Ginger planting environment big data analysis monitoring system and method

Abstract

The invention belongs to the technical field of biological monitoring, and discloses a ginger planting environment big data analysis monitoring system and method. Comprising the steps that biological data in a ginger planting environment are collected, and the biological data comprise image data and biological environment index data; performing data cleaning on the biological data to obtain accurate biological data; performing data analysis on the accurate biological data to obtain an optimal environmental parameter combination; various environmental parameters in the ginger planting environment are adjusted based on the optimal environmental parameter combination; an early warning device is arranged to monitor the adjusted ginger planting environment, and when environmental parameters of the adjusted ginger planting environment exceed a preset safety range, early warning information is automatically sent to a preset database; the ginger planting environment is accurately monitored, and scientific development of ginger planting is promoted.

Description

Technical Field

[0001] The present invention relates to the technical field of biological monitoring. More specifically, the present invention relates to a big data analysis and monitoring system and method for ginger planting environment. Background Art

[0002] The patent with the application publication number CN119314572A discloses a biological reaction intelligent monitoring system and method, belonging to the field of intelligent monitoring. The system includes: a reaction stage division unit for dividing and determining multiple reaction stages; a variable relationship mining unit for determining the dominant variable and auxiliary variables, mining the linear relationship between variables, collaborating with an insensitive loss function, and supervising the training of a soft monitoring module; a continuous monitoring and acquisition unit for configuring a monitoring device group, performing continuous monitoring, and determining reaction monitoring data; a data analysis and screening unit for preprocessing data, performing linear analysis and abnormal reaction positioning, and determining effective monitoring data; a game equilibrium decision-making unit for making a game equilibrium decision on the abnormal reaction identifier and determining a reaction adjustment strategy; and a reaction monitoring management unit for responding to the adjustment strategy, performing feedback monitoring analysis and reaction monitoring management, improving the accuracy of biological reaction monitoring, realizing real-time online monitoring, and performing intelligent control for abnormal situations.

[0003] Due to equipment aging and environmental factors during the monitoring data acquisition process, the quality of the collected data is often uneven, lacking an efficient data preprocessing method, which has a great impact on the monitoring accuracy of the ginger planting environment; since the growth of ginger is jointly affected by multiple environmental parameters, traditional data preprocessing methods are difficult to capture the complex relationships between these environmental parameters. For example, environmental parameters all have temporality, with certain trends and periodicities. When dealing with missing values, simply using an interpolation algorithm to fill in the missing values will ignore the temporality of environmental parameters, resulting in inaccurate missing value processing results; it is difficult to control the growth environment of crops through the analysis of monitoring data. The adjustment of environmental parameters in the ginger planting environment often relies on experience, lacking a scientific and effective method to determine the optimal growth conditions, which easily leads to poor ginger planting conditions and ultimately low ginger yield and quality.

[0004] In view of this, the present invention proposes a big data analysis and monitoring system and method for ginger planting environment to solve the above problems. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art and to achieve the above object, the present invention provides the following technical solution: A big data analysis and monitoring method for ginger planting environment, including:

[0006] S1. Collect biological data in the ginger planting environment, where the biological data includes image data and biological environment index data;

[0007] S2. Clean the biological data to obtain accurate biological data;

[0008] S3. Analyze the accurate biological data to obtain the optimal combination of environmental parameters; adjust each environmental parameter in the ginger planting environment based on the optimal combination of environmental parameters;

[0009] S4. Set up a warning device to monitor the adjusted ginger planting environment. When the environmental parameters of the adjusted ginger planting environment exceed the preset safe range, automatically send a warning message to the preset database.

[0010] Further, the acquisition method of the biological data includes: using a high-resolution camera to collect image data, equipped with a wide-angle lens and thermal imaging function, and setting it at a suitable position in the ginger planting area; collecting corresponding types of environmental parameters by deploying a variety of sensors in the ginger planting area, and the environmental parameters include temperature, humidity, light intensity, carbon dioxide concentration, and soil pH value; integrating all types of environmental parameters to obtain biological environment index data; querying the timestamps of each image in the image data and the timestamps of each environmental index data point in the biological environment index data, matching the image data and the biological environment index data based on the timestamps, and adding the same feature code to the image data and the biological environment index data that can be matched.

[0011] Further, the method of cleaning the biological data includes:

[0012] Perform missing value processing on the biological environment index data to obtain complete index data; perform outlier processing on the complete index data to obtain normal index data; perform denoising processing on the image data to obtain enhanced image data; integrate the normal index data and the enhanced image data to obtain accurate biological data;

[0013] The method of performing missing value processing on the biological environment index data includes:

[0014] Construct a set of prediction models, using the decision tree model as the basic framework for each prediction model; use the set of prediction models to predict the missing values of the biological environment index data to obtain the predicted values of each missing index data point in the biological environment index data; use the predicted values to fill the missing values of each missing index data point to obtain complete index data;

[0015] The method of constructing a set of prediction models includes:

[0016] Collect complete historical indicator data as the training set of the prediction model; use different types of historical environmental parameters in the complete historical indicator data as the training labels of different types of prediction models; train corresponding types of prediction models based on the complete historical indicator data and different types of historical environmental parameters, define corresponding loss functions for each type of prediction model, calculate the function values of each loss function, and when the function value of any loss function no longer decreases, fix the parameters to obtain a trained prediction model of any type, and integrate all the trained prediction models to obtain a prediction model set.

[0017] Further, the method for processing outliers in the complete indicator data includes:

[0018] Combine all the complete indicator data points belonging to any one timestamp in the complete indicator data into a multi-dimensional vector; perform clustering processing on all the multi-dimensional vectors to obtain a set of grouped multi-dimensional vectors; perform outlier marking on the set of grouped multi-dimensional vectors to obtain clustering differences, and determine outliers based on the clustering differences; adjust the outliers by querying a preset ginger planting knowledge base to obtain normal indicator data;

[0019] The method for performing clustering processing on all the multi-dimensional vectors includes:

[0020] Randomly select K multi-dimensional vectors as the initial clustering centers; calculate the Euclidean distance between each multi-dimensional vector and any one of the initial clustering centers; if the Euclidean distance between any one multi-dimensional vector and any one of the initial clustering centers is less than a preset distance threshold, then classify this multi-dimensional vector and this initial clustering center into one cluster class;

[0021] Calculate the average value of the Euclidean distances between any one multi-dimensional vector and other multi-dimensional vectors in its cluster class to obtain the average Euclidean distance of this multi-dimensional vector; calculate the average value of the Euclidean distances between any one multi-dimensional vector and the clustering centers of other cluster classes to obtain the average cluster distance of this multi-dimensional vector; based on the average Euclidean distance and the average cluster distance of this multi-dimensional vector, obtain the clustering coefficient of this multi-dimensional vector; calculate the clustering coefficients of all the multi-dimensional vectors and take the average value to obtain the average clustering coefficient;

[0022] Continuously adjust the number of initial clustering centers, and calculate the average clustering coefficient after each adjustment. When the average clustering coefficient reaches the maximum value, fix the number of clusters at this time to obtain the optimal number of clustering centers. Update the initial clustering centers based on the optimal number of clustering centers, and re - perform the clustering process to obtain the grouped multi - dimensional vector set. Calculate the Euclidean distance between each multi - dimensional vector in the grouped multi - dimensional vector set and the clustering center of its belonging cluster. Mark the multi - dimensional vectors with Euclidean distances greater than the preset classification threshold as abnormal, and record the abnormally marked multi - dimensional vectors as marked multi - dimensional vectors. Calculate the Euclidean distance between each element in each marked multi - dimensional vector and the clustering center of the cluster to which the marked multi - dimensional vector belongs, obtain the clustering difference of each element in each marked multi - dimensional vector, and determine the elements with clustering differences greater than the preset difference threshold as outliers.

[0023] Further, the methods for denoising the image data include:

[0024] Preset a wavelet basis function, and perform multi - scale wavelet decomposition on each image in the image data based on the preset wavelet basis function. First, perform low - pass and high - pass filtering in the horizontal direction to obtain the low - frequency component and high - frequency component of each image respectively. Then, perform vertical filtering on the low - frequency component of each image to obtain the low - frequency sub - image and high - frequency sub - image of each image. Perform horizontal and vertical filtering on the low - frequency sub - image of each layer again to obtain lower - level low - frequency sub - images and high - frequency sub - images. Repeat the multi - scale wavelet decomposition for each image until the maximum decomposition level is reached. Extract wavelet coefficients from the sub - images after each decomposition, and integrate all wavelet coefficients to obtain a wavelet coefficient set. Perform threshold processing on the wavelet coefficient set to obtain denoised frequency - domain data. Use inverse wavelet transform to reconstruct the signal for the denoised frequency - domain data to obtain denoised image data;

[0025] Use the RGB weighted average method to grayscale each denoised image in the denoised image data to obtain grayscale image data. Perform Gaussian filtering on each grayscale image in the grayscale image data to obtain filtered grayscale image data. Use an edge detection operator to calculate the pixel value change gradient of each filtered grayscale image in the filtered grayscale image data, and extract the edge features of each filtered grayscale image based on the pixel value change gradient. Sharpen each filtered grayscale image based on the edge features to obtain sharpened image data. Perform contrast enhancement processing on each sharpened image in the sharpened image data to obtain enhanced image data.

[0026] Further, the methods for performing threshold processing on the wavelet coefficient set include:

[0027] Preset a wavelet coefficient threshold, and construct an adaptive threshold function based on the wavelet coefficient threshold. Update each wavelet coefficient in the wavelet coefficient set based on the adaptive threshold function to obtain an updated wavelet coefficient set, and obtain denoised frequency - domain data;

[0028] The calculation formula of the adaptive threshold function is as follows:

[0029] where μ i,j represents the wavelet coefficient at the j-th translation position of the i-th level; represents the updated wavelet coefficient at the j-th translation position of the i-th level; A represents a judgment function; γ represents a preset wavelet coefficient threshold; α and β are dynamic adjustment factors.

[0030] Furthermore, the method for image sharpening each filtered grayscale image includes:

[0031] Construct a rectangular coordinate system with the grayscale pixel point at the lower left corner of each filtered grayscale image as the origin; calculate the Laplacian operator of each filtered grayscale image to obtain the initial Laplacian operator, and integrate all the initial Laplacian operators to obtain a set of Laplacian operators; construct a sliding window much smaller than the size of each filtered grayscale image and make the sliding window slide on any filtered grayscale image; within the area formed by the sliding window on any filtered grayscale image, calculate the pixel value change gradient of all grayscale pixel points within this area and take the average value to obtain the local average change gradient of this area; define an operator adjustment coefficient and preset a change gradient threshold; adjust the size of the operator adjustment coefficient based on the local average change gradient of this area and the preset change gradient threshold to obtain the operator adjustment coefficient of this area;

[0032] Update the initial Laplacian operator based on the operator adjustment coefficient of this area to obtain the updated Laplacian operator of this area; use the updated Laplacian operator of this area to perform a convolution operation with this area to complete the image sharpening of this area; translate the sliding window along the coordinate axis direction to perform image sharpening on other areas until the sliding window traverses the entire filtered grayscale image to complete the image sharpening and obtain the sharpened image; perform image sharpening on each filtered grayscale image, and integrate all the sharpened images to obtain sharpened image data;

[0033] The method for contrast enhancement processing of each sharpened image in the sharpened image data includes:

[0034] Construct a rectangular coordinate system with the sharpened pixel point at the lower left corner of each sharpened image as the origin; perform histogram equalization processing on any sharpened image to obtain the equalized pixel value of any sharpened pixel point in any sharpened image; calculate the pixel value change gradient of any sharpened pixel point, and calculate the weighting coefficient based on the pixel value change gradient of any sharpened pixel point; calculate the contrast enhancement pixel value based on the equalized pixel value of any sharpened pixel point and the weighting coefficient of this sharpened pixel point; calculate the contrast enhancement pixel value of each sharpened pixel point in each sharpened image to obtain enhanced image data.

[0035] Furthermore, the method for analyzing the precise biological data includes:

[0036] Constructing an analysis model to process the precise biological data to obtain a suitable combination of environmental parameters; optimizing the suitable combination of environmental parameters to obtain an optimal combination of environmental parameters;

[0037] The method for constructing the analysis model includes:

[0038] Taking the CNN convolutional neural network model as the basic framework of the analysis model; collecting historical precise biological data as the training set of the analysis model; constructing a growth condition scoring function, calculating the growth condition score of each piece of historical precise biological data, and taking the historical precise biological data with the highest growth condition score as the training label; training the analysis model with the historical precise biological data, defining an analysis loss function; calculating the function value of the analysis loss function until the function value of the analysis loss function no longer decreases, and fixing the parameters at this time to obtain the trained analysis model.

[0039] Furthermore, the method for optimizing the suitable combination of environmental parameters includes:

[0040] Initializing the particle swarm parameters, where the particle swarm parameters include the population size, the number of iterations, the best position of each particle individual, and the best position of the particle swarm; each particle individual in the particle swarm represents a suitable combination of environmental parameters;

[0041] Defining an optimization objective function where NUM represents the maximum number of iterations; t represents the number of iterations; m represents the population size; ω p represents the weight of the p-th element in any particle individual; represents the p-th element of the l-th particle individual in the t-th iteration; h represents any particle individual; represents the ideal value of the p-th element of the l-th particle individual in the t-th iteration; R cd represents the interaction intensity between the c-th element and the d-th element in any particle individual;

[0042] Calculating the position of each particle individual and the function value of the optimization objective function at the current position in each iteration. When the function value of the optimization objective function is less than the function value of the optimization objective function at the previous position of the particle individual, the current position is taken as the new best position of the particle individual; after a round of iteration, the best position of the particle individual where the particle with the minimum function value of the optimization objective function is located is taken as the new best position of the particle swarm;

[0043] Iterate the particle population repeatedly until the maximum number of iterations is reached; select the particle individual that can minimize the function value of the optimization objective function from all particle individuals as the optimal particle individual, and this optimal particle individual is the optimal combination of environmental parameters.

[0044] A big data analysis and monitoring system for ginger planting environment, which is used to implement a big data analysis and monitoring method for ginger planting environment, including:

[0045] A data acquisition module, which is used to acquire biological data in the ginger planting environment, and the biological data includes image data and biological environment index data;

[0046] A data cleaning module, which is used to clean the biological data to obtain accurate biological data;

[0047] A data analysis module, which is used to analyze the accurate biological data to obtain the optimal combination of environmental parameters; adjust various environmental parameters in the ginger planting environment based on the optimal combination of environmental parameters;

[0048] An environmental monitoring module, which sets an alarm to monitor the adjusted ginger planting environment. When the environmental parameters of the adjusted ginger planting environment exceed the preset safety range, an alarm message is automatically sent to the preset database; each module is connected by wired and / or wireless means.

[0049] The technical effects and advantages of the big data analysis and monitoring system and method for ginger planting environment of the present invention:

[0050] By collecting image data and biological environment index data in the ginger planting environment to form biological data, cleaning and analyzing the biological data, obtaining the optimal environmental parameter combination and applying it to the ginger planting environment, and setting an alarm to monitor the optimized ginger planting environment, a big data analysis and monitoring method for ginger planting environment is realized. Compared with the existing experience, the data collection is more extensive. It not only collects image data to observe the growth of ginger, but also collects biological environment index data with the same time stamp as the image data, and adds the same feature code to the image data and biological environment index data with the same time stamp, which is convenient to match the two kinds of data in subsequent operations. In terms of data cleaning, for the biological environment index data, the missing value processing and outlier processing are carried out respectively by using the model and clustering algorithm, making the biological environment index data more complete, the value conforming to the normal range, reducing manual intervention at the same time, and improving the data quality. For the image data, wavelet transform is used for denoising, and each image is sharpened and the contrast is enhanced, improving the quality of the image data and making the image clearer. In the process of data analysis, the optimal environmental parameter combination is obtained by constructing a model and using an optimization algorithm, and the ginger planting environment is adjusted based on the optimal parameter combination to optimize the ginger planting environment. Finally, an alarm is set to monitor the parameters of the optimized ginger planting environment. If it exceeds the preset range, a warning message will be automatically sent, realizing the function of monitoring the ginger planting environment. Brief Description of the Drawings

[0051] Figure 1 It is a schematic diagram of a big data analysis and monitoring method for ginger planting environment of the present invention;

[0052] Figure 2 It is a schematic diagram of a big data analysis and monitoring system for ginger planting environment of the present invention. Detailed Embodiments

[0053] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work shall fall within the protection scope of the present invention.

[0054] Embodiment 1;

[0055] Please refer to Figure 1 As shown, the big data analysis and monitoring method for ginger planting environment described in this embodiment includes:

[0056] S1. Collect biological data in the ginger planting environment, and the biological data includes image data and biological environment index data;

[0057] S2. Clean the biological data to obtain accurate biological data;

[0058] S3. Analyze the accurate biological data to obtain the optimal combination of environmental parameters; adjust each environmental parameter in the ginger planting environment based on the optimal combination of environmental parameters;

[0059] S4. Set up a warning device to monitor the adjusted ginger planting environment. When the environmental parameters of the adjusted ginger planting environment exceed the preset safe range, automatically send a warning message to the preset database.

[0060] The acquisition methods of biological data include: using a high-resolution camera to collect image data, equipped with a wide-angle lens and thermal imaging function, and setting it at a suitable position in the ginger planting area (collecting image data is to observe the growth of ginger at a certain moment. The wide-angle lens can cover a large area of the ginger planting area, and the thermal imaging function is added to monitor the temperature change of the soil to infer the growth status of ginger. After the above-ground part of ginger grows, the temperature change on the surface of the above-ground part of ginger can be observed through thermal imaging. Based on this temperature change, it is possible to identify whether there is insufficient water supply or pest and disease infestation in ginger); by deploying a variety of sensors in the ginger planting area to collect corresponding types of environmental parameters respectively, and the environmental parameters include temperature, humidity, light intensity, carbon dioxide concentration, and soil pH value; (for example, a temperature sensor measures temperature data, a humidity sensor measures humidity data, a light sensor measures light intensity data, a gas sensor measures carbon dioxide concentration, and a soil sensor measures soil pH value; by combining the use of various types of sensors to collect multi-dimensional environmental parameters, it is conducive to understanding the various conditions required for ginger growth) integrate all types of environmental parameters to obtain biological environment index data; query the timestamp of each image in the image data and the timestamp of each environmental index data point in the biological environment index data, and match the image data and the biological environment index data based on the timestamp, and add the same feature code to the image data and the biological environment index data that can be matched (for example, match and add the same feature code to the image data and the biological environment index data collected within the same time period; matching the image data and the biological environment index data based on the timestamp is to facilitate observing the growth of ginger and the change of environmental parameters within the same time period, ensuring the time consistency of biological data and realizing the synchronization and association of image data and biological environment index data).

[0061] The methods for cleaning biological data include:

[0062] During the acquisition process of biological data, some unexpected situations may occur. For example, sensor aging may lead to inaccurate measurement results of biological environment index data, resulting in abnormal values, or sensor failures may cause the inability to collect biological environment index data, resulting in missing values. For image data, there may be cases where the collected image data has noise and is not clear enough, or the contrast of the image data is too low, making the target object not obvious enough in the image. Therefore, it is necessary to clean the collected biological data to obtain higher-quality biological data for subsequent processing.

[0063] Process the missing values in the biological environment index data to obtain complete index data; process the abnormal values in the complete index data to obtain normal index data; perform denoising processing on the image data to obtain enhanced image data; integrate the normal index data and the enhanced image data to obtain accurate biological data.

[0064] The methods for processing the missing values in the biological environment index data include:

[0065] Since various environmental parameters in the biological environment index data change over time. For example, the light intensity gradually increases during the day and gradually decreases at night, repeating this process periodically. Therefore, the biological environment index data are all time-series type data. If only the interpolation algorithm is used to process the missing values in the biological environment index data, it is easy to ignore the trends and periodic changes existing in the time-series type data. Therefore, constructing a prediction model can provide a more effective solution.

[0066] Construct a set of prediction models, with the decision tree model as the basic framework for each prediction model; use the set of prediction models to predict the missing values in the biological environment index data to obtain the predicted values of each missing index data point in the biological environment index data; use the predicted values to fill in the missing values of each missing index data point to obtain complete index data.

[0067] The methods for constructing a set of prediction models include:

[0068] Collect the complete historical indicator data as the training set of the prediction model; use different types of historical environmental parameters in the complete historical indicator data as the training labels of different types of prediction models respectively; train the corresponding types of prediction models based on the complete historical indicator data and different types of historical environmental parameters (construct different types of prediction models for different types of missing values; for example, when the type of missing value is temperature, use the temperature in the historical environmental parameters as the training label, and other historical environmental parameters collected at the same moment as features, and the prediction model realizes the prediction of the missing value by learning the mapping relationship between the features and the training label; since this prediction model is used to predict the missing value of temperature, that is, the temperature prediction model, construct a corresponding prediction model for each environmental parameter, and use different prediction models to process the corresponding types of missing values), define a corresponding loss function for each type of prediction model (for example, define the temperature mean square error function for the temperature prediction model, and the variables of this function are the predicted value of temperature and the true value of temperature in the complete historical indicator data), calculate the function value of each loss function, when the function value of any loss function no longer decreases, fix the parameters to obtain a trained prediction model of any type, and integrate all the trained prediction models to obtain a set of prediction models.

[0069] Construct corresponding types of prediction models for different types of missing values. Each prediction model can be customized according to the characteristics of the target variable, improving the flexibility of missing value processing; each prediction model in the constructed set of prediction models only predicts one type of missing value, and can capture the dependence relationship between different variables more accurately.

[0070] The methods for handling outliers in the complete indicator data include:

[0071] In the growth environment monitoring data in the agricultural field, different environmental parameters may not have obvious distribution characteristics, and these data may show different change trends under different growth environments and different weather conditions of different crops; for this type of data for outlier processing, the clustering algorithm is an effective method, and the K-means clustering algorithm is selected to process the outliers in the complete indicator data; for the complete indicator data, there are multiple environmental parameters involved, and each environmental parameter has a unique change trend; the K-means clustering algorithm can perform overall clustering on the multi-dimensional vectors composed of these environmental parameters, comprehensively considering the relationship between multiple features, so as to distinguish normal values and outliers.

[0072] Combine all the complete indicator data points belonging to any one timestamp in the complete indicator data into a multi-dimensional vector (for example, the environmental parameters corresponding to a certain timestamp are temperature T0, humidity W0, light intensity H0, carbon dioxide concentration C0, and soil pH value P0 respectively, then the multi-dimensional vector corresponding to this timestamp is X O=[T0,W0,H0,C0,P0]; where X O Represents the multidimensional vector corresponding to the Oth timestamp); cluster all multidimensional vectors to obtain a grouped multidimensional vector set; mark the grouped multidimensional vector set as abnormal to obtain cluster differences, and determine the abnormal values ​​based on the cluster differences; adjust the abnormal values ​​by querying the preset ginger planting knowledge base to obtain normal indicator data (each environmental parameter should have a suitable range of variation and a normal trend of variation during the growth of ginger, and it is impossible to suddenly increase or decrease significantly. Therefore, it is selected to query the preset ginger planting knowledge base to obtain the real value range and change trend of the abnormal values, and adjust the abnormal values ​​based on this).

[0073] The methods for clustering all multidimensional vectors include:

[0074] Randomly select K multidimensional vectors as initial cluster centers; calculate the Euclidean distance between each multidimensional vector and any initial cluster center; if the Euclidean distance between any multidimensional vector and any initial cluster center is less than a preset distance threshold, the multidimensional vector and the initial cluster center are divided into one cluster.

[0075] Calculate the Euclidean distance between any multidimensional vector and other multidimensional vectors in its cluster and take the average value to obtain the average Euclidean distance of the multidimensional vector; calculate the Euclidean distance between any multidimensional vector and the cluster center of other clusters and take the average value to obtain the cluster average distance of the multidimensional vector; based on the average Euclidean distance of the multidimensional vector and the cluster average distance, obtain the clustering coefficient of the multidimensional vector; calculate the clustering coefficients of all multidimensional vectors and take the average value to obtain the average clustering coefficient.

[0076] The number of initial cluster centers is continuously adjusted, and the average clustering coefficient after each adjustment is calculated. When the average clustering coefficient reaches the maximum value, the number of clusters at this time is fixed to obtain the optimal number of cluster centers; the initial cluster centers are updated based on the optimal number of cluster centers, and clustering is performed again to obtain a set of grouped multidimensional vectors; the Euclidean distance between each multidimensional vector in the set of grouped multidimensional vectors and the cluster center of the cluster to which it belongs is calculated, and the multidimensional vectors with a Euclidean distance greater than a preset classification threshold are marked as abnormal, and the multidimensional vectors marked as abnormal are recorded as marked multidimensional vectors; the Euclidean distance between each element in each marked multidimensional vector and the cluster center of the cluster to which the marked multidimensional vector belongs is calculated to obtain the clustering difference of each element in each marked multidimensional vector, and the elements with clustering differences greater than a preset difference threshold are judged as abnormal values.

[0077] Using a clustering algorithm for outlier processing is suitable for processing multi-dimensional data, facilitating the handling of complex interrelationships among different environmental parameters, and being able to process outliers in large datasets more efficiently, improving work efficiency and reducing manual intervention.

[0078] The methods for denoising image data include:

[0079] Compared with traditional denoising methods (such as the mean filtering algorithm), wavelet transform has a better denoising effect; in order to more accurately separate noise and effective signals, multi-scale wavelet decomposition is performed on the image data to obtain subgraphs at different levels. According to the decomposition level and appropriate threshold processing method, noise at different frequencies can be removed at different levels; at the same time, using wavelet transform for denoising can effectively retain the edge and texture information in the image, avoiding the loss of image information caused by denoising.

[0080] Preset a wavelet basis function, and perform multi-scale wavelet decomposition on each image in the image data based on the preset wavelet basis function; first, perform low-pass and high-pass filtering in the horizontal direction to obtain the low-frequency component and high-frequency component of each image respectively, and then perform vertical filtering on the low-frequency component of each image to obtain the low-frequency subgraph and high-frequency subgraph of each image (select an appropriate wavelet basis function, such as the Daubechies wavelet basis function; use the filter corresponding to the wavelet basis function for filtering, such as using the Daubechies filter for filtering; the low-frequency subgraph mainly includes the main information and general outline of the image, and the high-frequency subgraph includes three parts, namely the horizontal high-frequency part, the vertical high-frequency part, and the diagonal high-frequency part, which respectively contain the details and noise information of the image in the horizontal, vertical, and diagonal directions); perform horizontal and vertical filtering on the low-frequency subgraph of each layer again to obtain a higher-level low-frequency subgraph and high-frequency subgraph; repeat the multi-scale wavelet decomposition for each image until the maximum decomposition level is reached (the maximum decomposition level floor = log 2 [min(CH, KA)]; where CH represents the length of the image and KA represents the width of the image; for example, for an image of 512×1024, floor = log 2 (512) = 9, that is, the maximum decomposition level is 9), extract wavelet coefficients from the subgraphs after each decomposition, and integrate all wavelet coefficients to obtain a wavelet coefficient set (the pixel values in the subgraphs formed after each decomposition reflect the frequency components of the subgraph at a certain scale and position, and the wavelet coefficients are the numerical representation forms of these frequency components); perform threshold processing on the wavelet coefficient set to obtain denoised frequency-domain data; use inverse wavelet transform to perform signal reconstruction on the denoised frequency-domain data to obtain denoised image data.

[0081] In each denoised image of the denoised image data, the noise has been removed, but some unclear areas in the image are not caused by noise; there are many factors in complex environmental conditions that cause the denoised image to be unclear. For example, too low brightness, too low contrast, or low resolution of the captured image due to instrument failure can all cause the denoised image to be blurred. Therefore, edge detection, image sharpening, and contrast enhancement need to be performed on the denoised image.

[0082] Use the RGB weighted average method (a commonly used image grayscale algorithm that converts a color image into a grayscale image by assigning different weights to different color channels of the color image) to grayscale each denoised image in the denoised image data to obtain grayscale image data; perform Gaussian filtering (Gaussian filtering is used to smooth the image, making the image smoother and more continuous, which can reduce edge detection errors caused by small-scale details and make the main structures and contours in the image easier to identify) on each grayscale image in the grayscale image data to obtain filtered grayscale image data; use an edge detection operator (such as the Sobel operator) to calculate the pixel value change gradient (the pixel value change gradient in a grayscale image refers to the change rate of grayscale) of each filtered grayscale image in the filtered grayscale image data, and extract the edge features (edge features refer to the contours, boundaries, and structures of objects in the image. The larger the pixel value change gradient value at a certain point, the more likely it is that this point represents the edge features in the image) of each filtered grayscale image based on the pixel value change gradient; perform image sharpening on each filtered grayscale image based on the edge features to obtain sharpened image data; perform contrast enhancement processing on each sharpened image in the sharpened image data to obtain enhanced image data.

[0083] The ways to perform threshold processing on the wavelet coefficient set include:

[0084] There are two traditional methods for performing threshold processing on wavelet coefficients, including the hard threshold processing method and the soft threshold processing method, which are used to handle different situations respectively; the traditional methods may not be able to adapt to different types and intensities of noise, so an adaptive threshold function is constructed for threshold processing; by adjusting the dynamic adjustment factor in the adaptive threshold function, the adaptive threshold function can adaptively update the wavelet coefficients to be denoised; using this function for threshold processing provides higher flexibility when processing image noise and also improves the denoising efficiency.

[0085] Preset a wavelet coefficient threshold, and construct an adaptive threshold function based on the wavelet coefficient threshold; update each wavelet coefficient in the wavelet coefficient set based on the adaptive threshold function to obtain an updated wavelet coefficient set and obtain denoised frequency domain data.

[0086] The calculation formula of the adaptive threshold function is:

[0087] where, μ i,j represents the wavelet coefficient at the j-th translation position in the i-th level (since multi-scale wavelet decomposition is repeated multiple times, wavelet coefficients exist at each level of decomposition, and this level is less than or equal to the maximum decomposition level; the translation position refers to the specific position of the wavelet coefficient in the sub-image obtained by decomposition, that is, the specific coordinates); represents the updated wavelet coefficient at the j-th translation position in the i-th level; A represents a judgment function (when the variable in the judgment function is greater than 0, the function takes the value of 1; when the variable in the judgment function is equal to 0, the function takes the value of 0; when the variable in the judgment function is less than 0, the function takes the value of -1); γ represents a preset wavelet coefficient threshold ( where, σ i represents the noise standard deviation of all wavelet coefficients in the i-th level; N represents the size of the sub-image to which the wavelet coefficients in the i-th level belong); α and β are dynamic adjustment factors (α ∈ [0, 1]; β ∈ [0, +∞]; α is used to control the degree of change of wavelet coefficients, and β is used to ensure the continuity of the adaptive threshold function, making the function smoother and more continuous).

[0088] The ways to perform image sharpening on each filtered grayscale image include:

[0089] Performing image sharpening on the filtered grayscale image is to enhance the edges and details in the image, making the image clearer. Therefore, the Laplace operator is selected for image sharpening processing; in order to make the Laplace operator adapt to the changes in different regions of the image, a method for adjusting the Laplace operator based on the pixel value change gradient in the local region is designed; the image is processed block by block by constructing a sliding window, and each region is analyzed and processed to ensure that each detail can be appropriately enhanced; an operator adjustment coefficient is defined, and this value is updated based on the pixel value change gradient in the sliding window region, and then the Laplace operator in the sliding window region is adjusted using the updated operator adjustment coefficient, so that the Laplace operator can apply different degrees of sharpening to different regions; this not only improves the calculation efficiency of the Laplace operator but also improves the flexibility of image sharpening.

[0090] Taking the grayscale pixel point at the lower left corner of each filtered grayscale image as the origin, a rectangular coordinate system is constructed (the X-axis direction is to the right, and the Y-axis direction is upward); calculate the Laplace operator of each filtered grayscale image to obtain the initial Laplace operator, and integrate all the initial Laplace operators to obtain a set of Laplace operators (the Laplace operator Among them, I(x, y) represents the pixel value of the image at the position (x, y); each Laplacian operator acts on a local area of the entire image; a sliding window much smaller than the size of each filtered grayscale image is constructed (much smaller means that the size of the filtered grayscale image is at least 100 times the size of the sliding window; for example, the size of the sliding window is 3×3, and the size of a certain filtered grayscale image is 512×512), and the sliding window is made to slide on any filtered grayscale image; within the area formed by the sliding window on any filtered grayscale image, calculate the pixel value change gradient of all grayscale pixel points within this area and take the average value to obtain the local average change gradient of this area; define an operator adjustment coefficient and preset a change gradient threshold; based on the local average change gradient of this area and the preset change gradient threshold, adjust the size of the operator adjustment coefficient to obtain the operator adjustment coefficient of this area (if the local average change gradient of any area is greater than the preset change gradient threshold, the operator adjustment coefficient B = 1 + log|G avg -G|; where G avg represents the local average change gradient of any area; G represents the preset change gradient threshold; if the local average change gradient of any area is less than or equal to the preset change gradient threshold, the operator adjustment coefficient B = 0.5).

[0091] Based on the operator adjustment coefficient of this area, update the initial Laplacian operator to obtain the updated Laplacian operator of this area (the value of the updated Laplacian operator is the product of the operator adjustment coefficient of this area and the initial Laplacian operator); use the updated Laplacian operator of this area to perform a convolution operation with this area to complete the image sharpening of this area; translate the sliding window along the coordinate axis direction to perform image sharpening on other areas until the sliding window traverses the entire filtered grayscale image to complete the image sharpening and obtain the sharpened image; perform image sharpening on each filtered grayscale image and integrate all the sharpened images to obtain the sharpened image data.

[0092] The methods for enhancing the contrast of each sharpened image in the sharpened image data include:[[]]

[0093] During the process of enhancing the contrast of the sharpened image, by performing histogram equalization on the sharpened image and dynamically adjusting the weighting coefficient to control the degree of contrast enhancement, ensure that the sharpened image is appropriately contrast-enhanced while being able to preserve the image details.

[0094] Construct a rectangular coordinate system with the sharpened pixel point at the lower left corner of each sharpened image as the origin; perform histogram equalization on any sharpened image to obtain the equalized pixel value of any sharpened pixel point in any sharpened image (histogram equalization is a commonly used contrast enhancement algorithm, which is mapped through the cumulative distribution function, and the calculation formula is: U(a,b) = CDF[U0(a,b)] × [max(U0) - min(U0)]; where, U0(a,b) represents the pixel value of the sharpened image U0 at the position (a,b); max(U0) represents the maximum pixel value in the sharpened image U0, min(U0) represents the minimum pixel value in the sharpened image U0; CDF represents the cumulative distribution function; U(a,b) represents the equalized pixel value of the sharpened image U0 at the position (a,b)); calculate the pixel value change gradient of any sharpened pixel point, and calculate the weighting coefficient based on the pixel value change gradient of any sharpened pixel point (weighting coefficient where, G0(a,b) represents the pixel value change gradient of the sharpened pixel point at the position (a,b); δ represents a constant, δ ∈ (-∞, +∞)); calculate the contrast enhancement pixel value based on the equalized pixel value of any sharpened pixel point and the weighting coefficient of this sharpened pixel point (the contrast enhancement pixel value calculation formula is: U after (a,b) = U(a,b) + W(a,b) × |U(a,b) - U0(a,b)|; where, U after (a,b) represents the contrast enhancement pixel value of the sharpened pixel point at the position (a,b)); calculate the contrast enhancement pixel value of each sharpened pixel point of each sharpened image to obtain the enhanced image data.

[0095] By performing image sharpening and contrast enhancement on the filtered grayscale image successively, the edges and details in the image become clearer, the image quality is improved, and the working efficiency of subsequent operations is indirectly improved.

[0096] The methods for data analysis of precise biological data include:

[0097] The purpose of data analysis of precise biological data is to obtain the optimal environmental parameter combination and apply this optimal environmental parameter combination to the ginger planting environment; therefore, first use the model to process the precise biological data. The image features reflect the external growth state of ginger, and the environmental parameter combination reflects the internal influencing factors of ginger growth. The model combines the two to conduct a more comprehensive evaluation of the optimal growth state, and obtains the image data that can reflect the optimal growth situation and the corresponding suitable environmental parameter combination; for the suitable environmental parameter combination, then use the optimization algorithm to optimize it, and obtain the optimal environmental parameter combination by constructing an optimization objective function and minimizing this optimization objective function; apply the optimal environmental parameter combination to the ginger planting environment to adjust the growth environment of ginger.

[0098] Build an analysis model to process accurate biological data and obtain a suitable combination of environmental parameters; optimize the suitable combination of environmental parameters to obtain an optimal combination of environmental parameters.

[0099] The ways to build an analysis model include:

[0100] Take the CNN convolutional neural network model as the basic framework of the analysis model; collect historical accurate biological data (this data includes historical enhanced image data and historical normal index data; the acquisition methods of historical enhanced image data and historical normal index data are: collect historical image data and historical index data, add feature codes based on timestamps, perform denoising processing on historical image data to obtain historical enhanced image data, and perform missing value processing and outlier processing on historical index data in sequence to obtain historical normal index data) as the training set of the analysis model; construct a growth situation scoring function, calculate the growth situation score of each piece of historical accurate biological data, and take the historical accurate biological data with the highest growth situation score as the training label (the growth situation scoring function is the sum of the growth situation scores of historical enhanced image data and historical index data with the same feature code; the acquisition method of the growth situation score of historical index data is: perform weighted summation on all environmental parameters, judge the importance of each environmental parameter by querying the preset ginger planting knowledge base, and assign a greater weight to the environmental parameter with greater importance; the acquisition method of the growth situation score of historical enhanced image data is: judge the growth situation of ginger at this moment by comparing each historical enhanced image with the image data in the preset ginger planting knowledge base, such as checking the color and shape of the leaves, the size of the tubers, and whether there is rot or damage, and checking whether there are spots or abnormal colors caused by pests and diseases on the ginger; use OpenCV to extract features from the historical enhanced image data to obtain the growth features of each historical enhanced image; define a scoring standard based on the image data in the preset ginger planting knowledge base, set different score ranges and weights for health, tuber size, and disease conditions respectively, and calculate the growth situation score of each historical enhanced image based on the scoring standard); use the historical accurate biological data to train the analysis model, define an analysis loss function (such as the mean square error function, and the variable in this function is the growth situation score); calculate the function value of the analysis loss function until the function value of the analysis loss function no longer decreases, and fix the parameters at this time to obtain the trained analysis model; use the trained analysis model to process the accurate biological data, convert the normal index data and enhanced image data with the same feature code in the accurate biological data into an index vector and an image vector respectively, the index vector is the vector composed of environmental parameters, and the image vector is the vector composed of image features; combine the two vectors into an accurate biological vector, and the accurate biological vector and the growth situation score corresponding to this vector are the input data of the analysis model; the model output data also appears in vector form, restore the index vector and the image vector in it to the environmental parameter combination and image data respectively, and the environmental parameter combination at this time is the appropriate environmental parameter combination.

[0101] The ways to optimize the appropriate environmental parameter combination include:

[0102] Since optimizing the combination of appropriate environmental parameters is an optimization problem of high-dimensional continuous variables, and each environmental parameter has different values, forming a complex search space, the particle swarm algorithm is selected to optimize the combination of appropriate environmental parameters. Considering that ginger has different stages during the entire growth cycle and there may be mutual influences among environmental parameters, an optimization objective function is constructed as the fitness function of the particle swarm algorithm, continuously searching for feasible solutions in the search space to minimize the function value of the optimization objective function, and finally obtaining the optimal solution, which is the optimal combination of environmental parameters.

[0103] Initialize the parameters of the particle population, including the population size, the number of iterations, the best position of each particle individual, and the best position of the particle population. Each particle individual in the particle population represents a combination of appropriate environmental parameters.

[0104] Define the optimization objective function where NUM represents the maximum number of iterations; t represents the number of iterations; m represents the population size; ω p represents the weight of the p-th element in any particle individual; represents the p-th element of the l-th particle individual in the t-th iteration; h represents any particle individual (any particle individual can be represented as [h 1 , h 2 , h 3 , h 4 , h 5 ; where h 1 , h 2 , h 3 , h 4 and h 5 respectively represent the temperature, humidity, light intensity, carbon dioxide concentration, and soil pH value in any combination of appropriate environmental parameters); represents the ideal value of the p-th element of the l-th particle individual in the t-th iteration (this ideal value is adjusted based on the changes in the ginger growth cycle, and the ginger growth cycle is divided into the seedling stage and the tuber growth stage; the values of environmental parameters corresponding to the best growth state of ginger in different growth cycles obtained by querying the preset ginger planting knowledge base); R cdIt represents the interaction intensity between the c-th element and the d-th element in any particle individual (c, d ∈ [1, 5]; the interaction intensity refers to the strength of the interaction relationship between two environmental parameters that can influence each other in any combination of environmental parameters. For example, when the temperature rises, the humidity drops, or when the light intensity increases, the carbon dioxide concentration rises, etc.; the interaction intensity is assigned a value by querying the relevant field experience in the preset ginger planting knowledge base. For example, the interaction intensity between temperature and humidity is 0.8); the smaller the function value of the optimization objective function, the smaller the difference between the elements in the particle individual corresponding to this function value and the ideal state.

[0105] At each iteration, calculate the position of each particle individual and the function value of the optimization objective function at the current position of this particle. When the function value of the optimization objective function is less than the function value of the optimization objective function at the previous position of the particle individual, then take the current position as the new best position of the particle individual (at each iteration, calculate the velocity of the particle individual at the current iteration based on the best position of the particle individual and the best position of the particle population at the previous iteration. The calculation formula is: Among them, represents the velocity of the l-th particle individual at the (t + 1)-th iteration; ω V represents the inertia weight of the particle individual velocity, which is a variable constant with a value range of (0, 1) and is used to balance the global search and local search capabilities; represents the best position of the l-th particle individual at the t-th iteration; gBest t represents the best position of the particle population at the t-th iteration; represents the position of the l-th particle individual at the t-th iteration; c 1 and c 2 represent the acceleration factors, both of which are variable constants with a value range of [0, 4] and are used to control the speed at which the particle approaches the best position of the particle individual and the best position of the particle population respectively; e 1 and e 2 are both random constants with a value range of [0, 1)); after a round of iteration, take the best position of the particle individual where the particle with the smallest function value of the optimization objective function is located as the new best position of the particle population.

[0106] Repeat the iteration of the particle population until the maximum number of iterations is reached; select the particle individual that can make the function value of the optimization objective function reach the minimum from all particle individuals as the optimal particle individual, and this optimal particle individual is the optimal combination of environmental parameters.

[0107] Using the particle swarm algorithm for optimization is easy to implement and has high computational efficiency; the adaptability of this optimization algorithm is very strong, can well adapt to the complex requirements of environmental parameter combination optimization, and has strong robustness.

[0108] In this embodiment, image data and biological environment index data in the ginger planting environment are collected and combined into biological data. Then, the biological data is subjected to data cleaning and data analysis to obtain the optimal environmental parameter combination, which is applied to the ginger planting environment. An alarm is set to monitor the optimized ginger planting environment, realizing a big data analysis and monitoring method for the ginger planting environment. Compared with existing experience, the data collection is more extensive. Not only image data is collected to observe the growth of ginger, but also biological environment index data with the same time stamp as the image data is collected. The same feature code is added to the image data and biological environment index data with the same time stamp, facilitating the matching of the two types of data in subsequent operations. In terms of data cleaning, for the biological environment index data, missing value processing and outlier processing are respectively carried out using models and clustering algorithms, making the biological environment index data more complete, with values within the normal range. At the same time, manual intervention is reduced and the data quality is improved. For the image data, denoising processing is carried out using wavelet transform, and each image is sharpened and the contrast is enhanced, improving the quality of the image data and making the images clearer. In the process of data analysis, the optimal environmental parameter combination is obtained by constructing models and using optimization algorithms, and the ginger planting environment is adjusted based on the optimal parameter combination to optimize the ginger planting environment. Finally, an alarm is set to monitor the parameters of the optimized ginger planting environment. If the parameters exceed the preset range, a warning message is automatically sent, realizing the function of monitoring the ginger planting environment.

[0109] Embodiment 2;

[0110] Please refer to Figure 2 As shown, for the parts not described in detail in this embodiment, refer to the description content of Embodiment 1. A big data analysis and monitoring system for the ginger planting environment is provided, including:

[0111] A data collection module, used to collect biological data in the ginger planting environment, where the biological data includes image data and biological environment index data;

[0112] A data cleaning module, used to clean the biological data to obtain accurate biological data;

[0113] A data analysis module, used to analyze the accurate biological data to obtain the optimal environmental parameter combination; adjust each environmental parameter in the ginger planting environment based on the optimal environmental parameter combination;

[0114] An environmental monitoring module, sets a warning device to monitor the adjusted ginger planting environment. When the environmental parameters of the adjusted ginger planting environment exceed the preset safety range, a warning message is automatically sent to the preset database; each module is connected by wired and / or wireless means.

[0115] Embodiment 3

[0116] This embodiment publicly provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, it implements the operation mode of the above-provided big data analysis and monitoring method for ginger planting environment.

[0117] Since the electronic device introduced in this embodiment is the electronic device used to implement the big data analysis and monitoring method for ginger planting environment in the embodiments of the present application, based on the big data analysis and monitoring method for ginger planting environment introduced in the embodiments of the present application, those skilled in the art can understand the specific implementation manners and various variations of the electronic device in this embodiment. Therefore, the specific implementation of how this electronic device implements the method in the embodiments of the present application will not be described in detail here. As long as those skilled in the art implement the electronic device used in the big data analysis and monitoring method for ginger planting environment in the embodiments of the present application, it belongs to the scope protected by the present application.

[0118] The above formulas are all calculated by taking the numerical values after dimensionless. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain a formula closest to the actual situation. The preset parameters and threshold selection in the formulas are set by those skilled in the art according to the actual situation.

[0119] The above is only the preferred embodiment of the present invention. The protection scope of the present invention is not limited to the above embodiments. All technical solutions within the idea of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technical users in the technical field, several improvements and refinements made without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A ginger planting environment big data analysis and monitoring method, characterized in that: include: S1. Collect biological data in the ginger planting environment, the biological data includes image data and biological environment indicator data; S2. Clean the biological data to obtain accurate biological data; S3. Analyze the precise biological data to obtain the optimal environmental parameter combination; adjust various environmental parameters in the ginger planting environment based on the optimal environmental parameter combination; S4. An early warning device is set to monitor the adjusted ginger planting environment. When the environmental parameters of the adjusted ginger planting environment exceed the preset safety range, an early warning message is automatically sent to a preset database.

2. A ginger planting environment big data analysis and monitoring method according to claim 1, characterized in that, The biological data acquisition method includes: using a high-resolution camera to collect image data, equipped with a wide-angle lens and thermal imaging function, and selecting a suitable location in the ginger planting area for setting; by deploying a variety of sensors in the ginger planting area, respectively collecting corresponding types of environmental parameters, the environmental parameters include temperature, humidity, light intensity, carbon dioxide concentration and soil pH value; integrating all types of environmental parameters to obtain biological environmental indicator data; querying the timestamp of each image in the image data and the timestamp of each environmental indicator data point in the biological environmental indicator data, matching the image data and the biological environmental indicator data based on the timestamp, and adding the same feature code to the image data and the biological environmental indicator data that can be matched.

3. A ginger planting environment big data analysis and monitoring method according to claim 2, characterized in that, The method of cleaning biological data includes: Process missing values ​​of biological environmental indicator data to obtain complete indicator data; process outliers of complete indicator data to obtain normal indicator data; perform denoising on image data to obtain enhanced image data; integrate normal indicator data and enhanced image data to obtain accurate biological data; The methods for handling missing values ​​of biological environmental indicator data include: Construct a set of prediction models, and use the decision tree model as the basic framework of each prediction model; use the prediction model set to predict the missing values ​​of the biological environmental indicator data, and obtain the predicted value of each missing indicator data point in the biological environmental indicator data; use the predicted value to fill the missing value of each missing indicator data point to obtain complete indicator data; Ways to build a prediction model ensemble include: Collect historical complete indicator data as the training set of the prediction model; use different types of historical environmental parameters in the historical complete indicator data as training labels for different types of prediction models; train corresponding types of prediction models based on historical complete indicator data and different types of historical environmental parameters, define the corresponding loss function for each type of prediction model, calculate the function value of each loss function, and when the function value of any loss function no longer becomes smaller, fix the parameters to obtain any type of trained prediction model, and integrate all trained prediction models to obtain a prediction model set.

4. A ginger planting environment big data analysis and monitoring method according to claim 3, characterized in that, The method of processing outliers on the complete indicator data includes: All complete indicator data points belonging to any timestamp in the complete indicator data are combined into a multidimensional vector; all multidimensional vectors are clustered to obtain a grouped multidimensional vector set; the grouped multidimensional vector set is abnormally marked to obtain cluster differences, and abnormal values ​​are determined based on the cluster differences; abnormal values ​​are adjusted by querying a preset ginger planting knowledge base to obtain normal indicator data; The methods for clustering all multidimensional vectors include: Randomly select K multidimensional vectors as initial cluster centers; calculate the Euclidean distance between each multidimensional vector and any initial cluster center; if the Euclidean distance between any multidimensional vector and any initial cluster center is less than a preset distance threshold, the multidimensional vector and the initial cluster center are divided into one cluster; Calculate the Euclidean distance between any multidimensional vector and other multidimensional vectors in its cluster and take the average value to obtain the average Euclidean distance of the multidimensional vector; calculate the Euclidean distance between any multidimensional vector and the cluster center of other clusters and take the average value to obtain the average cluster distance of the multidimensional vector; based on the average Euclidean distance of the multidimensional vector and the average cluster distance, obtain the clustering coefficient of the multidimensional vector; calculate the clustering coefficients of all multidimensional vectors and take the average value to obtain the average clustering coefficient; The number of initial cluster centers is continuously adjusted, and the average clustering coefficient after each adjustment is calculated. When the average clustering coefficient reaches the maximum value, the number of clusters at this time is fixed to obtain the optimal number of cluster centers; the initial cluster centers are updated based on the optimal number of cluster centers, and clustering is performed again to obtain a set of grouped multidimensional vectors; the Euclidean distance between each multidimensional vector in the set of grouped multidimensional vectors and the cluster center of the cluster to which it belongs is calculated, and the multidimensional vectors with a Euclidean distance greater than a preset classification threshold are marked as abnormal, and the multidimensional vectors marked as abnormal are recorded as marked multidimensional vectors; the Euclidean distance between each element in each marked multidimensional vector and the cluster center of the cluster to which the marked multidimensional vector belongs is calculated to obtain the clustering difference of each element in each marked multidimensional vector, and the elements with clustering differences greater than a preset difference threshold are judged as abnormal values.

5. A ginger planting environment big data analysis and monitoring method according to claim 4, characterized in that: Methods for denoising image data include: Preset wavelet basis functions, and perform multi-scale wavelet decomposition on each image in the image data based on the preset wavelet basis functions; firstly perform low-pass and high-pass filtering in the horizontal direction to obtain the low-frequency component and high-frequency component of each image respectively, and then perform vertical filtering on the low-frequency component of each image to obtain the low-frequency sub-image and high-frequency sub-image of each image; perform horizontal and vertical filtering on the low-frequency sub-image of each layer again to obtain low-frequency sub-images and high-frequency sub-images of higher levels; repeat the multi-scale wavelet decomposition of each image until the maximum number of decomposition layers is reached, extract wavelet coefficients from the sub-images after each decomposition, and integrate all wavelet coefficients to obtain a wavelet coefficient set; perform threshold processing on the wavelet coefficient set to obtain denoised frequency domain data; use inverse wavelet transform to reconstruct the denoised frequency domain data to obtain denoised image data; Each denoised image in the denoised image data is grayed by using the RGB weighted average method to obtain gray image data; each gray image in the gray image data is Gaussian filtered to obtain filtered gray image data; an edge detection operator is used to calculate the pixel value change gradient of each filtered gray image in the filtered gray image data, and the edge features of each filtered gray image are extracted based on the pixel value change gradient; each filtered gray image is sharpened based on the edge features to obtain sharpened image data; each sharpened image in the sharpened image data is contrast enhanced to obtain enhanced image data.

6. A ginger planting environment big data analysis and monitoring method according to claim 5, characterized in that: Ways to perform threshold processing on a set of wavelet coefficients include: Preset a wavelet coefficient threshold, and construct an adaptive threshold function based on the wavelet coefficient threshold; update each wavelet coefficient in the wavelet coefficient set based on the adaptive threshold function to obtain an updated wavelet coefficient set, and obtain denoised frequency domain data; The calculation formula of the adaptive threshold function is: Among them, μi,j represents the wavelet coefficient of the jth translation position at the i-th level; represents the updated wavelet coefficient of the jth translation position at the i-th level; A represents a judgment function; γ represents the preset wavelet coefficient threshold; α and β are dynamic adjustment factors.

7. A ginger planting environment big data analysis and monitoring method according to claim 6, characterized in that: The method of performing image sharpening on each filtered grayscale image comprises: A rectangular coordinate system is constructed with the grayscale pixel point at the lower left corner of each filtered grayscale image as the origin; the Laplace operator of each filtered grayscale image is calculated to obtain an initial Laplace operator, and all initial Laplace operators are integrated to obtain a Laplace operator set; a sliding window much smaller than the size of each filtered grayscale image is constructed, and the sliding window is made to slide on any filtered grayscale image; within an area formed by the sliding window on any filtered grayscale image, the pixel value change gradient of all grayscale pixels in the area is calculated and the average value is taken to obtain the local average change gradient of the area; an operator adjustment coefficient is defined, and a change gradient threshold is preset; the size of the operator adjustment coefficient is adjusted based on the local average change gradient of the area and the preset change gradient threshold to obtain the operator adjustment coefficient of the area; The initial Laplace operator is updated based on the operator adjustment coefficient of the region to obtain the updated Laplace operator of the region; the updated Laplace operator of the region is used to perform a convolution operation with the region to complete the image sharpening of the region; the sliding window is translated along the coordinate axis direction to perform image sharpening on other regions until the sliding window traverses the entire filtered grayscale image, the image sharpening is completed, and a sharpened image is obtained; each filtered grayscale image is sharpened, and all sharpened images are integrated to obtain sharpened image data; The method of performing contrast enhancement processing on each sharpened image in the sharpened image data includes: A rectangular coordinate system is constructed with the sharpened pixel point at the lower left corner of each sharpened image as the origin; a histogram equalization process is performed on any sharpened image to obtain an equalized pixel value of any sharpened pixel point in any sharpened image; a pixel value change gradient of any sharpened pixel point is calculated, and a weighting coefficient is calculated based on the pixel value change gradient of any sharpened pixel point; a contrast enhancement pixel value is calculated based on the equalized pixel value of any sharpened pixel point and the weighting coefficient of the sharpened pixel point; and a contrast enhancement pixel value of each sharpened pixel point of each sharpened image is calculated to obtain enhanced image data.

8. A ginger planting environment big data analysis and monitoring method according to claim 7, characterized in that: The method of performing data analysis on precise biological data includes: Construct an analytical model to process precise biological data and obtain a suitable combination of environmental parameters; optimize the suitable combination of environmental parameters to obtain the optimal combination of environmental parameters; Ways to build analytical models include: The CNN convolutional neural network model is used as the basic framework of the analysis model; historical precise biological data are collected as the training set of the analysis model; a growth scoring function is constructed to calculate the growth score of each piece of historical precise biological data, and the historical precise biological data with the highest growth score is used as the training label; the analysis model is trained using historical precise biological data, and the analysis loss function is defined; the function value of the analysis loss function is calculated until the function value of the analysis loss function no longer decreases, and the parameters are fixed at this time to obtain a trained analysis model.

9. A ginger planting environment big data analysis and monitoring method according to claim 8, characterized in that: The method for optimizing the suitable environmental parameter combination includes: Initialize the particle population parameters, which include population size, number of iterations, optimal position of individual particles and optimal position of particle population; each individual particle in the particle population represents a suitable combination of environmental parameters; Define the optimization objective function Among them, NUM represents the maximum number of iterations; t represents the number of iterations; m represents the population size; ωp represents the weight of the pth element in any individual particle; represents the pth element of the lth particle individual in the tth iteration; h represents any particle individual; represents the ideal value of the pth element of the lth particle individual at the tth iteration; Rcd represents the interaction strength between the cth element and the dth element in any particle individual; In each iteration, the position of each individual particle and the function value of the optimization objective function of the particle at the current position are calculated. When the function value of the optimization objective function is less than the function value of the optimization objective function of the particle at the previous position, the current position is taken as the new optimal position of the individual particle. After a round of iterations, the optimal position of the individual particle with the smallest function value of the optimization objective function is taken as the new optimal position of the particle population. Repeat the iteration of the particle population until the maximum number of iterations is reached; select the particle individual that can minimize the function value of the optimization objective function from all particle individuals as the optimal particle individual, and the optimal particle individual is the optimal environmental parameter combination.

10. A ginger planting environment big data analysis and monitoring system, which is used to implement a ginger planting environment big data analysis and monitoring method according to any one of claims 1 to 9, characterized in that: include: A data acquisition module is used to collect biological data in the ginger planting environment, and the biological data includes image data and biological environment indicator data; Data cleaning module, used to clean biological data to obtain accurate biological data; The data analysis module is used to analyze precise biological data to obtain the optimal environmental parameter combination; based on the optimal environmental parameter combination, various environmental parameters in the ginger planting environment are adjusted; The environment monitoring module is provided with an early warning device to monitor the adjusted ginger planting environment. When the environmental parameters of the adjusted ginger planting environment exceed the preset safety range, an early warning message is automatically sent to a preset database. Each module is connected to each other by wired and / or wireless means.

Images