Cucurbit vegetable virus disease identification method and system based on sensor array construction

Abstract

The invention discloses a melon vegetable virus disease identification method and system constructed based on a sensor array. The method comprises the following steps: generating a dynamic thermodynamic diagram by collecting insect activity data of a target planting area, constructing a virus disease propagation distribution diagram in combination with an entomophilous propagation probability model, and deploying an initial sensor array according to the virus disease propagation distribution diagram; plant physiological parameters and environmental data are obtained in real time through multiple sensors, a virus disease recognition model fused with entomophilous activity characteristics is constructed, and early diagnosis of diseases is achieved; and evaluating the efficiency of the sensor array based on a recognition result, and dynamically adjusting the array layout by adopting a self-adaptive optimization algorithm to form an optimized network with balanced monitoring precision and resource efficiency. The virus disease detection sensitivity and accuracy are remarkably improved, the monitoring cost is reduced, and a digital solution is provided for prevention and control of diseases of melon crops.

Description

Technical Field

[0001] The present invention relates to the technical field of melon vegetable disease identification, and particularly to a method and system for identifying viral diseases of melon vegetables based on a sensor array. Background Art

[0002] With the advancement of agricultural modernization, especially in the cultivation of melon vegetables, the prevention and control of viral diseases have increasingly become a key factor affecting crop yield and quality. Traditional methods for monitoring and controlling viral diseases mainly rely on manual inspections and traditional detection techniques, and the accuracy and timeliness of detection results cannot be effectively guaranteed. Therefore, developing an efficient, accurate, and automated viral disease identification technology is of great significance for improving agricultural production efficiency and reducing losses.

[0003] Melon vegetables are often invaded by various viral diseases during the planting process, seriously affecting crop yield and quality. Research shows that insect vectors, especially aphids, whiteflies, etc., are one of the main vectors for the spread of viral diseases. These insects can carry the virus from diseased plants to healthy plants during feeding, resulting in the rapid spread of the virus in the field. Traditional methods for identifying viral diseases mostly rely on manual inspections and laboratory tests, with low efficiency and poor accuracy, and it is difficult to timely grasp the development trend of the disease.

[0004] In recent years, with the development of agricultural intelligence, disease monitoring technologies based on sensor arrays have been gradually applied to field management. By deploying spectral sensors, high-resolution imaging devices, etc., phenotypic characteristics and growth conditions of crops can be obtained, providing data support for disease identification. However, existing technologies generally ignore the spatial characteristics and dynamic laws of the vector-borne transmission of viral diseases, and the deployment of sensors lacks pertinence, and the identification accuracy and coverage efficiency need to be improved.

[0005] Therefore, constructing a method for identifying viral diseases that integrates insect vector activity information, transmission probability modeling, and sensor array optimization has become a key direction for current precision agriculture prevention and control. The present invention precisely aims at the above problems and proposes a method and system for identifying viral diseases of melon vegetables that integrates insect activity heat maps, transmission probability distributions, and intelligent recognition models, significantly improving the timeliness and accuracy of identification. Summary of the Invention

[0006] To solve the above at least one technical problem, the present invention proposes a method and system for identifying viral diseases of melon vegetables based on a sensor array.

[0007] In a first aspect of the present invention, a method for identifying viral diseases of melon vegetables based on a sensor array is provided, including:

[0008] Obtaining data on the insect activity situation in the target planting area of melon vegetables, and constructing an insect activity heat map of the target planting area according to the insect activity situation data;

[0009] Construct a probability distribution map of the vector-borne transmission of viral diseases in cucurbit vegetables in the target planting area according to the insect activity heat map, and construct an initial sensor array for the identification of viral diseases in cucurbit vegetables in the target planting area according to the probability distribution map of vector-borne transmission.

[0010] Obtain the viral disease monitoring data of each sensor, construct a viral disease identification model for cucurbit vegetables, and identify the viral disease monitoring data according to the viral disease identification model to obtain the identification result of cucurbit viral diseases.

[0011] Determine the identification quality of viral diseases of the initial sensor array according to the identification result, and adjust the initial sensor array according to the identification quality of viral diseases to obtain an optimized sensor array.

[0012] In this solution, the method for obtaining the insect activity data in the target planting area of cucurbit vegetables and constructing the insect activity heat map of the target planting area according to the insect activity data is as follows:

[0013] Obtain the historical insect pest record data of the target planting area of cucurbit vegetables, determine the types of active insects in the target planting area according to the historical insect pest record data, obtain the standard image data of the types of active insects, and perform insect name annotation on the standard image data to obtain annotated image data.

[0014] Construct an insect identification model based on a convolutional neural network, construct the convolutional layer, pooling layer and fully connected layer of the insect identification model, and construct a cross-entropy loss function and an Adam optimizer, and import the annotated image data into the insect identification model for training.

[0015] Divide the target planting area into N sub-regions of a preset size, obtain the image data of each sub-region in a preset time period based on a high-definition camera device, and import the image data into the insect identification model for insect identification to obtain the insect activity data of each sub-region. The insect activity data includes the type of insect and the number of insects.

[0016] Based on the kernel density estimation algorithm, regard each sub-region as an observation point, and the number of insects as the weight value of the observation point. Select a Gaussian kernel function and determine the bandwidth parameter of the Gaussian kernel function, and construct an insect density distribution map of a continuous space in the target planting area according to the kernel density estimation algorithm.

[0017] Perform color mapping on the insect density at each position according to the insect density distribution map, and construct an insect activity heat map of the target planting area in a preset time period.

[0018] In this solution, the probability distribution map of the vector-borne transmission of the viral disease of cucurbit vegetables in the target planting area is constructed based on the insect activity heat map, and the initial sensor array for the identification of the viral disease of cucurbit vegetables in the target planting area is constructed according to the probability distribution map of the vector-borne transmission, specifically as follows:

[0019] Obtain the data on the incidence of viral diseases in cucurbit vegetable plants in the target planting area during the preset time period. The incidence data includes the incidence location, the type of viral disease, the severity of the incidence, and the spread path of the incidence location;

[0020] Perform a spatial overlay analysis on the insect type and insect density in the insect activity heat map of each sub-region and the severity of the incidence in the corresponding sub-region, calculate the Pearson correlation coefficient between the density of each insect type and the degree of incidence, construct a correlation coefficient matrix, and screen out the target insect types with a correlation coefficient greater than the preset value for the transmission of the viral disease based on the correlation coefficient matrix;

[0021] Extract the distribution density data of the target insect types in the target planting area based on the insect activity heat map, and combine the direction characteristics of the spread path of the incidence location to construct a virus transmission and diffusion model based on the migration path of the target insects;

[0022] Calculate the virus transmission probability weight of each sub-region according to the virus transmission and diffusion model, where the sub-regions closer to the incidence location and with a higher density of target insects are given a higher transmission probability weight;

[0023] Perform spatial interpolation calculation on the transmission probability weights of each sub-region to generate a continuous probability surface covering the target planting area, divide the high, medium, and low transmission risk regions according to the probability surface threshold, and construct a probability distribution map of vector-borne transmission;

[0024] Determine the arrangement density of the hyperspectral imaging sensors in each sub-region of the target planting area according to the probability distribution map of the vector-borne transmission, determine the arrangement positions according to the arrangement density, and construct the initial sensor array for the identification of the viral disease of cucurbit vegetables in the target planting area according to the arrangement positions.

[0025] In this solution, the virus disease monitoring data of each sensor is obtained, a virus disease identification model for cucurbit vegetables is constructed, and the virus disease monitoring data is identified according to the virus disease identification model to obtain the identification result of the cucurbit virus disease, specifically as follows:

[0026] Obtain the appearance morphology data of diseased plants of different virus disease types at different growth stages, obtain the spectral reflection data of the appearance morphology, extract the spectral characteristics of the spectral reflection data based on the PCA algorithm, and integrate the spectral characteristics of the same virus disease type to construct a spectral feature vector of the diseased plants of the virus disease;

[0027] Label the spectral feature vectors with the types and severities of viral diseases to obtain the training set label spectral feature data;

[0028] Construct a viral disease classification model based on the support vector machine algorithm, select the Gaussian kernel function and define the hinge loss function and the L2 regularization term, and use the sequential minimal optimization algorithm to train the model with the training set label spectral feature data. Adjust the kernel function parameters and the regularization coefficient through cross-validation until the classification accuracy exceeds the preset threshold;

[0029] Deploy the trained viral disease classification model to each sensor node in each of the initial sensor arrays, obtain the spectral reflection data monitored by the sensors in real time and perform spectral feature extraction, input the extracted spectral features into the viral disease classification model to predict the types and severity levels of viral diseases, summarize the prediction results of all sensors and map them to the corresponding sub-regions, and generate the viral disease recognition results for the target planting area.

[0030] In this solution, determine the viral disease recognition quality of the initial sensor array according to the recognition results, and adjust the initial sensor array according to the viral disease recognition quality to obtain an optimized sensor array. Specifically:

[0031] Obtain the actual viral disease infection situation data of the melon and vegetable plants in each sub-region of the target planting area, compare the actual viral disease infection situation data with the viral disease recognition results, and judge the accuracy and missed detection rate of the initial sensor array for the viral diseases in each sub-region;

[0032] Evaluate the viral disease recognition quality of the melon and vegetable viral diseases in each sub-region according to the accuracy and missed detection rate of the viral disease recognition, and obtain the viral disease recognition quality data of the initial sensor array for each sub-region;

[0033] Perform clustering operations on the viral disease recognition quality data based on the BIRCH clustering algorithm, divide the target planting area into high recognition quality areas, medium recognition quality areas and low recognition quality areas, and generate a recognition quality clustering map of spatial distribution;

[0034] Perform spatial overlay analysis on the recognition quality clustering map and the insect vector transmission probability distribution map, extract the area ratio covered by the high recognition quality area within the high-risk area of insect vector transmission, and calculate the recognition coverage rate of the insect vector transmission risk area;

[0035] If the recognition coverage rate of the insect vector transmission risk area is less than the preset threshold, perform a sensor deployment defect analysis on the uncovered high-risk sub-regions of insect vector transmission, and obtain the average value of the viral transmission probability weights and the corresponding viral disease recognition quality scores of the defect sub-regions;

[0036] When the mean value of the virus transmission probability weights in the defective sub-region is greater than the preset risk threshold and the virus disease recognition quality score is lower than the preset quality threshold, it is determined as a sensor deployment failure area. The ant colony optimization algorithm is used to re-plan the sensor layout path in the failure area, and the hyperspectral imaging sensor density is increased at the peak point of the virus transmission probability weights to obtain the first optimization strategy;

[0037] When the mean value of the virus transmission probability weights in the defective sub-region is less than the preset risk threshold but the recognition quality score is lower than the preset quality threshold, it is determined as a sensor monitoring blind area. The monitoring angle and spectral acquisition frequency of the sensors in the blind area are adjusted according to the migration direction of the target insects in the insect activity heat map to obtain the second optimization strategy;

[0038] The initial sensor array is adjusted according to the first optimization strategy and the second optimization strategy to obtain an optimized sensor array.

[0039] In this solution, the clustering operation on the virus disease recognition quality data based on the BIRCH clustering algorithm is specifically as follows:

[0040] Perform data standardization processing on the virus disease recognition quality data, and convert the feature values of each dimension into a zero-mean and unit-variance distribution to obtain standardized virus disease data;

[0041] Introduce the BIRCH clustering algorithm, calculate the branching factor threshold according to the data dimension and sample number of the standardized virus disease data, and determine the maximum branching factor and sub-cluster diameter threshold of the clustering feature tree;

[0042] Initialize the clustering feature tree structure according to the maximum branching factor and sub-cluster diameter threshold, and create a root node containing an empty clustering feature vector. The clustering feature vector consists of the sample number, the linear summation vector, and the sum-of-squares scalar;

[0043] Traverse each data point in the standardized virus disease data, start from the root node and search downward along the tree structure, calculate the Euclidean distance between the current data point and the clustering feature vectors of each sub-cluster, and select the sub-cluster path with the closest distance for recursive access;

[0044] If the radius of the sub-cluster in the leaf node does not exceed the sub-cluster diameter threshold, merge the current data point into the sub-cluster and update the clustering feature vector. Otherwise, create a new sub-cluster and check the branching factor limit. If the branching factor is exceeded, split the node to generate a new leaf node;

[0045] After all data points are inserted, perform a condensed clustering operation on the clustering feature tree, and merge the sub-clusters that meet the diameter threshold in adjacent leaf nodes to form a coarse-grained clustering;

[0046] Extract the clustering feature vectors of all sub - clusters based on the coarse - grained clustering results as a new data set, and perform global clustering operations using the hierarchical clustering algorithm. Calculate the distance matrix between sub - clusters and determine the final number of clusters by cutting the dendrogram;

[0047] Map the virus disease recognition quality data to the global clustering results according to the belonging relationship of the sub - clusters to obtain the clustering results of the virus disease recognition quality data.

[0048] In the second aspect of the present invention, there is also provided a recognition system for virus diseases of melon vegetables constructed based on a sensor array. The system includes: a memory and a processor. The memory includes a program for the method of recognizing virus diseases of melon vegetables constructed based on a sensor array. When the program for the method of recognizing virus diseases of melon vegetables constructed based on a sensor array is executed by the processor, the following steps are implemented:

[0049] Obtain the data on the insect activities in the target planting area of melon vegetables, and construct a heat map of insect activities in the target planting area according to the data on the insect activities;

[0050] Construct a probability distribution map of the vector - borne transmission of virus diseases of melon vegetables in the target planting area according to the heat map of insect activities, and construct an initial sensor array for recognizing virus diseases of melon vegetables in the target planting area according to the probability distribution map of the vector - borne transmission;

[0051] Obtain the virus disease monitoring data of each sensor, construct a recognition model for virus diseases of melon vegetables, and recognize the virus disease monitoring data according to the recognition model to obtain the recognition results of melon virus diseases;

[0052] Determine the recognition quality of the virus diseases of the initial sensor array according to the recognition results, and adjust the initial sensor array according to the recognition quality of the virus diseases to obtain an optimized sensor array.

[0053] The present invention discloses a method and a system for recognizing virus diseases of melon vegetables constructed based on a sensor array. The method generates a dynamic heat map by collecting insect activity data in the target planting area, constructs a virus disease transmission distribution map in combination with a vector - borne transmission probability model, and deploys an initial sensor array accordingly; obtains plant physiological parameters and environmental data in real - time through multiple sensors, constructs a virus disease recognition model integrating insect activity characteristics, and realizes early diagnosis of diseases; evaluates the effectiveness of the sensor array based on the recognition results, and dynamically adjusts the array layout using an adaptive optimization algorithm to form an optimized network with balanced monitoring accuracy and resource efficiency. It significantly improves the sensitivity and accuracy of virus disease detection, reduces the monitoring cost, and provides a digital solution for the prevention and control of melon crop diseases. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1Shows a flowchart of a method for identifying viral diseases of melon vegetables based on a sensor array according to the present invention;

[0055] Figure 2 Shows a flowchart of constructing a heat map of insect activities in the target planting area according to the present invention;

[0056] Figure 3 Shows a flowchart of obtaining the recognition result of melon viral diseases according to the present invention;

[0057] Figure 4 Shows a block diagram of a system for identifying viral diseases of melon vegetables based on a sensor array according to the present invention. Detailed implementation manners

[0058] In order to more clearly understand the above objects, features and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific implementation manners. It should be noted that, without conflict, the embodiments of the present application and the features in the embodiments may be combined with each other.

[0059] In the following description, many specific details are set forth in order to fully understand the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited by the specific embodiments disclosed below.

[0060] Figure 1 Shows a flowchart of a method for identifying viral diseases of melon vegetables based on a sensor array according to the present invention.

[0061] As Figure 1 shown, the first aspect of the present invention provides a method for identifying viral diseases of melon vegetables based on a sensor array, including:

[0062] S102, obtaining data on the insect activity situation in the target planting area of melon vegetables, and constructing a heat map of insect activities in the target planting area according to the data on the insect activity situation;

[0063] S104, constructing a probability distribution map of vector-borne transmission of viral diseases of melon vegetables in the target planting area according to the heat map of insect activities, and constructing an initial sensor array for identifying viral diseases of melon vegetables in the target planting area according to the probability distribution map of vector-borne transmission;

[0064] S106, obtaining virus disease monitoring data of each sensor, constructing a virus disease recognition model for melon vegetables, and identifying the virus disease monitoring data according to the virus disease recognition model to obtain a recognition result of melon viral diseases;

[0065] S108. Determine the recognition quality of viral diseases of the initial sensor array according to the recognition result, and adjust the initial sensor array according to the recognition quality of viral diseases to obtain an optimized sensor array.

[0066] It should be noted that based on the construction of the insect activity heat map, the distribution density and migration law of insect vectors can be captured, the potential source areas and diffusion directions of virus transmission can be accurately located, and the spatio-temporal limitations of traditional manual inspections can be broken through. Secondly, by combining the insect vector transmission probability distribution map to deploy the initial sensor array, hyperspectral imaging sensors are preferentially arranged in areas with high transmission probability, effectively enhancing the monitoring sensitivity of high-risk areas and avoiding the waste of resources caused by uniform layout and insufficient monitoring of key areas. Subsequently, through the integrated analysis of multi-source sensor data by the viral disease recognition model, the spectral reflection characteristics of plants and the insect vector activity parameters are synchronously analyzed to achieve refined discrimination of the type, infection stage and severity of viral diseases, significantly improving the classification accuracy. Finally, based on the spatial quality assessment of the recognition results, an intelligent optimization algorithm is used to dynamically adjust the sensor layout: increase the sensor density and optimize the monitoring parameters in areas with high transmission risk but low recognition quality, and reduce redundant devices in areas with low risk and high accuracy, so as to construct an adaptive monitoring network, while ensuring full coverage of the core areas of virus transmission, optimize the resource allocation efficiency.

[0067] Figure 2 The flowchart of constructing the insect activity heat map of the target planting area of the present invention is shown.

[0068] According to an embodiment of the present invention, the obtaining of the insect activity data of the target planting area of cucurbit vegetables and the construction of the insect activity heat map of the target planting area according to the insect activity data are specifically as follows:

[0069] S202. Obtain the historical insect pest record data of the target planting area of cucurbit vegetables, determine the types of active insects in the target planting area according to the historical insect pest record data, obtain the standard image data of the types of active insects, and perform insect name annotation on the standard image data to obtain annotated image data;

[0070] S204. Build an insect recognition model based on a convolutional neural network, build the convolutional layer, pooling layer and fully connected layer of the insect recognition model, and build a cross-entropy loss function and an Adam optimizer, and import the annotated image data into the insect recognition model for training;

[0071] S206. Divide the target planting area into N sub-areas of preset size, obtain the image data of each sub-area in a preset time period based on a high-definition camera device, import the image data into the insect recognition model for insect recognition, and obtain the insect activity data of each sub-area, where the insect activity data includes insect type and insect quantity;

[0072] S208, regarding each sub-region as an observation point based on the kernel density estimation algorithm, taking the number of insects as the weight value of the observation point, selecting a Gaussian kernel function and determining the bandwidth parameter of the Gaussian kernel function, and constructing a continuous-space insect density distribution map in the target planting area according to the kernel density estimation algorithm;

[0073] S210, performing color mapping on the insect density at each position according to the insect density distribution map, and constructing an insect activity heat map of the target planting area during a preset time period.

[0074] It should be noted that an insect type library is established based on historical pest situation data and standard image annotation. Combining the convolutional-pooling-fully connected layer structure of the convolutional neural network model (CNN) and cross-entropy loss function optimization, high-precision identification of insect species in complex lighting and occlusion environments in the field is realized, solving the problems of high misjudgment rate and poor timeliness existing in traditional manual counting. Secondly, by dynamically collecting image data in sub-regions through high-definition camera devices and inputting them into the trained insect recognition model, accurate distribution data of specific insect types and quantities in each sub-region can be obtained in real time, breaking through the spatial limitations of traditional sampling surveys. Further using the kernel density estimation algorithm, the discrete sub-region observation data is converted into a continuous-space insect density distribution map. Through the dynamic optimization of the Gaussian kernel function bandwidth parameter, the density gradient change and migration aggregation characteristics of insect vectors are accurately characterized, overcoming the deficiency of the traditional grid statistical method in expressing spatial continuity. Finally, an intuitive insect activity heat map is generated through color mapping, dynamically reflecting the peak period of insect vector activity and the core diffusion path, providing high spatio-temporal resolution insect vector dynamic data support for modeling the transmission probability of viral diseases, and forming a full-chain automated monitoring system from data collection to visual analysis.

[0075] According to an embodiment of the present invention, constructing a vector transmission probability distribution map of melon vegetable viral diseases in the target planting area according to the insect activity heat map, and constructing an initial sensor array for identifying melon vegetable viral diseases in the target planting area according to the vector transmission probability distribution map, specifically:

[0076] Obtaining the data on the incidence of viral diseases of melon vegetable plants in the target planting area during the preset time period, where the incidence data includes the incidence location, viral disease type, severity of incidence, and the spread path of the incidence location;

[0077] Performing spatial overlay analysis on the insect type, insect density in the insect activity heat map of each sub-region and the severity of incidence in the corresponding sub-region, calculating the Pearson correlation coefficient between the density of each insect type and the severity of incidence, constructing a correlation coefficient matrix, and screening out target insect types whose correlation coefficient with the transmission of viral diseases is greater than a preset value according to the correlation coefficient matrix;

[0078] Extract the distribution density data of the target insect type in the target planting area according to the heat map of insect activities, and combine the direction characteristics of the spread path of the disease incidence location to construct a virus transmission and diffusion model based on the migration path of the target insect;

[0079] Calculate the virus transmission probability weight of each sub-region according to the virus transmission and diffusion model, and assign a higher transmission probability weight to the sub-region that is closer to the disease incidence location and has a higher target insect density;

[0080] Perform spatial interpolation calculation on the transmission probability weights of each sub-region to generate a continuous probability surface covering the target planting area, and divide the high, medium, and low transmission risk areas according to the probability surface threshold to construct a probability distribution map of vector-borne transmission;

[0081] Determine the arrangement density of the hyperspectral imaging sensors for each sub-region in the target planting area according to the probability distribution map of vector-borne transmission, determine the arrangement positions according to the arrangement density, and construct an initial sensor array for the identification of viral diseases of cucurbit vegetables in the target planting area according to the arrangement positions.

[0082] It should be noted that in the target planting area of melon vegetables, due to the activities of insects, viruses can be transmitted to the target planting area. The transmission of viruses by insect vectors, which leads to the infection of melon vegetables with virus diseases, is an important virus transmission route. Therefore, through the spatial overlay analysis of the insect activity heat map and the virus disease incidence data, the target insect vector types significantly related to virus transmission are screened out, and the interference of non-transmission vector insects is eliminated to ensure the biological rationality of model construction. Secondly, combining the target insect vector density distribution and the characteristics of the disease incidence and spread path, a virus transmission and diffusion model is constructed to quantitatively evaluate the transmission probability weights of different sub-regions, breaking through the limitation of ignoring the spatio-temporal correlation between insect migration and disease spread in traditional uniform deployment. By spatial interpolation, a continuous probability surface is generated and the risk level is divided, converting the abstract insect activity data into a visualized virus transmission heat zone map to accurately identify the virus diffusion front area and potential high-incidence infection areas. Based on this, it dynamically guides the differential deployment of hyperspectral imaging sensors: encrypting sensor nodes in high-transmission probability areas to enhance the spectral data acquisition density, and sparsely arranging them in low-risk areas to reduce resource redundancy, forming a monitoring network highly matching the virus transmission path. This solution not only improves the capture ability of the sensor array for the early infection characteristics of virus diseases, but also provides real-time data support for blocking the virus diffusion path by dynamically tracking the insect vector-driven transmission trend. The virus transmission and diffusion model is based on the density distribution data of target insects in the planting area, and a continuous insect density field is generated by spatial interpolation method. At the same time, the spatial distribution characteristics of virus disease incidence points are systematically analyzed, and the dominant direction of disease spread is extracted. On this basis, the model comprehensively considers three key factors: insect density, the distance attenuation effect from the incidence point, and the direction matching degree. Among them, the distance attenuation is characterized by a non-linear attenuation function to represent the characteristic that virus transmission weakens with the increase of distance, and the direction matching degree is quantified by calculating the angle difference to measure the consistency between the transmission path and the main spread direction. Through a multi-factor weighted fusion algorithm, the above elements are integrated into a unified transmission probability calculation system, and a spatial anisotropy correction mechanism is introduced to reflect the transmission differences in different directions. The finally generated model can dynamically simulate the diffusion process of viruses through insect vectors in three-dimensional space. The insect vector transmission probability distribution map is a visualized heat map reflecting the transmission risk levels of virus diseases at different positions in the planting area.

[0083] Figure 3 The flowchart showing the recognition result of the melon virus disease obtained by the present invention is shown.

[0084] According to an embodiment of the present invention, the obtaining of the virus disease monitoring data of each sensor, constructing a virus disease recognition model for melon vegetables, and recognizing the virus disease monitoring data according to the virus disease recognition model to obtain the recognition result of the melon virus disease are specifically as follows:

[0085] S302. Obtain the appearance morphological data of diseased plants of different virus disease types at different growth stages, obtain the spectral reflection data of the appearance morphology, extract the spectral features of the spectral reflection data based on the PCA algorithm, integrate the spectral features of the same virus disease type, and construct a spectral feature vector of diseased plants with virus diseases;

[0086] S304. Label the virus disease type and severity level for the spectral feature vector to obtain the training set label spectral feature data;

[0087] S306. Construct a virus disease classification model based on the support vector machine algorithm, select the Gaussian kernel function and define the hinge loss function and L2 regularization term, use the sequential minimal optimization algorithm to train the model with the training set label spectral feature data, and adjust the kernel function parameters and regularization coefficients through cross-validation until the classification accuracy exceeds the preset threshold;

[0088] S308. Deploy the trained virus disease classification model to each sensor node in each of the initial sensor arrays, obtain the spectral reflection data monitored by the sensors in real time and extract the spectral features, input the extracted spectral features into the virus disease classification model to predict the virus disease type and severity level, summarize the prediction results of all sensors and map them to the corresponding sub-regions, and generate the virus disease recognition result of the target planting area.

[0089] It should be noted that by constructing a virus disease classification model based on the support vector machine algorithm and using the kernel function to map the high-dimensional spectral data to the feature space, the problem of spectral non-linearly separable that is difficult to handle by traditional methods is effectively solved, and different virus disease types and their disease severity levels can be accurately distinguished. By adopting the multi-classification strategy and the ordered regression method, the system realizes the accurate discrimination of multiple virus diseases and their different disease stages, and is especially good at dealing with the difficult problem of identifying virus types with similar spectral features. The built-in regularization mechanism and soft margin classification design in the model significantly improve the anti-interference ability, enabling it to adapt to the data acquisition conditions in the complex field environment. At the same time, the model supports online learning and incremental update functions, and can continuously optimize the classification performance as new data accumulates. The appearance morphological data includes plant dwarfing, leaf chlorosis, mosaic, curling, deformity, retarded plant development, and fruit deformity; the spectral features include absorption peaks, reflection peaks, emission peaks, wavelength positions, intensities, widths, shapes, continuity, slopes, and characteristic spectral bands.

[0090] According to an embodiment of the present invention, the virus disease recognition quality of the initial sensor array is determined according to the recognition result, and the initial sensor array is adjusted according to the virus disease recognition quality to obtain an optimized sensor array, specifically:

[0091] Obtain the actual data on the viral disease infection status of cucurbit vegetable plants in each sub-region of the target planting area, compare the actual data on the viral disease infection status with the viral disease recognition results, and determine the accuracy and missed detection rate of the initial sensor array in recognizing viral diseases in each sub-region;

[0092] Evaluate the recognition quality of viral diseases in cucurbit vegetables in each sub-region based on the accuracy and missed detection rate of viral disease recognition, and obtain the viral disease recognition quality data of the initial sensor array for each sub-region;

[0093] Perform a clustering operation on the viral disease recognition quality data based on the BIRCH clustering algorithm, divide the target planting area into high recognition quality areas, medium recognition quality areas, and low recognition quality areas, and generate a recognition quality clustering map of spatial distribution;

[0094] Perform a spatial overlay analysis on the recognition quality clustering map and the insect vector transmission probability distribution map, extract the area ratio covered by the high recognition quality area within the high-risk area of insect vector transmission, and calculate the recognition coverage rate of the insect vector transmission risk area;

[0095] It should be noted that due to problems with the installation angle or installation position of the initial sensor array, the initial sensor array may not be able to fully monitor and recognize the entire insect vector transmission risk area. Performing a clustering operation on the viral disease recognition quality data based on the BIRCH clustering algorithm and constructing a recognition quality clustering map can achieve rapid hierarchical clustering of large-scale farmland monitoring data. By dynamically adjusting the threshold parameters of the clustering feature tree, the boundary ranges of high, medium, and low recognition quality areas can be accurately divided. The generated recognition quality clustering map can intuitively reflect the weak and strong areas of the monitoring efficiency of the sensor array. After performing a spatial overlay analysis in combination with the insect vector transmission probability distribution map, by calculating the recognition coverage rate of the insect vector transmission risk area, the effective coverage degree of the sensor monitoring network in the high-risk area can be quantitatively evaluated.

[0096] If the recognition coverage rate of the insect vector transmission risk area is less than the preset threshold, perform an analysis of the sensor deployment defects in the un-covered high-risk sub-areas of insect vector transmission, and obtain the average value of the viral transmission probability weights and the corresponding viral disease recognition quality scores of the defective sub-areas;

[0097] When the average value of the viral transmission probability weights in the defective sub-area is greater than the preset risk threshold and the viral disease recognition quality score is lower than the preset quality threshold, it is determined as a sensor deployment failure area. Use the ant colony optimization algorithm to re-plan the sensor layout path in the failure area, and increase the density of hyperspectral imaging sensors at the peak points of the viral transmission probability weights to obtain the first optimization strategy;

[0098] When the average value of the virus transmission probability weights in the defective sub-region is less than the preset risk threshold but the recognition quality score is lower than the preset quality threshold, it is determined as a sensor monitoring blind area, and the monitoring angle and spectral acquisition frequency of the sensors in the blind area are adjusted according to the migration direction of the target insects in the insect activity heat map to obtain the second optimization strategy;

[0099] Adjust the initial sensor array according to the first optimization strategy and the second optimization strategy to obtain an optimized sensor array.

[0100] It should be noted that when the average value of the virus transmission probability weights in the defective sub-region is greater than the preset risk threshold and the recognition quality score is lower than the preset quality threshold, it indicates that this region is on the core path of virus diffusion driven by insect vectors. However, due to insufficient sensor deployment density, misalignment between node distribution and the hot area of insect vector activities, or mismatching of device performance parameters, the high-risk transmission areas are not effectively monitored, forming key vulnerabilities in the monitoring network. Therefore, through the intelligent optimization mechanism of the ant colony optimization algorithm to simulate the foraging path of insects, a sensor layout path highly consistent with the virus transmission probability weight distribution is generated in the area where sensor deployment fails, and hyperspectral imaging sensor nodes are preferentially encrypted at the weight peak points at the forefront of virus diffusion. This strategy breaks through the rigid constraints of traditional grid layout, realizes the precise inclination of sensor resources to the core virus transmission channels, effectively improves the monitoring sensitivity and data integrity of high-risk areas, and blocks the risk of missed detection of virus diffusion caused by monitoring blind areas. When the average value of the virus transmission probability weights in the defective sub-region is lower than the risk threshold but the recognition quality still does not meet the standard, it reflects that although this region is not the main current virus diffusion path, due to deviations in sensor monitoring angles, mismatching between spectral acquisition frequency and the activity rhythm of insect vectors, or environmental interference (such as plant occlusion), the data quality deteriorates, forming a non-risk-oriented monitoring blind area. For the monitoring blind area of the non-main risk path, the system dynamically adjusts the sensor monitoring angle according to the migration direction of insect vectors revealed by the insect activity heat map (such as aligning the spectral acquisition perspective with the migration direction of the insect swarm), and simultaneously increases the spectral acquisition frequency to match the peak activity period of the target insects. This strategy significantly enhances the capture ability of sporadic transmission events and low-density insect vector activities through the spatio-temporal coupling optimization of sensor parameters and insect vector behavior characteristics, reduces data loss caused by device response lag or monitoring direction deviation, and realizes the balanced improvement of the monitoring efficiency of the entire region.

[0101] According to the embodiments of the present invention, the clustering operation on the virus disease recognition quality data based on the BIRCH clustering algorithm is specifically as follows:

[0102] Perform data standardization processing on the virus disease recognition quality data, convert the feature values of each dimension into a zero-mean and unit-variance distribution to obtain standardized virus disease data;

[0103] Introduce the BIRCH clustering algorithm, calculate the branching factor threshold according to the data dimension and sample number of the standardized virus disease data, and determine the maximum branching factor and sub-cluster diameter threshold of the clustering feature tree;

[0104] Initialize the clustering feature tree structure according to the maximum branching factor and sub-cluster diameter threshold, create a root node containing an empty clustering feature vector, and the clustering feature vector is composed of the sample number, linear summation vector and sum of squares scalar;

[0105] Traverse each data point in the standardized virus disease data, start from the root node and search downward along the tree structure, calculate the Euclidean distance between the current data point and the clustering feature vectors of each sub-cluster, and select the sub-cluster path with the closest distance for recursive access;

[0106] If the radius of the sub-cluster in the leaf node does not exceed the sub-cluster diameter threshold, merge the current data point into the sub-cluster and update the clustering feature vector, otherwise create a new sub-cluster and check the branching factor limit. If the branching factor is exceeded, split the node to generate a new leaf node;

[0107] After inserting all data points, perform a condensed clustering operation on the clustering feature tree, and merge the sub-clusters that meet the diameter threshold in adjacent leaf nodes to form a coarse-grained clustering;

[0108] Extract the clustering feature vectors of all sub-clusters as a new data set based on the coarse-grained clustering result, perform a global clustering operation using the hierarchical clustering algorithm, calculate the distance matrix between sub-clusters, and determine the final number of clusters by cutting the dendrogram;

[0109] Map the virus disease recognition quality data to the global clustering result according to the belonging relationship of the sub-clusters to obtain the clustering result of the virus disease recognition quality data.

[0110] It should be noted that after eliminating the dimensional differences of multi-dimensional quality indicators through data standardization processing, introduce the BIRCH clustering algorithm to calculate the branching factor threshold and sub-cluster diameter threshold to construct a clustering feature tree, adaptively divide the data dense area and sparse area, and effectively solve the overfitting or underfitting problems of traditional algorithms for non-uniformly distributed data. Combine the recursive insertion and node splitting mechanisms to achieve incremental data clustering, significantly reduce the computational overhead in dynamic monitoring scenarios, and at the same time compress redundant sub-clusters through the condensed clustering operation, and combine the dendrogram cutting strategy of hierarchical clustering to accurately identify spatial heterogeneity features, forming a high-discrimination quality partition. The BIRCH clustering algorithm realizes the efficient processing of large-scale virus disease recognition quality data by constructing a clustering feature tree with dynamic branches. This algorithm uses the sub-cluster diameter threshold to automatically control the clustering granularity, effectively identifies the data distribution characteristics, and significantly reduces the computational complexity while ensuring accuracy.

[0111] Figure 4The block diagram of a recognition system for viral diseases of melon vegetables based on a sensor array according to the present invention is shown.

[0112] In a second aspect of the present invention, there is also provided a recognition system 4 for viral diseases of melon vegetables based on a sensor array. The system includes: a memory 41 and a processor 42. The memory includes a program for a recognition method of viral diseases of melon vegetables based on a sensor array. When the program for the recognition method of viral diseases of melon vegetables based on a sensor array is executed by the processor, the following steps are implemented:

[0113] Obtain data on the insect activity in the target planting area of melon vegetables, and construct a heat map of insect activity in the target planting area according to the data on the insect activity.

[0114] Construct a probability distribution map of the vector-borne transmission of viral diseases of melon vegetables in the target planting area according to the heat map of insect activity, and construct an initial sensor array for the recognition of viral diseases of melon vegetables in the target planting area according to the probability distribution map of the vector-borne transmission.

[0115] Obtain the monitoring data of viral diseases of each sensor, construct a recognition model of viral diseases of melon vegetables, and recognize the monitoring data of viral diseases according to the recognition model of viral diseases to obtain a recognition result of melon viral diseases.

[0116] Determine the recognition quality of the initial sensor array according to the recognition result, and adjust the initial sensor array according to the recognition quality of the viral diseases to obtain an optimized sensor array.

[0117] The present invention discloses a recognition method and system for viral diseases of melon vegetables based on a sensor array. The method generates a dynamic heat map by collecting insect activity data in the target planting area, constructs a distribution map of the spread of viral diseases in combination with a vector-borne transmission probability model, and deploys an initial sensor array accordingly; obtains plant physiological parameters and environmental data in real time through multiple sensors, constructs a recognition model of viral diseases integrating insect activity characteristics, and realizes early diagnosis of diseases; evaluates the effectiveness of the sensor array based on the recognition result, and dynamically adjusts the array layout by using an adaptive optimization algorithm to form an optimized network with balanced monitoring accuracy and resource efficiency. Significantly improve the sensitivity and accuracy of viral disease detection, reduce the monitoring cost, and provide a digital solution for the prevention and control of melon crop diseases.

[0118] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined, or can be integrated into another system, or some features can be ignored, or not executed. In addition, the couplings between the components shown or discussed, or direct couplings, or communication connections can be through some interfaces. The indirect couplings or communication connections of devices or units can be electrical, mechanical, or other forms.

[0119] The units described above as separate components may or may not be physically separated. The components shown as units may or may not be physical units. They can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0120] In addition, each functional unit in the embodiments of the present invention can be all integrated in a processing unit, or each unit can be separately used as a unit, or two or more units can be integrated in one unit. The above integrated units can be implemented in the form of hardware, or in the form of a combination of hardware and software functional units.

[0121] Those of ordinary skill in the art can understand that all or part of the steps to implement the above method embodiments can be completed by hardware related to program instructions. The foregoing program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps including the above method embodiments. The foregoing storage medium includes various media that can store program codes, such as removable storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0122] Alternatively, if the above integrated units of the present invention are implemented in the form of software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in the embodiments of the present invention. The foregoing storage medium includes various media that can store program codes, such as removable storage devices, ROM, RAM, magnetic disks, or optical discs.

[0123] As described above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims described above.

Claims

1. A method for identifying viral diseases of melon vegetables based on a sensor array, characterized in that, Including the following steps: Obtain data on the insect activity in the target planting area of cucurbit vegetables, and construct an insect activity heat map of the target planting area according to the insect activity data; Construct a probability distribution map of the vector-borne transmission of virus diseases of cucurbit vegetables in the target planting area according to the insect activity heat map, and construct an initial sensor array for the identification of virus diseases of cucurbit vegetables in the target planting area according to the probability distribution map of vector-borne transmission; Obtain the virus disease monitoring data of each sensor, construct a virus disease identification model for cucurbit vegetables, and identify the virus disease monitoring data according to the virus disease identification model to obtain the identification result of cucurbit virus diseases; Determine the virus disease identification quality of the initial sensor array according to the identification result, and adjust the initial sensor array according to the virus disease identification quality to obtain an optimized sensor array.

2. The method for identifying viral diseases of melon vegetables based on a sensor array according to claim 1, wherein The obtaining of the data on the insect activity in the target planting area of cucurbit vegetables and the construction of the insect activity heat map of the target planting area according to the insect activity data are specifically as follows: Obtain the historical insect situation record data of the target planting area of cucurbit vegetables, determine the types of active insects in the target planting area according to the historical insect situation record data, obtain the standard image data of the types of active insects, and perform insect name annotation on the standard image data to obtain annotated image data; Construct an insect identification model based on a convolutional neural network, construct the convolutional layer, pooling layer and fully connected layer of the insect identification model, and construct a cross-entropy loss function and an Adam optimizer, and import the annotated image data into the insect identification model for training; Divide the target planting area into N sub-regions of a preset size, obtain the image data of each sub-region in a preset time period based on a high-definition camera device, import the image data into the insect identification model for insect identification, and obtain the insect activity data of each sub-region, where the insect activity data includes the insect type and the number of insects; Regard each sub-region as an observation point based on the kernel density estimation algorithm, use the number of insects as the weight value of the observation point, select a Gaussian kernel function and determine the bandwidth parameter of the Gaussian kernel function, and construct an insect density distribution map of a continuous space in the target planting area according to the kernel density estimation algorithm; Perform color mapping on the insect density at each position according to the insect density distribution map, and construct an insect activity heat map of the target planting area in a preset time period.

3. The method for identifying viral diseases of melon vegetables based on a sensor array according to claim 2, wherein The construction of the probability distribution map of the vector-borne transmission of virus diseases of cucurbit vegetables in the target planting area according to the insect activity heat map and the construction of the initial sensor array for the identification of virus diseases of cucurbit vegetables in the target planting area according to the probability distribution map of vector-borne transmission are specifically as follows: Obtain the data on the incidence of virus diseases of cucurbit vegetable plants in the target planting area during the preset time period, where the incidence data includes the incidence location, virus disease type, severity of the disease, and the spread path of the incidence location; Perform a spatial overlay analysis of the insect type, insect density in the insect activity heat map of each sub-region and the disease severity of the corresponding sub-region, calculate the Pearson correlation coefficient between the density of each insect type and the disease severity, construct a correlation coefficient matrix, and select the target insect types with a correlation coefficient greater than the preset value related to the spread of viral diseases according to the correlation coefficient matrix; Extract the distribution density data of the target insect types in the target planting area according to the insect activity heat map, and construct a virus transmission and diffusion model based on the migration path of the target insects in combination with the direction characteristics of the spread path of the disease occurrence location; Calculate the virus transmission probability weight of each sub-region according to the virus transmission and diffusion model, where the sub-region closer to the disease occurrence location and with a higher density of the target insects is given a higher transmission probability weight; Perform spatial interpolation calculation on the transmission probability weights of each sub-region to generate a continuous probability surface covering the target planting area, divide the high, medium, and low transmission risk areas according to the probability surface threshold, and construct a probability distribution map of vector-borne transmission; Determine the arrangement density of the hyperspectral imaging sensors for each sub-region in the target planting area according to the probability distribution map of vector-borne transmission, determine the arrangement positions according to the arrangement density, and construct an initial sensor array for the identification of viral diseases in cucurbit vegetables in the target planting area according to the arrangement positions; 4. A method for identifying viral diseases of melon vegetables based on a sensor array according to claim 1, characterized in that, Obtain the virus disease monitoring data of each sensor, construct a virus disease identification model for cucurbit vegetables, and identify the virus disease monitoring data according to the virus disease identification model to obtain the identification result of cucurbit virus diseases, specifically: Obtain the appearance morphology data of diseased plants of different virus disease types at different growth stages, obtain the spectral reflection data of the appearance morphology, extract the spectral characteristics of the spectral reflection data based on the PCA algorithm, and integrate the spectral characteristics of the same virus disease type to construct a spectral feature vector of diseased plants with virus diseases; Label the spectral feature vector with the virus disease type and severity to obtain the training set label spectral feature data; Construct a virus disease classification model based on the support vector machine algorithm, select the Gaussian kernel function and define the hinge loss function and L2 regularization term, and use the sequential minimal optimization algorithm to train the model with the training set label spectral feature data, and adjust the kernel function parameters and regularization coefficients through cross-validation until the classification accuracy exceeds the preset threshold; Deploy the trained virus disease classification model to each sensor node in each of the initial sensor arrays, obtain the spectral reflection data monitored by the sensors in real time and perform spectral feature extraction, input the extracted spectral features into the virus disease classification model to predict the virus disease type and severity level, summarize the prediction results of all sensors and map them to the corresponding sub-regions to generate the identification result of virus diseases in the target planting area; 5. The melon vegetable virus disease recognition method based on a sensor array according to claim 1, characterized in that, Determine the virus disease identification quality of the initial sensor array according to the identification result, and adjust the initial sensor array according to the virus disease identification quality to obtain an optimized sensor array, specifically: Obtain the actual data on the viral disease infection status of melon and vegetable plants in each sub-region of the target planting area, compare the actual viral disease infection status data with the viral disease recognition results, and judge the accuracy and miss detection rate of the initial sensor array for the viral disease recognition in each sub-region; Evaluate the quality of the viral disease recognition of melon and vegetable in each sub-region according to the accuracy and miss detection rate of the viral disease recognition, and obtain the viral disease recognition quality data of the initial sensor array for each sub-region; Based on the BIRCH clustering algorithm, perform clustering operations on the viral disease recognition quality data, divide the target planting area into high recognition quality areas, medium recognition quality areas and low recognition quality areas, and generate a recognition quality clustering map of spatial distribution; Conduct a spatial overlay analysis of the recognition quality clustering map and the insect-borne transmission probability distribution map, extract the area ratio covered by the high recognition quality area within the high-risk area of insect-borne transmission, and calculate the recognition coverage rate of the insect-borne transmission risk area; If the recognition coverage rate of the insect-borne transmission risk area is less than the preset threshold, conduct a defect analysis of the sensor deployment for the uncovered high-risk sub-regions of insect-borne transmission, and obtain the average value of the viral transmission probability weights and the corresponding viral disease recognition quality scores of the defect sub-regions; When the average value of the viral transmission probability weights of the defect sub-region is greater than the preset risk threshold and the viral disease recognition quality score is lower than the preset quality threshold, it is determined as a sensor deployment failure area, and the ant colony optimization algorithm is used to re-plan the sensor layout path within the failure area, and increase the density of hyperspectral imaging sensors at the peak point of the viral transmission probability weight to obtain the first optimization strategy; When the average value of the viral transmission probability weights of the defect sub-region is less than the preset risk threshold but the recognition quality score is lower than the preset quality threshold, it is determined as a sensor monitoring blind area, and adjust the monitoring angle and spectral acquisition frequency of the sensors in the blind area according to the migration direction of the target insects in the insect activity heat map to obtain the second optimization strategy; Adjust the initial sensor array according to the first optimization strategy and the second optimization strategy to obtain an optimized sensor array.

6. The method for identifying viral diseases of melon vegetables based on a sensor array according to claim 5, characterized in that, The specific operation of performing clustering on the viral disease recognition quality data based on the BIRCH clustering algorithm is as follows: Perform data standardization processing on the viral disease recognition quality data, transform the feature values of each dimension into a zero-mean and unit-variance distribution to obtain standardized viral disease data; Introduce the BIRCH clustering algorithm, calculate the branching factor threshold according to the data dimension and sample number of the standardized viral disease data, and determine the maximum branching factor and sub-cluster diameter threshold of the clustering feature tree; Initialize the clustering feature tree structure according to the maximum branching factor and sub-cluster diameter threshold, create a root node containing an empty clustering feature vector, and the clustering feature vector is composed of the sample number, the linear summation vector and the sum of squares scalar; Traverse each data point in the standardized viral disease data, start from the root node and search downward along the tree structure, calculate the Euclidean distance between the current data point and the clustering feature vectors of each sub-cluster, and select the sub-cluster path with the closest distance for recursive access; If the radius of the sub-cluster in the leaf node does not exceed the sub-cluster diameter threshold, the current data point is merged into the sub-cluster and the clustering feature vector is updated; otherwise, a new sub-cluster is created and the branching factor limit is checked. If the branching factor is exceeded, the node is split to generate a new leaf node. After all data points are inserted, a clustering operation is performed on the clustering feature tree to merge sub-clusters that meet the diameter threshold in adjacent leaf nodes to form a coarse-grained clustering. Based on the coarse-grained clustering results, the clustering feature vectors of all sub-clusters are extracted as a new data set, and a hierarchical clustering algorithm is used for global clustering operations. The distance matrix between sub-clusters is calculated and the final number of clusters is determined by cutting the dendrogram. The virus disease recognition quality data is mapped to the global clustering results according to the belonging relationship of the sub-clusters to obtain the clustering results of the virus disease recognition quality data.

7. A recognition system for viral diseases of melon vegetables constructed based on a sensor array, characterized in that, The melon vegetable virus disease recognition system based on a sensor array includes a memory and a processor. The memory includes a program for the melon vegetable virus disease recognition method based on a sensor array. When the program for the melon vegetable virus disease recognition method based on a sensor array is executed by the processor, the following steps are implemented: Obtain the insect activity data in the target planting area of melon vegetables, and construct an insect activity heat map of the target planting area according to the insect activity data. Construct a probability distribution map of the vector-borne transmission of melon vegetable virus diseases in the target planting area according to the insect activity heat map, and construct an initial sensor array for the recognition of melon vegetable virus diseases in the target planting area according to the probability distribution map of the vector-borne transmission. Obtain the virus disease monitoring data of each sensor, construct a virus disease recognition model for melon vegetables, and identify the virus disease monitoring data according to the virus disease recognition model to obtain the recognition results of melon virus diseases. Determine the virus disease recognition quality of the initial sensor array according to the recognition results, and adjust the initial sensor array according to the virus disease recognition quality to obtain an optimized sensor array.

Images