tolerance performance prediction map, the specific method is as follows: Spatial encoding: Convert the UTM grid coordinates to the WGS84 geographic coordinate system (EPSG:4326), and convert the elevation data to altitude (based on the EGM96 geoid model); Thermal rendering: Use the HSV color space to map the score value (0 - 1 corresponding to the blue - green - red gradient) to generate a three-dimensional thermal isosurface (threshold interval 0.1); Metadata integration: Associate the multi-dimensional feature matrix of S24 (salt osmotic pressure gradient, stomatal conductance difference) as an auxiliary analysis layer for the map; Output interface definition: Store the map in the Cesium 3D Tiles format (including LOD levels), and attach Adaptive Scoring JSON metadata (fields bound to variety numbers, score values, and confidence levels).

[0032] As Figure 2 shown, a rice saline-alkali tolerance performance prediction system based on big data analysis is used to implement the above-mentioned rice saline-alkali tolerance performance prediction method based on big data analysis, including the following modules: Multi-source heterogeneous data acquisition module: It is used to acquire multi-source heterogeneous data, including dynamic monitoring data of saline-alkali land environment, rice phenomics data and historical stress response data, and record the spatio-temporal coordinates of the data through the GNSS positioning unit to achieve the positioning and alignment of the data; Spatio-temporal alignment and feature matrix construction module: It is used to perform frequency-domain and time-domain joint calibration on the acquired multi-source heterogeneous data, eliminate the time deviation of data acquisition through the dynamic time warping algorithm, and construct a multi-dimensional feature matrix with a unified spatio-temporal coordinate system; Physiological response feature extraction module: It is used to establish a dynamic coupling model of saline-alkali gradient-physiological response, extract the saline-alkali tolerance physiological response features of rice, including Na+ / K+ ion transmembrane transport features and PSII exciton transfer efficiency, and generate a spatio-temporal coupling feature vector set; Quantized feature analysis module: It is used to perform Hilbert space reconstruction on the generated spatio-temporal coupling feature vector set, calculate the proton pump kinetic parameters of root cells, and generate a quantum probability cloud map in combination with the analysis of the ion channel state transition path; Regional adaptability prediction module: It is used to fuse the generated quantum probability cloud map and the acquired soil surface salt concentration data, analyze the matching degree between the saline-alkali stress feature space and the regional soil parameters through the geodesic domain adaptation model, calculate the adaptability score of rice varieties, and finally generate a three-dimensional saline-alkali tolerance performance prediction map; Data storage and visualization module: It is used to store the generated three-dimensional saline-alkali tolerance performance prediction map and support the visual display of the data.

[0033] The present invention covers any substitutions, modifications, equivalent methods and solutions made on the essence and scope of the present invention. In order to enable the public to have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, and those skilled in the art can fully understand the present invention without the description of these details. In addition, well-known methods, processes, procedures, components and circuits are not described in detail to avoid unnecessary confusion to the essence of the present invention.

[0034] The above are only the preferred embodiments of the present invention. It should be pointed out that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A method for predicting rice salt-alkali tolerance based on big data analysis, characterized in that: The following steps are involved: S1, multi-source heterogeneous data acquisition: multi-source heterogeneous data are acquired through the multi-source heterogeneous data acquisition module, including dynamic monitoring data of saline-alkali land environment, rice phenotypic data and historical stress response data; S2, spatiotemporal alignment and feature matrix construction: The spatiotemporal alignment engine is used to perform frequency-domain-time domain joint calibration on the acquired multi-source heterogeneous data, and the millisecond-level timing deviation between drone hyperspectral sampling and groundwater level monitoring is eliminated through the dynamic time warping algorithm to generate a multi-dimensional feature matrix with a unified spatiotemporal coordinate system; S3, physiological response feature extraction: construct a dynamic coupling model of salinity gradient and physiological response, input the generated multi-dimensional feature matrix into the bidirectional LSTM network based on the attention mechanism of gated recurrent units, extract the correlation features of Na+ / K+ ion transmembrane transport and chloroplast PSII reaction center exciton transmission, and output a set of spatiotemporal coupling feature vectors; S4, quantized feature analysis: The output spatiotemporal coupling feature vector set is reconstructed in Hilbert space through the quantized feature analysis module, the tunneling probability distribution of the root cell proton pump dynamics is calculated, and a quantum probability cloud map containing the ion channel state transfer path is generated; S5, regional adaptability prediction: The regional adaptability prediction module based on transfer learning performs multi-scale convolution operations on the generated quantum probability cloud map and the soil conductivity thermodynamic map of the target area, and constructs a geodesic domain adaptation model of the salt-alkali stress feature space and regional soil parameters through the geodesic mapping algorithm, and outputs a three-dimensional salt-alkali tolerance prediction map containing the variety adaptability score.

2. The method for predicting rice salt-alkali tolerance based on big data analysis according to claim 1, characterized in that: The S1 includes: S11, acquisition of dynamic monitoring data of saline-alkali land environment: acquisition of dynamic monitoring data of saline-alkali land environment by a multi-source heterogeneous data acquisition module, wherein the dynamic monitoring data of saline-alkali land environment includes soil surface reflectance spectrum sequence and groundwater depth-mineralization time series data; S12, rice phenotypic data collection: collecting rice phenotypic data through a multi-source heterogeneous data collection module, wherein the rice phenotypic data includes a canopy chlorophyll content matrix, a plant height dynamic change sequence, a stomatal conductance time series data, and a root ion flow rate matrix; S13, historical stress response data collection: collecting historical stress response data through a multi-source heterogeneous data collection module, wherein the historical stress response data includes a salt-alkali stress event record table, an ion transport efficiency mapping table, and a PSII photochemical efficiency grade table; S14, spatiotemporal benchmark alignment of multi-source data: perform spatiotemporal benchmark alignment on the collected saline-alkali land environment dynamic monitoring data, rice phenotypic data and historical stress response data to generate a spatiotemporal alignment data stream; S15, Heterogeneous Data Standardization and Structured Cube Generation: Convert the spatiotemporal aligned data stream into a standardized multidimensional data cube and generate a structured data cube. The structured data cube includes the following independent data channels: Soil parameter layer: including soil surface salt concentration data and groundwater mineralization data; Phenotypic parameter layer: including temperature-compensated stomatal conductance data and standardized root ion flow rate data; Historical parameter layer: including PSII stress level data.

3. The method for predicting rice salt-alkali tolerance based on big data analysis according to claim 2, characterized in that: The S2 includes: S21, frequency domain-time domain joint calibration of multi-source data: frequency domain-time domain joint calibration is performed on the structured data cube through the spatiotemporal alignment engine to generate a synchronized calibration data stream, which includes a synchronized time domain data stream and a frequency domain coupling feature vector; S22, dynamic time warping algorithm optimization: applying a dynamic time warping algorithm to the synchronized time domain data stream in the synchronized calibration data stream to generate a dynamic time warping optimized data stream; S23, spatiotemporal coupling coding generation: Based on the generated dynamic time warping optimization data stream, combined with the frequency domain coupling eigenvector, a spatiotemporal coupling coefficient matrix is ​​generated.

4. The method for predicting rice salt-alkali tolerance based on big data analysis according to claim 3, characterized in that: The S2 further includes: S24, multi-dimensional feature matrix generation: the spatiotemporal coupling coefficient matrix generated by S23 is integrated with the phenotypic parameter layer and the historical parameter layer of S15 to generate a multi-dimensional feature matrix; S25, data quality verification: perform integrity check on the generated multi-dimensional feature matrix to generate a verified feature matrix.

5. The method for predicting rice salt-alkali tolerance based on big data analysis according to claim 4, characterized in that: The S3 includes: S31, Construction of dynamic coupling model of salinity gradient-physiological response: Based on the output verified feature matrix, the dynamic coupling model of salinity gradient-physiological response is constructed to generate the dynamic coupling model architecture; S32, joint training and optimization of spatiotemporal features: perform joint training of spatiotemporal features on the dynamic coupling model to generate a trained coupling model; S33, Na+ / K+ ion transmembrane transport feature extraction: Na+ / K+ ion transmembrane transport features are extracted through the post-training coupling model to generate ion transport associated feature vectors.

6. The method for predicting rice salt-alkali tolerance based on big data analysis according to claim 5, characterized in that: The S3 further includes: S34, chloroplast PSII exciton transfer related features extraction: extract exciton transfer related features based on PSII stress level data, and generate PSII exciton transfer feature vector; S35, generation of spatiotemporal coupling feature vector set: fusion of ion transport association feature vector and PSII exciton transfer feature vector to generate spatiotemporal coupling feature vector set.

7. The method for predicting rice salt-alkali tolerance based on big data analysis according to claim 6, characterized in that: The S4 includes: S41, Hilbert space reconstruction: The Hilbert space is reconstructed for the space-time coupling feature vector set through the quantized feature analysis module to generate the quantum state vector space; S42, Root cell proton pump dynamics modeling: Calculate the root cell proton pump dynamics parameters based on the quantum state vector space and generate a proton pump tunneling probability model; S43, ion channel state transition path analysis: Combined with the generated ion transport association feature vector, the ion channel state transition path is analyzed to generate a state transition Markov chain; S44, quantum probability cloud map generation: integrating the proton pump tunneling probability model with the state transition Markov chain to generate a quantum probability cloud map; S45, packaging of quantized analysis results: standardize and package the quantum probability cloud map to generate a quantum probability cloud map of ion channel state transition.

8. The method for predicting rice salt-alkali tolerance based on big data analysis according to claim 7, characterized in that: The S5 includes: S51, generating a heat map of soil conductivity in the target area: generating a heat map of soil conductivity in the target area based on the soil surface salt concentration data and the spatiotemporal coupling coefficient matrix, and performing Gaussian pyramid decomposition on the heat map of soil conductivity in the target area to generate a multi-scale soil conductivity feature tensor; S52, quantum-soil multi-scale convolution operation: the generated ion channel state transition quantum probability cloud map is fused with the multi-scale soil conductivity feature tensor to generate a cross-domain fusion feature map; S53, construction of geodesic domain adaptation model: construct a domain adaptation model of saline-alkali stress feature space and regional soil parameters based on the cross-domain fusion feature map to generate a geodesic domain adaptation model.

9. The method for predicting rice salt-alkali tolerance based on big data analysis according to claim 8, characterized in that: The S5 further includes: S54, calculation of rice variety adaptability score: predicting the salt-alkali tolerance of rice varieties through the geodesic domain adaptation model and generating an adaptability score matrix; S55, generation of three-dimensional salt-alkali resistance prediction map: encapsulate the adaptability scoring matrix into a visual map to generate a three-dimensional salt-alkali resistance prediction map.

10. A rice salt-alkali tolerance prediction system based on big data analysis, used to implement a rice salt-alkali tolerance prediction method based on big data analysis as claimed in any one of claims 1 to 9, characterized in that: Includes the following modules: Multi-source heterogeneous data acquisition module: used to collect multi-source heterogeneous data, including dynamic monitoring data of saline-alkali land environment, rice phenotypic data and historical stress response data, and record the spatiotemporal coordinates of the data through the GNSS positioning unit to achieve data positioning and alignment; Spatiotemporal alignment and feature matrix construction module: used to perform frequency-domain-time domain joint calibration on the collected multi-source heterogeneous data, eliminate the time deviation of data collection through the dynamic time warping algorithm, and construct a multi-dimensional feature matrix with a unified spatiotemporal coordinate system; Physiological response feature extraction module: used to establish a dynamic coupling model of salt-alkali gradient-physiological response, extract the physiological response characteristics of rice salt-alkali tolerance, including Na+ / K+ ion transmembrane transport characteristics and PSII exciton transfer efficiency, and generate a set of spatiotemporal coupling feature vectors; Quantum feature analysis module: used to reconstruct the Hilbert space of the generated spatiotemporal coupling feature vector set, calculate the root cell proton pump dynamics parameters, and generate a quantum probability cloud map in combination with ion channel state transition path analysis; Regional adaptability prediction module: used to fuse the generated quantum probability cloud map with the collected soil surface salt concentration data, analyze the matching degree between the saline-alkali stress characteristic space and the regional soil parameters through the geodesic domain adaptation model, calculate the adaptability score of rice varieties, and finally generate a three-dimensional salt-alkali tolerance prediction map; Data storage and visualization module: used to store the generated three-dimensional salt-alkali resistance performance prediction map and support the visualization of data.

Images