Plant photoperiod sensitivity real-time monitoring method and system

Abstract

The invention discloses a plant photoperiod sensitivity real-time monitoring method and system, and the method comprises the steps: collecting spectral reflection time sequence data of each organ of a plant under various photoperiod conditions, extracting photoperiod sensitive features, and generating a space-time specific marker set of organ and cell levels; and based on the marker, performing response classification on the signal intensity gradient of the dynamic fluctuation of the cell layer to form a light response signal distribution diagram, and analyzing the signal frequency and the fluctuation amplitude to realize the classification of the light sensitive characteristics of the organ. Further constructing a light sensitive characteristic positioning identification set based on cell ID and space coordinates, integrating light response fluctuation relevance of the organ and a cell layer by combining a data relevance network diagram, and finally establishing a light period response system of the plant organ. And predicting the frequency and spatial distribution dynamic trend of the photoresponse signal through a time sequence analysis method, and outputting a photoperiod dynamic response prediction map. According to the invention, real-time and high-precision photoperiod sensitivity analysis from the organ level to the cell level is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of plant photoperiod sensitivity monitoring, in particular to a plant photoperiod sensitivity real-time monitoring method and system. BACKGROUND

[0002] Plant photoperiod sensitivity research is an important direction in the fields of agriculture and biological science, and has key significance for improving crop yield, optimizing growth cycle and adapting to environmental changes. As a core factor affecting plant growth and development, photoperiod is directly related to the regulation of flowering, fruiting and overall life cycle of plants, therefore, real-time monitoring of plant response to photoperiod has become an urgent need to promote the development of precision agriculture.

[0003] Chinese patent, publication number: CN106258542A, publication date: January 4, 2017, discloses a rice variety photoperiod sensitivity identification method. Rice is planted in pots, and rice planting management is the same as that in the field; a stainless steel frame is built, and a light-shielding cloth that is breathable and light-proof is used to shield light; rice in the 7.0-9.1 leaf stage is moved into a dark room for short-day treatment for 7-10 days; according to the length of time from sowing to rice earing, rice varieties are divided into three categories: high sensitivity, medium sensitivity and low sensitivity.

[0004] The above technical solution has the following disadvantages: only the overall growth state of the plant is concerned, and the differential response of different organs and even cells to the photoperiod is ignored. This extensive monitoring method cannot capture the subtle changes in the plant, resulting in a lack of depth and pertinence of the data, and thus affecting the accurate judgment of the plant growth pattern and the effectiveness of the intervention measures. SUMMARY

[0005] In order to solve the problems of the prior art, the purpose of the present application is to provide a plant photoperiod sensitivity real-time monitoring method and system, which overcomes the defects of lack of differential response monitoring of organs and lack of dynamic fluctuation analysis at the cellular level, and establishes a full-dimensional analysis system of plant photoperiod response from organs to cells.

[0006] To achieve the above purpose, the technical solution adopted by the present application is as follows:

[0007] The present application provides a plant photoperiod sensitivity real-time monitoring method, which comprises:

[0008] S101, collecting spectral reflectance time series data of each organ of the plant under multiple photoperiod conditions;

[0009] S102, extracting photoperiod sensitivity features according to the spectral reflectance time series data, and obtaining a specific marker set of the photoperiod sensitivity features of each organ of the plant based on time and spatial dimensions;

[0010] S103, constructing a plant photoperiod sensitivity response model based on the specific marker set;

[0010] S103, responding classification of the signal intensity gradient of the dynamic fluctuation of the cell layer of each organ of the plant according to the specific marker set, to generate the light response signal distribution map of the cell layer of each organ of the plant based on time and space dimensions;

[0011] S104, obtaining the periodic frequency and fluctuation amplitude of the light response signal of the cell layer of each organ of the plant based on the light response signal distribution map, and classifying the light sensitivity characteristics of each organ of the plant;

[0012] S105, generating the light sensitivity characteristic spatial positioning identification set based on cell ID, spatial coordinates and corresponding sensitive characteristics according to the result of the light sensitivity characteristic classification;

[0013] S106, constructing the data correlation network graph about the light response signal fluctuation and the light sensitivity characteristic fluctuation of the cell layer of each organ of the plant according to the light sensitivity characteristic spatial positioning identification set;

[0014] S107, constructing the photoperiod response system of the plant organ according to the data correlation network graph;

[0015] S108, obtaining the plant organ photoperiod dynamic response prediction map by dynamically predicting the frequency and spatial distribution correlation of the light response signal of the cell layer of the plant organ according to the photoperiod response system of the plant organ through the time series analysis method.

[0016] As a preferred technical solution, in step S102, the light period sensitive feature is extracted according to the spectral reflectance time series data, which includes: obtaining the gene expression of each organ of the plant according to the spectral reflectance time series data; the gene expression of each organ of the plant is processed by principal component analysis method to obtain the light period sensitive feature of the principal component, and the corresponding gene and gene weight of the light period sensitive feature are determined; based on the gene, gene weight and time point, the contribution degree of each gene in the principal component is calculated

[0017] As a preferred technical solution, in step S102, the specific marker set of the light period sensitive feature of each organ of the plant based on time and space dimensions includes: calculating the variance of the light period sensitive feature set of each organ of the plant, if the variance contribution rate is greater than a preset threshold, then the sensitive feature set is classified according to organ type; the classified sensitive feature set is divided into several time intervals, and the spatial distribution heterogeneity index of the sensitive feature of each time interval is calculated by using clustering algorithm; based on the spatial distribution heterogeneity index of the sensitive feature of each time interval, the specific marker set of the light period sensitive feature of each organ of the plant based on time and space dimensions is obtained.

[0018] As a preferred technical solution, in step S103, the response classification of the signal intensity gradient of the cell layer dynamic fluctuation of each organ of the plant according to the specific marker set comprises: judging whether the sensitive feature and the spatial distribution heterogeneity index in the specific marker set exceed the preset threshold; if the sensitive feature and the spatial distribution heterogeneity index exceed the preset threshold, the convolutional neural network is used to extract the signal intensity gradient feature of the cell layer dynamic fluctuation; the signal intensity gradient feature is classified by the convolutional neural network to obtain the signal classification result of the cell layer dynamic fluctuation of each organ of the plant; and the generation of the light response signal distribution graph of the cell layer of each organ of the plant based on the time and space dimensions comprises: mapping the signal classification result to the space dimension by the space mapping algorithm to generate the light response signal distribution graph.

[0019] As a preferred technical solution, in step S104, based on the light response signal distribution graph, the periodic frequency and fluctuation amplitude of the light response signal of the cell layer of each organ of the plant are obtained by using the fast Fourier transform algorithm; and the light-sensitive characteristic classification of each organ of the plant comprises: calculating the spatial distribution correlation of each organ of the plant by using the Pearson correlation coefficient method to obtain the response contrast feature between organs; and respectively performing cluster analysis on the fluctuation amplitude of the light response signal and the response contrast feature between organs to realize the light-sensitive characteristic classification of each organ of the plant.

[0020] As a preferred technical solution, step S105 comprises: for each result set of the light-sensitive characteristic classification, extracting a two-dimensional feature vector of the periodic frequency and fluctuation amplitude of the light response signal; dividing the two-dimensional feature vector into several feature levels by a clustering algorithm, each feature level representing a different light response mode; calculating the signal intensity ratio of each cell at different time points for each feature level, determining the spatial distribution correlation of the cell in the three-dimensional space by spatial autocorrelation analysis, and generating a spatial correlation matrix; and based on the spatial correlation matrix, the spatial interpolation algorithm is used to finely adjust the cell position to generate the light-sensitive characteristic spatial positioning identification set.

[0021] As a preferred technical solution, step S106 comprises: according to the light-sensitive characteristic spatial positioning identification set, for each result set of the light-sensitive characteristic classification, extracting the periodic frequency and fluctuation amplitude of the light response signal, the spatial distribution correlation of each organ, and the overall sensitive feature; and using a weighted network construction algorithm to generate a data correlation network graph by taking the spatial distribution correlation of each organ as an edge weight.

[0022] As a preferred technical solution, step S107 comprises: for each class result set of photosensitive characteristic classification, extracting the signal intensity gradient of the dynamic fluctuation of the cell layer of each organ of the plant, and calculating the signal intensity ratio of each organ; inputting the signal intensity ratio of each organ into the data correlation network graph, taking organs as nodes and signal intensity ratio as edges, and optimizing the data correlation network graph; combining the spatial coordinates in the photosensitive characteristic spatial positioning identifier set, mapping the spatial coordinates to the optimized data correlation network graph, to obtain the photoperiod response system of the plant organs.

[0023] As a preferred technical solution, step S108 comprises: collecting the spectral reflection time series data of the plant organs under a specific photoperiod, extracting the periodic frequency and fluctuation amplitude of the light response signal, and if the fluctuation amplitude is greater than a preset threshold, then subsequent prediction is performed; through time series analysis, using an autoregressive moving average model to predict the frequency of the light response signal of the plant organs in a future period of time; performing grid processing on the spatial distribution correlation of the plant organs, to generate a light response intensity distribution matrix; combining the predicted light response signal frequency and the light response intensity distribution matrix, using a convolutional neural network model to perform visual modeling, and outputting a photoperiod dynamic response prediction map of the plant organs.

[0024] The application also provides a plant photoperiod sensitivity real-time monitoring system, comprising: a data acquisition module, the data acquisition module is used for: acquiring spectral reflectance time series data of each organ of the plant under multiple photoperiod conditions; a sensitive feature extraction module, the sensitive feature extraction module is used for: extracting photoperiod sensitive features according to the spectral reflectance time series data, obtaining a specific marker set of the photoperiod sensitive features of each organ of the plant based on time and space dimensions; a light response signal analysis module, the light response signal analysis module is used for: responding to the signal intensity gradient of the cell layer dynamic fluctuation of each organ of the plant according to the specific marker set, and generating a light response signal distribution map of the cell layer of each organ of the plant based on time and space dimensions; an organ light sensitivity classification module, the organ light sensitivity classification module is used for: obtaining the periodic frequency and fluctuation amplitude of the light response signal of the cell layer of each organ of the plant based on the light response signal distribution map, and classifying the light sensitivity characteristics of each organ of the plant; a spatial positioning identification construction module, the spatial positioning identification construction module is used for: generating a light sensitivity characteristic spatial positioning identification set based on the cell ID, spatial coordinates and corresponding sensitive features according to the result of the light sensitivity characteristic classification; a data correlation network construction module, the data correlation network construction module is used for: constructing a data correlation network graph about the light response signal fluctuation and the light sensitivity characteristic fluctuation of the cell layer of each organ of the plant according to the light sensitivity characteristic spatial positioning identification set; a photoperiod response system generation module, the photoperiod response system generation module is used for: constructing the photoperiod response system of the plant organ according to the data correlation network graph; and a dynamic response prediction module, the dynamic response prediction module is used for: dynamically predicting the frequency and spatial distribution correlation of the light response signal of the cell layer of the plant organ through time series analysis method according to the photoperiod response system of the plant organ, and obtaining a plant organ photoperiod dynamic response prediction map.

[0025] Compared with the prior art, the application has the beneficial effects that:

[0026] The application extracts organ-level photoperiod sensitive features through spectral reflectance time series, can identify specific response markers of multiple organs under different photoperiods, and therefore solves the problem that traditional research only focuses on the response of the whole plant and ignores the organ heterogeneity.

[0027] Based on the light response signal distribution map and the classification result, the application can locate the dynamic fluctuation rule of the light sensitivity signal at the cell layer, and therefore can verify the time and space synchronization of the light signal gradient and the core biological clock gene, and analyze the cell basis of photoperiod adaptability.

[0028] The light sensitivity characteristic spatial positioning identification set and the correlation network graph of the application integrate the cell ID, spatial coordinates and dynamic parameters, and for the first time construct a quantitative model of "light signal input-cell response-organ output", and therefore the application can identify the light intensity threshold response rule of specific cell groups in the plant organ.

[0029] In addition, the photoperiod dynamic response prediction map of the present application can guide the greenhouse light regulation strategy, realize the breeding and adaptive improvement of stress-resistant varieties, etc. BRIEF DESCRIPTION OF DRAWINGS

[0030] Figure 1 A flow chart of the steps of a plant photoperiod sensitivity real-time monitoring method of the present application. DETAILED DESCRIPTION

[0031] In order to enable personnel in the art to better understand the scheme of the present application, the technical scheme in the specific embodiments of the present application will be described clearly and completely below in combination with the drawings in the embodiments of the present application.

[0032] As shown in Figure 1 The present application provides a plant photoperiod sensitivity real-time monitoring method, which comprises:

[0033] S101, collecting spectral reflectance time series data of each organ of the plant under multiple photoperiod conditions;

[0034] S102, extracting photoperiod sensitivity features according to the spectral reflectance time series data, and obtaining a specific marker set of the photoperiod sensitivity features of each organ of the plant based on time and space dimensions;

[0035] S103, responding to the classification of the signal intensity gradient of the cell layer dynamic fluctuation of each organ of the plant according to the specific marker set, and generating a light response signal distribution map of the cell layer of each organ of the plant based on time and space dimensions;

[0036] S104, obtaining the periodic frequency and fluctuation amplitude of the light response signal of the cell layer of each organ of the plant based on the light response signal distribution map, and classifying the light sensitivity characteristics of each organ of the plant;

[0037] S105, generating a light sensitivity characteristic space positioning mark set based on cell ID, spatial coordinates and corresponding sensitivity characteristics according to the result of the light sensitivity characteristic classification;

[0038] S106, constructing a data correlation network map about the light response signal fluctuation and light sensitivity characteristic fluctuation of the cell layer of each organ of the plant according to the light sensitivity characteristic space positioning mark set;

[0039] S107, constructing a photoperiod response system of the plant organ according to the data correlation network map;

[0040] S108, dynamically predicting the frequency and spatial distribution correlation of the light response signal of the cell layer of the plant organ by time series analysis method according to the photoperiod response system of the plant organ, and obtaining a plant organ photoperiod dynamic response prediction map.

[0041] In the present application, the plant organs include leaves, stems, and flower buds, etc.

[0042] In the present application, organ-level photoperiod-sensitive characteristics are extracted through spectral reflectance time series, and specific response markers of multiple organs under different photoperiods can be identified, thus, the present application solves the problem that traditional research only focuses on the response of the whole plant and ignores the heterogeneity of organs.

[0043] Based on the light response signal distribution map and the classification result, the present application can locate the dynamic fluctuation rule of the light-sensitive signal at the cell layer, thus, the present application can verify the spatiotemporal synchronicity of the light signal gradient and the core biological clock gene, and analyze the cell basis of photoperiod adaptability.

[0044] The light-sensitive characteristic spatial positioning identification set and the associated network graph of the present application integrate cell ID, spatial coordinates, and dynamic parameters, and for the first time, a quantitative model of “light signal input-cell response-organ output” is constructed, thus, the present application can identify the light intensity threshold response rule of specific cell groups in plant organs.

[0045] In addition, the photoperiod dynamic response prediction map of the present application can guide the greenhouse light regulation strategy, and realize the breeding and adaptability improvement of stress-resistant varieties, etc.

[0046] In step S101, the spectral reflectance time series data is collected by a spectral sensor array. For example, the spectral sensor array is ASD FieldSpec 4, the wavelength range is 350-2500 nm, and the resolution is 3 nm-700 nm.

[0047] In the present application, the spectral sensor array is used to record the reflectance spectrum data at a high frequency acquisition rate of 10 times per second, and the reflectance spectrum data of leaves, stems, and flower buds under different photoperiods (8h light / 16h darkness, 12h light / 12h darkness, and 16h light / 8h darkness) is recorded for 72 hours, and a spectral reflectance time series data set containing 4000 waveband points is generated.

[0048] Further, step S101 further includes: adopting a data preprocessing method to denoise and standardize the spectral reflectance time series data, and obtaining the corrected spectral reflectance time series data. Further, the data preprocessing adopts a Savitzky-Golay filtering algorithm to smooth the spectral curve, eliminate the noise influence, and retain the characteristic signals of the cell layer dynamic fluctuation. In the present application, the Savitzky-Golay filtering algorithm window size is 11, and a quadratic polynomial fitting is adopted. The characteristic signals of the cell layer dynamic fluctuation include gene expression, cell distribution density, etc. In the present application, the characteristic signals of the cell layer dynamic fluctuation under the photoperiod response are collected from three organs of leaves, stems and flower buds, including the gene expression, cell distribution density and other indexes at each 2-hour time point within 24 hours, and the sample size is 100 time point data for each organ.

[0049] Further, step S101 further includes: extracting the main spectral features in the spectral reflectance time series data by a principal component analysis (PCA), calculating the reflectance rate of change of the plant organs at a specific absorption peak, and quantifying the influence of the photoperiod on the organ light response. For example, the reflectance rate of change of the leaves at 680 nm (chlorophyll absorption peak) and the stems at 1450 nm (water absorption peak) is calculated. In the present application, the fluctuation amplitude of the leaves at 680 nm under the 12h light period is 0.15±0.02, which is increased by 12.5% compared with the 8h period. The spectral reflectance time series data of the flower buds is modeled by a partial least squares regression (PLSR) to predict the correlation between the flower bud development stage and the spectral features. All the data are stored in an HDF5 format database, combined with a time stamp and a photoperiod label, and a multi-organ light response basic data set is constructed. The analysis results show that the signal intensity gradient of the leaves and the flower buds is more significant under the long photoperiod, and the change rate of the stem water-related features at 1450 nm is 0.08, which reveals the regulation mechanism of the photoperiod on the plant physiological dynamics. In order to ensure the logical rigor, the system automatically checks the data integrity, eliminates the abnormal values (reflectance out of the 3σ range), and classifies the spectral data by K-means clustering (K=3), verifies the response mode consistency of different organs under the photoperiod, and the clustering accuracy is 92%.

[0050] Further, step S101 further includes: constructing a time series feature matrix associated with the cell layer dynamic fluctuation and the signal intensity gradient. If the cell layer dynamic fluctuation and the signal intensity gradient of an organ in the time series feature matrix deviate from a preset threshold value, the spectral reflectance time series data of the organ is subjected to deep feature extraction, and a refined light response data set is obtained. According to the refined light response data set, a support vector machine algorithm is applied to classify and process the multi-organ light response data, and determine the response difference categories under different photoperiod conditions. The response difference categories after the classification processing are obtained, and a comprehensive light response basic data set containing the cell layer dynamic fluctuation and the signal intensity gradient is constructed.

[0051] The method for calculating the cell layer signal intensity gradient is as follows: after the gene expression quantity of each organ is normalized, the gradient descent algorithm is used to calculate the signal intensity change rate.

[0052] As a preferred technical solution, in step S102, the light period sensitive feature is extracted from the spectral reflectance time series data, including: obtaining the gene expression quantity of each organ of the plant according to the spectral reflectance time series data. The gene expression quantity of each organ of the plant is processed by dimension reduction through principal component analysis, the light period sensitive feature of the principal component is obtained, and the corresponding gene and gene weight of the light period sensitive feature are determined. Based on the gene, gene weight and time point, the contribution degree of each gene in the principal component is calculated by using the weighted average algorithm.

[0053] In this application, the PCA module in the sklearn library in Python is used to process the gene expression matrix of each organ (dimension 100*50, where 50 is the number of genes) by dimension reduction, and the number of principal components is set to 3. The calculation result shows that the first principal component of the leaf explains 65.2% of the variance, the stem explains 58.7%, and the flower bud explains 62.3%. The gene weight corresponding to the first three principal components of each organ is extracted through the principal component loading matrix, and the genes with an absolute weight value greater than 0.5 are selected as the light period sensitive features. Ten genes are selected for the leaf, eight for the stem, and twelve for the flower bud.

[0054] As a preferred technical solution, in step S102, the specific marker set of the light period sensitive feature of each organ of the plant based on the time and spatial dimensions includes: calculating the variance of the light period sensitive feature set of each organ of the plant, and if the variance contribution rate is greater than a preset threshold, the sensitive feature set is classified according to the organ type. The classified sensitive feature set is divided into several time intervals, and the spatial distribution heterogeneity index of the sensitive feature in each time interval is calculated by using the clustering algorithm. Based on the spatial distribution heterogeneity index of the sensitive feature in each time interval, the specific marker set of the light period sensitive feature of each organ of the plant based on the time and spatial dimensions is obtained.

[0055] In this application, 24 hours is divided into 4 6-hour intervals, and the spatial distribution heterogeneity in each time interval is classified using the K-means clustering algorithm (set K = 3), obtaining the heterogeneity index of the first interval of the leaf as 0.75, the stem as 0.68, and the flower bud as 0.82, indicating that the spatial response of the flower bud to the photoperiod is more significant. Finally, a threshold is set (gene weight > 0.6 and heterogeneity index > 0.7) to screen out 3 leaf-specific marker genes, 2 stem-specific marker genes, and 5 flower bud-specific marker genes, and the correlation of these markers with known light response pathways is verified through database comparison (the correlation coefficient is greater than 0.85), thereby forming a complete feature extraction and marker screening logic chain to ensure that the results can be used for subsequent photoperiod regulation research.

[0056] As a preferred technical solution, in step S103, the response classification of the signal intensity gradient of the cell layer dynamic fluctuation of each organ of the plant according to the specific marker set comprises: judging whether the sensitive feature and the spatial distribution heterogeneity index in the specific marker set exceed the preset threshold. If the sensitive feature and the spatial distribution heterogeneity index exceed the preset threshold, the convolutional neural network is used to extract the signal intensity gradient feature of the cell layer dynamic fluctuation. The signal classification result of the cell layer dynamic fluctuation of each organ of the plant is obtained by classifying the signal intensity gradient feature through the convolutional neural network. The generation of the light response signal distribution map of the cell layer of each organ of the plant based on the time and space dimensions comprises: mapping the signal classification result to the space dimension through the spatial mapping algorithm to generate the light response signal distribution map.

[0057] In this application, the convolutional neural network for extracting the signal intensity gradient feature of the cell layer dynamic fluctuation contains 5 convolutional layers and 3 fully connected layers, and the training data set contains 10,000 cell layer signal samples of leaves, stems and flower buds. The accuracy of the model on the validation set is 92.3%. The signal classification result of the cell layer dynamic fluctuation of each organ of the plant is: the leaf signal intensity gradient value is 2.8, the stem signal intensity gradient value is 1.9, and the flower bud signal intensity gradient value is 3.2. The classification result shows that the leaf and the flower bud belong to the high response category, and the stem is the low response category. The signal intensity gradient is mapped to a two-dimensional space using heat map visualization technology to generate a light response signal distribution map. The leaf area is displayed in red (high response, value > 2.5), the stem is blue (low response, value < 2.0), and the flower bud is orange (high response, value > 3.0).

[0058] Further, step S103 further comprises: optimizing the smoothness of the light response signal distribution map through a spatial interpolation algorithm to ensure the accuracy of the map. In this application, the Kriging interpolation method is used to optimize the smoothness of the light response signal distribution map, and the error range is controlled within 0.05.

[0059] Further, step S103 further comprises: associating the light response signal distribution map with the plant growth stage database, automatically comparing the current light response signal distribution map with the historical data, and analyzing the light response signal change trend. In this application, by comparing the current light response signal distribution map with the light response signal distribution map at the 30th day of growth (historical data), it is found that the leaf response value is increased by 15% compared with the historical data, thereby providing data support for subsequent light regulation. The entire process realizes data processing, model calling and atlas generation through automatic script, ensuring efficiency and accuracy.

[0060] As a preferred technical solution, in step S104, based on the light response signal distribution map, the periodic frequency and fluctuation amplitude of the light response signal of each organ cell layer of the plant are obtained by using the fast Fourier transform algorithm. The light-sensitive characteristic classification of the plant organs includes: using the Pearson correlation coefficient method to calculate the spatial distribution correlation of the plant organs, and obtaining the response contrast characteristics between organs. The fluctuation amplitude of the light response signal and the response contrast characteristics between organs are respectively subjected to cluster analysis, so as to realize the light-sensitive characteristic classification of the plant organs.

[0061] Specifically, when processing the light response distribution map data of the cell layer, the light response signal of each organ is first extracted by image processing technology.

[0062] For example, the light response signal data of organ A, organ B and organ C is extracted from the light response signal distribution map, the time span is 24 hours, the sampling frequency is 1 time per hour, the light response signal sequence of organ A is [5.2, 6.1, 7.3, …, 4.9], the light response signal sequence of organ B is [3.1, 4.2, 5.0, …, 3.0], and the light response signal sequence of organ C is [6.5, 7.8, 8.2, …, 6.0], a total of 24 data points. The periodic response frequency is analyzed by using the fast Fourier transform algorithm (FFT), and the calculation result shows that the main frequency of organ A is 0.0417 Hz (corresponding to a 24-hour period), the main frequency of organ B is 0.0833 Hz (corresponding to a 12-hour period), and the main frequency of organ C is 0.0417 Hz (24-hour period), indicating that organ A and organ C have diurnal response characteristics, and organ B has semi-diurnal response characteristics. Then, the spatial distribution correlation of each organ is calculated, and the Pearson correlation coefficient method is used to calculate the spatial distribution correlation coefficient of organ A and organ B, which is 0.65, the spatial distribution correlation coefficient of organ A and organ C is 0.85, and the spatial distribution correlation coefficient of organ B and organ C is 0.45, indicating that the spatial response distribution of organ A and organ C is highly correlated.

[0063] In this application, the main frequency of the leaf is 0.083 Hz, the stem is 0.042 Hz, and the flower bud is 0.025 Hz, indicating that the leaf is the most sensitive to the photoperiod. The spatial distribution correlation coefficient of the leaf and the stem is 0.75, the spatial distribution correlation coefficient of the stem and the flower bud is 0.62, and the spatial distribution correlation coefficient of the leaf and the flower bud is 0.58, indicating that the leaf and the stem have stronger photoperiod response correlation in spatial distribution.

[0064] In this application, the fluctuation amplitude and the response contrast between organs of the light response signal are respectively analyzed by clustering analysis, K-means clustering algorithm (K=2), and the fluctuation amplitude (standard deviation) and the average response intensity are used as features. The fluctuation amplitude of the leaf light response signal is 1.2, and the average is 6.0. The fluctuation amplitude of the stem light response signal is 0.9, and the average is 3.8. The fluctuation amplitude of the flower bud light response signal is 1.5, and the average is 7.0. The results cluster the leaf and the flower bud into one class (high fluctuation and high intensity), and the stem is a separate class (low fluctuation and low intensity). Finally, based on the clustering results and the frequency characteristics, the light sensitivity characteristic classification result set is determined, the leaf and the flower bud are classified into the "daily cycle high sensitivity class", and the stem is classified into the "semi-diurnal cycle low sensitivity class". The results are stored as a structured data table containing frequency, correlation coefficient, clustering category and other fields, forming a complete light sensitivity characteristic analysis report.

[0065] Through the above steps, from data extraction to classification results, a complete logical chain from original image to feature analysis to classification is formed, ensuring the scientificity and systematicness of the analysis.

[0066] As a preferred technical solution, step S105 includes: for each class result set of light sensitivity characteristic classification, extracting a two-dimensional feature vector of the periodic frequency and the fluctuation amplitude of the light response signal. The two-dimensional feature vector is divided into several feature levels by a clustering algorithm, and each feature level represents a different light response mode. In this application, the two-dimensional feature vectors are divided into 5 feature levels by K-means clustering algorithm (K=5). The signal intensity ratio of each cell at different time points is calculated for each feature level, and the spatial distribution correlation of the cell in the three-dimensional space is determined by spatial autocorrelation analysis to generate a spatial correlation matrix. In this application, the spatial distribution correlation of the cell in the three-dimensional space is determined by Moran's I index, and the calculation result is 0.65, indicating significant spatial aggregation. Based on the spatial correlation matrix, a spatial interpolation algorithm (such as Kriging interpolation) is used to fine-tune the cell position with an error controlled within 0.1 microns to generate the light sensitivity characteristic spatial positioning identification set.

[0067] To form a rigorous logical relationship, the spatial positioning identification set is associated with the biological function data of the cells (such as the gene expression amount), the correlation between the spatial distribution and the function is verified by the Pearson correlation coefficient, the feature mapping is ensured to be consistent with the actual biological significance, and thus the whole-process automatic processing from data extraction to spatial positioning is completed.

[0068] As a preferred technical solution, the step S106 includes: according to the spatial positioning identification set of the photosensitivity characteristics, for each result set of the photosensitivity characteristic classification, extracting the periodic frequency and fluctuation amplitude of the light response signal, the spatial distribution correlation of each organ, and the overall sensitivity characteristics. The periodic frequency and fluctuation amplitude of the light response signal and the overall sensitivity characteristics are generated into a data correlation network graph with the spatial distribution correlation of each organ as an edge weight by using a weighted network construction algorithm (such as Kruskal algorithm).

[0069] The data correlation network graph is output in the form of a heat map by a visualization tool, and the light period response correlation of each part is intuitively displayed. The above method forms a complete logical chain from data acquisition to result output through numerical quantification, algorithm analysis and model construction, and ensures the operability and scientificity of the technical implementation.

[0070] As a preferred technical solution, the step S107 includes: for each result set of the photosensitivity characteristic classification, extracting the signal intensity gradient of the dynamic fluctuation of the cells of each organ of the plant, and calculating the signal intensity ratio of each organ. The signal intensity ratio of each organ is input into the data correlation network graph, and the data correlation network graph is optimized with the organ as a node and the signal intensity ratio as an edge. The spatial coordinates in the spatial positioning identification set of the photosensitivity characteristics are combined to map the data correlation network graph after optimization, so as to obtain the light period response system of the plant organs.

[0071] The method for calculating the signal intensity ratio of each organ through the signal intensity gradient is: extracting the expression amount of a photosensitive gene (such as PHYB), and calculating the ratio of the expression amount of the photosensitive gene of each organ under different light periods.

[0072] The above process is realized by automatic data processing and algorithm analysis, ensuring a rigorous logic, mutual verification of front and back data, for example, the correlation between the signal intensity gradient and the ratio is embodied by the network graph, and the combination of the spatial positioning and the response probability further optimizes the regulation accuracy.

[0073] As a preferred technical solution, the step S108 comprises: collecting the spectral reflectance time series data of the plant organ to be predicted under a specific photoperiod, extracting the periodic frequency and fluctuation amplitude of the light response signal, and if the fluctuation amplitude is greater than a preset threshold (the fluctuation amplitude preset threshold in the present application is 0.85), then subsequent prediction is performed. By using the time series analysis method, the autoregressive moving average model (ARIMA) is used to predict the frequency of the light response signal of the plant organ in the future period of time. In combination with the spatial interpolation algorithm (Kriging method, etc.), the spatial distribution correlation of the plant organ is processed by gridding to generate a light response intensity distribution matrix (in the present application, the resolution of the light response intensity distribution matrix is 1 centimeter x 1 centimeter). In combination with the predicted frequency of the light response signal and the light response intensity distribution matrix, a convolutional neural network model is used for visual modeling to output a photoperiod dynamic response prediction map of the plant organ. The response intensity in the photoperiod dynamic response prediction map is displayed in the form of a heat map, and the intensity value range is 0.3 to 0.9, so as to ensure that the deviation between the prediction result and the actual observation is controlled within 5%.

[0074] The above process integrates data processing and algorithms to form a complete logical chain from data collection to prediction output, ensuring the continuity and accuracy of data flow and analysis results at each link.

[0075] The present application also provides a plant photoperiod sensitivity real-time monitoring system, which comprises: a data acquisition module, a sensitive feature extraction module, a light response signal analysis module, an organ light sensitivity classification module, a spatial positioning identification construction module, a data correlation network construction module, a photoperiod response system generation module, and a dynamic response prediction module.

[0076] The data acquisition module is used to collect the spectral reflectance time series data of each organ of the plant under various photoperiod conditions.

[0077] The sensitive feature extraction module is used to extract the photoperiod sensitive features according to the spectral reflectance time series data, and obtain a specific marker set of the photoperiod sensitive features of each organ of the plant based on the time and spatial dimensions.

[0078] The light response signal analysis module is used to respond to the signal intensity gradient of the dynamic fluctuation of each organ cell layer of the plant according to the specific marker set, and generate a light response signal distribution map of each organ cell layer of the plant based on the time and spatial dimensions.

[0079] The organ light sensitivity classification module is used to obtain the periodic frequency and fluctuation amplitude of the light response signal of each organ cell layer of the plant based on the light response signal distribution map, and classify the light sensitive characteristics of each organ of the plant.

[0080] The spatial positioning identification construction module is configured to generate a set of light-sensitive characteristic spatial positioning identifications based on the cell ID, the spatial coordinates, and the corresponding sensitive characteristics according to the result of the light-sensitive characteristic classification.

[0081] The data correlation network construction module is configured to construct a data correlation network diagram about the light response signal fluctuation and the light-sensitive characteristic fluctuation of each organ cell layer of the plant according to the set of light-sensitive characteristic spatial positioning identifications.

[0082] The photoperiod response system generation module is configured to construct a photoperiod response system of the plant organ according to the data correlation network diagram.

[0083] The dynamic response prediction module is configured to predict the dynamic trend of the frequency and the spatial distribution correlation of the light response signal of the plant organ cell layer by the time series analysis method according to the photoperiod response system of the plant organ, and obtain a photoperiod dynamic response prediction diagram of the plant organ.

[0084] It should be noted that the terms "first", "second", and similar terms used in the specification and claims of the present application do not indicate any order, number, or importance, but are only used to distinguish different components. Similarly, the terms "one" or "a" and the like do not indicate a quantity limitation, but indicate the existence of at least one. "Multiple" or "several" means at least two. Unless otherwise specified, the terms "before", "after", "left", "right", "below", and / or "above" and the like are only for ease of description, and are not limited to a position or a spatial orientation. The terms "include" or "contain" and the like mean that the elements or objects appearing before "include" or "contain" cover the elements or objects listed after "include" or "contain" and their equivalents, and do not exclude other elements or objects. The terms "connected" or "connected" and the like are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect.

[0085] The singular forms "a", "said", and "the" used in the specification and the appended claims of the present application are also intended to include the plural forms, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used herein means and includes any or all possible combinations of one or more associated listed items.

[0086] It should be understood that those skilled in the art can make improvements or changes according to the above description, and all such improvements and changes shall fall within the scope of protection of the appended claims of the present application.

Claims

1. A method for real-time monitoring of photoperiod sensitivity in plants, characterized in that, The method comprises: S101, collecting spectral reflection time series data of each organ of the plant under multiple photoperiod conditions; S102, extracting photoperiod sensitive features according to the spectral reflection time series data, and obtaining a specific marker set of the photoperiod sensitive features of each organ of the plant based on time and space dimensions; S103, responding to the signal intensity gradient of the dynamic fluctuation of the cell layer of each organ of the plant according to the specific marker set, and generating a light response signal distribution map of the cell layer of each organ of the plant based on time and space dimensions; S104, obtaining the periodic frequency and fluctuation amplitude of the light response signal of the cell layer of each organ of the plant based on the light response signal distribution map, and classifying the light sensitive characteristics of each organ of the plant; S105, generating a light sensitive characteristic space positioning identification set based on the cell ID, spatial coordinates and corresponding sensitive features according to the result of the light sensitive characteristic classification; S106, constructing a data correlation network graph about the light response signal fluctuation and the light sensitive characteristic fluctuation of the cell layer of each organ of the plant according to the light sensitive characteristic space positioning identification set; S107, constructing a photoperiod response system of the plant organ according to the data correlation network graph; S108, dynamically predicting the frequency and spatial distribution correlation of the light response signal of the cell layer of the plant organ by a time series analysis method according to the photoperiod response system of the plant organ, and obtaining a plant organ photoperiod dynamic response prediction map.

2. The method of claim 1, wherein the method is performed in real time. In step S102, the photoperiod sensitive features are extracted according to the spectral reflection time series data, which comprises: obtaining the gene expression amount of each organ of the plant according to the spectral reflection time series data; performing dimension reduction processing on the gene expression amount of each organ of the plant by principal component analysis to obtain the photoperiod sensitive features of the principal components, and determining the genes and gene weights corresponding to the photoperiod sensitive features; and calculating the contribution degree of each gene in the principal components based on the genes, gene weights and time points by using a weighted average algorithm.

3. The method of claim 2, wherein the method is performed in real time. In step S102, the specific marker set of the photoperiod sensitive features of each organ of the plant based on time and space dimensions is obtained, which comprises: calculating the variance of the photoperiod sensitive feature set of each organ of the plant, and if the variance contribution rate is greater than a preset threshold, classifying the sensitive feature set according to organ types; dividing the classified sensitive feature set into several time intervals, and calculating the spatial distribution heterogeneity index of the sensitive features in each time interval by using a clustering algorithm; and obtaining the specific marker set of the photoperiod sensitive features of each organ of the plant based on time and space dimensions based on the spatial distribution heterogeneity index of the sensitive features in each time interval.

4. The method of claim 1, wherein the method is performed in real time. In step S103, the signal intensity gradient of the dynamic fluctuation of the cell layer of each organ of the plant is classified according to the specific marker set, which comprises: judging whether the sensitive features and the spatial distribution heterogeneity index in the specific marker set exceed a preset threshold; if the sensitive features and the spatial distribution heterogeneity index exceed the preset threshold, extracting the signal intensity gradient feature of the dynamic fluctuation of the cell layer by using a convolutional neural network; and classifying the signal intensity gradient feature by the convolutional neural network to obtain the signal classification result of the dynamic fluctuation of the cell layer of each organ of the plant; The light response signal distribution map of each organ cell layer of the plant based on time and space dimensions comprises: mapping the signal classification result to the spatial dimension by a spatial mapping algorithm to generate the light response signal distribution map.

5. The method of claim 1, wherein the method is performed in real time. In step S104, based on the light response signal distribution map, the periodic frequency and fluctuation amplitude of the light response signal of each organ cell layer of the plant are obtained by using a fast Fourier transform algorithm. The light-sensitive characteristic classification of each organ of the plant comprises: calculating the spatial distribution correlation of each organ of the plant by using a Pearson correlation coefficient method to obtain inter-organ response contrast features; and performing clustering analysis on the fluctuation amplitude of the light response signal and the inter-organ response contrast features respectively to realize the light-sensitive characteristic classification of each organ of the plant.

6. The method of claim 1, wherein the method is performed in real time. Step S105 comprises: for each class result set of the light-sensitive characteristic classification, extracting a two-dimensional feature vector of the periodic frequency and fluctuation amplitude of the light response signal; dividing the two-dimensional feature vector into several feature levels by a clustering algorithm, each feature level representing a different light response mode; for each feature level, respectively calculating the signal intensity ratio of each cell at different time points, determining the spatial distribution correlation of the cells in the three-dimensional space by spatial autocorrelation analysis, and generating a spatial correlation matrix; based on the spatial correlation matrix, using a spatial interpolation algorithm to finely adjust the cell position to generate the light-sensitive characteristic spatial positioning identifier set.

7. The method of claim 1, wherein the method is performed in real time. Step S106 comprises: according to the light-sensitive characteristic spatial positioning identifier set, for each class result set of the light-sensitive characteristic classification, extracting the periodic frequency and fluctuation amplitude of the light response signal, the spatial distribution correlation of each organ, and the overall sensitive feature; using a weighted network construction algorithm to generate a data correlation network graph with the periodic frequency and fluctuation amplitude of the light response signal and the overall sensitive feature as the edge weight of the spatial distribution correlation of each organ.

8. The method of claim 1, wherein the method is performed in real time. Step S107 comprises: for each class result set of the light-sensitive characteristic classification, extracting the signal intensity gradient of the dynamic fluctuation of each organ cell layer of the plant, calculating the signal intensity ratio of each organ; inputting the signal intensity ratio of each organ into the data correlation network graph to optimize the data correlation network graph with the organ as the node and the signal intensity ratio as the edge; and combining the spatial coordinates in the light-sensitive characteristic spatial positioning identifier set to map the spatial coordinates to the optimized data correlation network graph to obtain the photoperiod response system of the plant organ.

9. The method of claim 1, wherein the method is performed in real time. Step S108 comprises: collecting the spectral reflectance time series data of the plant organ under a specific photoperiod, extracting the periodic frequency and fluctuation amplitude of the light response signal, and if the fluctuation amplitude is greater than a preset threshold, performing subsequent prediction; using an autoregressive moving average model to predict the frequency of the light response signal of the plant organ in a future period of time by time series analysis; performing grid processing on the spatial distribution correlation of the plant organ to generate a light response intensity distribution matrix; combining the predicted frequency of the light response signal and the light response intensity distribution matrix, and using a convolutional neural network model to perform visual modeling to output a photoperiod dynamic response prediction map of the plant organ.

10. A plant photoperiod sensitivity real-time monitoring system, characterized by, The system comprises: The data acquisition module is configured to: acquire spectral reflectance time series data of each organ of the plant under multiple photoperiod conditions. The sensitive feature extraction module is configured to: extract photoperiod-sensitive features according to the spectral reflectance time series data, and obtain a specific marker set of the photoperiod-sensitive features of each organ of the plant based on time and spatial dimensions. The light response signal analysis module is configured to: perform response classification on signal intensity gradients of cell layer dynamic fluctuations of each organ of the plant according to the specific marker set, and generate a light response signal distribution map of each organ of the plant based on time and spatial dimensions. The organ light-sensitive classification module is configured to: obtain periodic frequency and fluctuation amplitude of the light response signal of each organ of the plant based on the light response signal distribution map, and perform light-sensitive characteristic classification on each organ of the plant. The spatial positioning identification construction module is configured to: generate a light-sensitive characteristic spatial positioning identification set based on cell ID, spatial coordinates and corresponding sensitive features according to a result of the light-sensitive characteristic classification. The data correlation network construction module is configured to: construct a data correlation network graph about light response signal fluctuations and light-sensitive characteristic fluctuations of each organ of the plant according to the light-sensitive characteristic spatial positioning identification set. The photoperiod response system generation module is configured to: construct a photoperiod response system of the plant organs according to the data correlation network graph. The dynamic response prediction module is configured to: perform dynamic trend prediction on frequency and spatial distribution correlation of the light response signal of each organ of the plant by a time series analysis method according to the photoperiod response system of the plant organs, and obtain a photoperiod dynamic response prediction map of the plant organs.

Images