Method, device and storage medium for monitoring and predicting iron core fir insect pests

Abstract

This application discloses a method, device, and storage medium for monitoring and predicting pests in *Cephalotaxus fortunei*, relating to the field of pest monitoring technology. The method includes the following steps: collecting image data, environmental data, and pest trapping data from the *Cephalotaxus fortunei* growing area; standardizing and structuring the image data, environmental data, and pest trapping data to obtain structured image data, structured environmental data, and structured pest trapping data; inputting the structured image data into a pre-trained pest identification model to obtain *Cephalotaxus fortunei* plant status data; inputting the plant status data, structured environmental data, and structured pest trapping data into a pre-trained pest time-series prediction model to obtain a predicted pest evolution curve; and obtaining a predicted pest risk level based on the plant status data and the predicted pest evolution curve. This application enables objective analysis of the health status of *Cephalotaxus fortunei* plants and prediction of pest evolution.

Description

Technical Field

[0001] This application relates to the field of pest monitoring technology for ironwood, specifically a method, equipment, and storage medium for monitoring and predicting ironwood pests. Background Technology

[0002] Ironwood, as a rare native protection and timber species, possesses significant ecological and economic value. However, during its growth, it is susceptible to pests such as bagworms, twig rollers, longhorn beetles, and bark beetles, leading to problems like leaf drop, trunk and branch damage, stunted growth, and even complete tree death. These issues severely restrict the healthy growth and sustainable cultivation of ironwood. Traditional methods of monitoring ironwood pests rely on manual inspections, which have the following drawbacks: Manual inspections are highly subjective, have low monitoring frequency and limited coverage, and are difficult to capture the initial small-scale outbreak characteristics of pests, making it easy to miss the best control window. Furthermore, they cannot effectively analyze pest outbreaks based on pest distribution and the health status of ironwood plants, and pest management is often delayed after an outbreak occurs. Summary of the Invention

[0003] In view of the deficiencies in the existing technology, the technical problem to be solved by this application is: how to analyze the health status of ironwood plants and effectively predict pest outbreaks.

[0004] To achieve the above objectives, in a first aspect, embodiments of this application provide a method for monitoring and predicting pests in ironwood, the method comprising the following steps: Collect image data, environmental data, and pest trapping data of the growing area of ​​ironwood; The image data, environmental data, and pest trapping data are standardized and structured to obtain structured image data, structured environmental data, and structured pest trapping data. The image structured data is input into a pre-trained pest identification model to obtain the status data of ironwood plants; Input the plant status data, environmental structured data, and insect trapping quantity structured data of ironwood into a pre-trained insect pest time series prediction model to obtain the predicted insect pest evolution curve. Based on the plant status data of ironwood and the predicted pest evolution curve, the predicted pest risk level is obtained.

[0005] In conjunction with the first aspect, in one embodiment, the process of acquiring the image structured data includes: Images of ironwood branches and leaves were selected from the image data and retained as valid images; After performing size standardization and pixel normalization on the valid images, standardized image matrix data is obtained; By binding standardized image matrix data, timestamps of image data acquisition, and identifiers of image data acquisition areas, structured image data is obtained.

[0006] In conjunction with the first aspect, in one embodiment, the pre-trained pest identification model is constructed based on a convolutional neural network, with the objective loss function being the minimization of the sum of cross-entropy loss and detection box regression loss; The data on the condition of the ironwood plants include the percentage of withered and yellowed branches and leaves, the percentage of damaged branches and trunks, the percentage of area damaged by pests, and the quantitative level of pest damage.

[0007] In conjunction with the first aspect, in one embodiment, the environmental data includes air temperature and humidity, soil temperature and humidity, and light intensity; The process for acquiring the structured environmental data includes: The 3σ criterion was used to remove outliers in air temperature and humidity, soil temperature and humidity, and light intensity. After time-series alignment of air temperature and humidity, soil temperature and humidity, and light intensity, missing values ​​are filled in by the mean of adjacent time series, thus completing the preprocessing. The pre-processed air temperature and humidity, soil temperature and humidity, and light intensity were normalized to obtain environmental feature vectors. The environmental feature vectors are sorted according to a fixed field order, and the timestamps of environmental data collection and the identifiers of the environmental data collection areas are bound to obtain structured environmental data.

[0008] In conjunction with the first aspect, in one embodiment, the process for acquiring the structured data on pest trapping includes: Screen the total number of target pests from the pest trapping data; Based on the capture rate of the pre-calibrated pest trapping equipment and the total number of target pests, determine the stock of target pests in the corresponding collection area of ​​the pest trapping equipment; After binding the timestamps of the data collection on the target pest stock and the amount of pests trapped, and the identifier of the data collection area for the amount of pests trapped, structured data on the amount of pests trapped can be obtained.

[0009] In conjunction with the first aspect, in one embodiment, the pre-trained pest time-series prediction model is constructed based on an LSTM network, with the objective loss function being the minimization of mean squared error.

[0010] In conjunction with the first aspect, in one embodiment, the process of obtaining the predicted pest risk level based on the state data of the ironwood plant and the predicted pest evolution curve includes: The health index of ironwood plants was obtained by weighting and summing the percentage of yellowing branches and leaves, the percentage of damaged branches and trunks, the percentage of area damaged by pests, and the quantitative level of pest damage. Add one to the health index of the ironwood plant to get the health coefficient of the ironwood plant. Use the product of the health index of the ironwood plant and the health coefficient of the ironwood plant as the comprehensive pest risk index. The temporal growth rate and stock of target pests are quantified based on the predicted pest evolution curve. The environmental suitability coefficient is obtained by weighted summing of the normalized values ​​of pretreated air temperature and humidity, soil temperature and humidity, and light intensity. The comprehensive pest outbreak index is obtained by weighted summation of the health coefficient of ironwood plants, the temporal growth rate of target pests, the stock of target pests, and the environmental suitability coefficient. Based on the comprehensive pest risk index and the comprehensive pest outbreak index, the predicted pest risk level is obtained.

[0011] In conjunction with the first aspect, in one embodiment, the logic for determining the predicted pest risk level based on the comprehensive pest risk index and the comprehensive pest outbreak index includes: When the comprehensive pest risk index is 0 ≤ pest risk index < 0.6 and the comprehensive pest outbreak index is < 0.3, it is judged as no risk level; When the comprehensive pest risk index is between 0.6 and 1.2 and the comprehensive pest outbreak index is between 0.3 and 0.6, it is judged to be at a mild risk level. When the comprehensive pest risk index is between 1.2 and 1.6 and the comprehensive pest outbreak index is between 0.6 and 0.8, it is classified as a medium risk level. When the comprehensive pest risk index is ≥1.6 and the comprehensive pest outbreak index is ≥0.8, it is judged as a severe risk level.

[0012] Secondly, embodiments of this application provide a monitoring and prediction device for ironwood pests, the ironwood pest monitoring and prediction device including a processor, a memory, and an ironwood pest monitoring and prediction program stored in the memory and executable by the processor, wherein when the ironwood pest monitoring and prediction program is executed by the processor, the method provided in the first aspect is implemented.

[0013] Thirdly, embodiments of this application provide a computer-readable storage medium storing a monitoring and prediction program for pests of *Cephalotaxus fortunei*, wherein the monitoring and prediction program for pests of *Cephalotaxus fortunei* implements the method provided in the first aspect when executed.

[0014] Compared with the prior art, the advantages of this application are: (1) By collecting image data, the pest identification model is used to realize the quantitative analysis of the status of ironwood plants, providing objective and reliable data support for the health status analysis of ironwood plants. Combined with the collection of environmental data and pest trapping data, the pest evolution prediction model is used to realize the prediction of pest evolution, providing forward-looking data support for the pest management of ironwood, and effectively solving the problem of the lag of traditional manual inspection.

[0015] (2) First, the proportion of yellowing leaves, broken branches and trunks, insect damage area, and insect damage level are quantified by the pest identification model. Then, the health index and health level of the ironwood plant are calculated by weighting. This provides a dynamic ironwood plant health coefficient for predicting the pest risk level. Then, the plant status data, environmental structured data, and insect trapping quantity structured data are input into the pest time series prediction model to deduce the pest evolution trend. The plant status analysis is performed first and the pest trend prediction is performed later. This progressive analysis method realizes the comprehensive consideration of predicting the pest risk level and provides accurate and reliable data support for the pest control of ironwood. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart of the method for monitoring and predicting pests in ironwood in the embodiments of this application; Figure 2 This is a schematic diagram of the hardware structure of the ironwood pest monitoring and prediction device involved in the embodiments of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] The flowchart shown in the attached diagram is for illustrative purposes only and does not necessarily include all content and operations / steps, nor does it necessarily have to be performed in the order described. For example, some operations / steps can be broken down, combined, or partially merged, so the actual execution order may change depending on the actual situation.

[0020] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0021] Firstly, embodiments of this application provide a method for monitoring and predicting pests in ironwood, referring to... Figure 1 The method includes the following steps: S1. Collect image data, environmental data, and pest trapping data of the growing area of ​​ironwood. S2. Standardize and structure the image data, environmental data (air temperature and humidity, soil temperature and humidity, light intensity), and pest trapping data respectively to obtain structured image data, structured environmental data, and structured pest trapping data. S3. Input the image structured data into the pre-trained pest identification model to obtain the status data of the ironwood plant (proportion of withered and yellowed branches and leaves, proportion of damaged branches and trunks, proportion of pest-damaged area, and quantitative level of pest damage). S4. Input the plant status data, environmental structured data, and insect trapping quantity structured data of ironwood into the pre-trained insect pest time series prediction model to obtain the predicted insect pest evolution curve. S5. Based on the plant status data of ironwood and the predicted pest evolution curve, obtain the predicted pest risk level.

[0022] By collecting image data and using a pest identification model, the state of ironwood plants can be quantitatively analyzed, providing objective and reliable data support for the health status analysis of ironwood plants. Combined with the collection of environmental data and pest trapping data, a pest time-series prediction model can be used to predict pest evolution, providing forward-looking data support for pest management of ironwood and effectively solving the problem of the lag in traditional manual inspections.

[0023] In one embodiment, before collecting image data, environmental data, and pest trapping data of the growing area of ​​*Hemiberlesia lataniae*, the growing area of ​​*Hemiberlesia lataniae* is divided into regions, and corresponding labels are added to the divided regions. Camera equipment, sensor modules (temperature and humidity sensors, light intensity sensors), and pest trapping equipment are deployed in the divided regions with different labels to realize the collection of image data, environmental data, and pest trapping data.

[0024] Based on this, the process of acquiring structured image data includes: Images of ironwood branches and leaves were selected from the image data and retained as valid images; After performing size normalization (scaling the effective image to 640×640 pixels) and pixel normalization on the effective image, standardized image matrix data is obtained. The size normalization process involves scaling the effective image to 640×640 pixels; the pixel normalization calculation formula is as follows: ; In the formula, These are the normalized pixel values. These are the original pixel values; By binding standardized image matrix data, timestamps of image data acquisition, and identifiers of image data acquisition areas, structured image data is obtained.

[0025] Furthermore, the image structured data is input into a pre-trained pest identification model. The pre-trained pest identification model traverses a pre-stored database of ironwood-specific samples (containing standardized image feature vectors of ironwood plant withered yellowing labels, branch and trunk damage labels, and pest damage labels), and calculates the similarity between the sample and the image structured data (similarity algorithms such as Euclidean distance can be used). When the similarity is greater than a preset threshold (e.g., >0.8), the model outputs the percentage of withered yellowing leaves, the percentage of branch and trunk damage, the percentage of pest damage area, and the pest damage quantification level of the corresponding image structured data.

[0026] The percentage of withered and yellow leaves is the ratio of the total pixel area of ​​withered and yellow leaves to the total pixel area of ​​leaves of the ironwood plant; it is used to characterize the health of the plant's leaves, such as 0%-5% being healthy, 5%-15% being sub-healthy, and >15% being in a weak state.

[0027] The percentage of damaged branches is the ratio of the total pixel area of ​​damaged, bored, and insect-trail branches to the total pixel area of ​​branches of the ironwood plant; it is used to characterize the degree of damage to the integrity of branches, such as the quantitative range: 0%-3% is intact, 3%-10% is slightly damaged, and >10% is severely damaged.

[0028] The percentage of area damaged by pests is the ratio of the total area of ​​damage caused by insect infestation, gnawing, insect excrement contamination, and death to the total pixel area of ​​branches, leaves, and trunks in the ironwood plant: it is used to characterize the degree of direct damage caused by pests.

[0029] Based on this, the real-time pest damage level is determined according to the proportion of the damaged area. The criteria for this determination include: When the area damaged by pests is less than 5%, it is judged as a minor infestation, and the value is set to 0.1. When the percentage of insect-damaged area is less than 20% and 5% is considered moderate damage, the value is set to 0.5. When the area damaged by pests accounts for ≥20%, it is judged as a severe infestation, and the value is set to 1.0.

[0030] Furthermore, the health index of the ironwood plant was obtained by weighting and summing the percentages of withered and yellowed branches and leaves, damaged branches and trunks, insect-damaged area, and the quantitative level of insect damage (with the sum of weight coefficients being 1). The closer the health index of the ironwood plant is to 1, the more severe the damage and the worse the health status of the ironwood plant.

[0031] Based on this, the health level of the ironwood plants is determined according to the plant health index. The determination logic includes: When the health index of ironwood plants is 0 ≤ 0.2, they are considered healthy plants, meaning there is no obvious insect damage and the branches and leaves are in good condition. When the health index of ironwood plants is 0.2 ≤ 0.4, they are judged to be sub-healthy plants, which have slight yellowing and damage, and no moderate or above insect damage. When the health index of ironwood plants is 0.4 ≤ 0.7, they are considered weak plants. At this time, there are obvious signs of yellowing and withering of branches and leaves, damage to branches and trunks, and moderate pest damage. When the health index of ironwood plants is 0.7 ≤ 1.0, the plants are considered to be diseased, indicating that there is extensive damage, severe insect infestation, and weak plant growth.

[0032] The above-mentioned pest identification model is built on a convolutional neural network (such as the YOLO model), and its training process includes: Images of insect eggs, larvae, adults, boreholes, insect tunnels, withered plants, and damaged branches were collected from different seasons, times, and light conditions in the ironwood forest area to serve as a pest identification dataset. The pest identification dataset is divided into a training set and a validation set according to a preset ratio (8:2); The pest identification model was trained using the training set and iteratively trained using the Adam optimizer with an initial learning rate of 0.001, a batch size of 16, and a maximum number of iterations of 100. The objective loss function was to minimize the sum of cross-entropy loss and detection box regression loss. After the target loss function converges, the recognition accuracy of the pest identification model is determined using the validation set. If the accuracy is greater than the preset threshold (≥98%), the pre-training of the pest identification model is completed; otherwise, the parameters are adjusted and the pest identification model is trained again using the training set.

[0033] The expression for the above objective loss function is: ; In the formula, For the total loss, The cross-entropy loss is calculated as follows: ; In the formula, For the total sample size, This represents the total number of pest categories. For real category labels, Predict class probabilities for the model; The regression loss for the detection boxes is calculated as follows: ; In the formula, To perform an intersection-union comparison between the predicted bounding box and the ground truth bounding box, It is the smallest bounding rectangle region containing both the predicted bounding box and the ground truth bounding box. To predict the detection box region, This represents the actual labeled area.

[0034] In one embodiment, the process for acquiring structured environmental data includes: The 3σ criterion was used to remove outliers in air temperature and humidity, soil temperature and humidity, and light intensity. After time-series alignment of air temperature and humidity, soil temperature and humidity, and light intensity, missing values ​​are filled in by the mean of adjacent time series, thus completing the preprocessing. The pre-processed air temperature and humidity, soil temperature and humidity, and light intensity were normalized to obtain environmental feature vectors. The environmental feature vectors are sorted according to a fixed field order, and the timestamps of environmental data collection and the identifiers of the environmental data collection areas are bound to obtain structured environmental data.

[0035] In one embodiment, the process for obtaining structured data on pest trapping includes: Screen the total number of target pests (such as bagworm, pine shoot leafroller, scarab beetle, two-striped pine longhorn beetle, bark beetle, and pine stem weevil) from the pest trapping data; Based on the capture rate of the pre-calibrated pest trapping equipment and the total number of target pests, determine the stock of target pests in the corresponding collection area of ​​the pest trapping equipment; After binding the timestamps of the data collection on the target pest stock and the amount of pests trapped, and the identifier of the data collection area for the amount of pests trapped, structured data on the amount of pests trapped can be obtained.

[0036] The method for determining the capture rate of the above-mentioned pre-calibrated pest trapping equipment is as follows: After releasing a known number of pests into a closed hemlock growing area, the number of pests captured by the pest trapping device per unit time is calculated to obtain the capture rate. After repeating the operation several times, the average value is taken as the capture rate of the pest trapping device, thus completing the pre-calibration of the pest trapping device.

[0037] In one embodiment, the pest time series prediction model is built based on an LSTM network, with the objective loss function being the minimization of mean squared error.

[0038] The pest time series prediction model uses known pest results from the ironwood growing area for at least 3 consecutive years, along with corresponding image and environmental data, as training data. The pre-training of the pest time series prediction model is completed when the target loss function converges and the prediction deviation is less than a preset threshold (≤10%).

[0039] In one embodiment, the process of obtaining the predicted pest risk level based on the state data of ironwood plants and the predicted pest evolution curve includes: Add one to the health index of the ironwood plant to get the health coefficient of the ironwood plant. Use the product of the health index of the ironwood plant and the health coefficient of the ironwood plant as the comprehensive pest risk index. The temporal growth rate and stock of target pests are quantified based on the predicted pest evolution curve. The environmental suitability coefficient is obtained by summing the normalized values ​​of pretreated air temperature and humidity, soil temperature and humidity, and light intensity (the sum of the weighting coefficients is 1). The pest outbreak comprehensive index is obtained by summing the weighted averages of the health coefficient of the ironwood plant, the sequential growth rate of the target pest, the stock of the target pest, and the environmental suitability coefficient (the sum of the weighted coefficients is 1); the higher the pest outbreak comprehensive index, the higher the risk of pest outbreak. The predicted pest risk level is obtained based on the comprehensive pest risk index and the comprehensive pest outbreak index. The determination logic includes: When the comprehensive pest risk index is 0 ≤ pest risk index < 0.6 and the comprehensive pest outbreak index is < 0.3, it is judged as no risk level. Regardless of the health status of the plant, the pest population grows slowly and there is no outbreak trend, so there is no risk of harm to ironwood. When the comprehensive pest risk index is 0.6 ≤ pest risk index < 1.2 and the comprehensive pest outbreak index is 0.3 ≤ pest outbreak index < 0.6, it is judged as a mild risk level. At this time, the pests increase slightly, and only sub-healthy, weak, and diseased plants are at slight risk of being damaged, while healthy plants are not significantly harmed. When the comprehensive pest risk index is 1.2 ≤ 1.6 and the comprehensive pest outbreak index is 0.6 ≤ 0.8, it is judged as a medium risk level. At this time, the pest population expands rapidly. Combined with the shortcomings of plant health, it causes significant damage to sub-healthy and weak plants, and diseased plants are severely damaged. When the comprehensive pest risk index is ≥1.6 and the comprehensive pest outbreak index is ≥0.8, it is judged as a severe risk level. At this time, the pest outbreak trend is significant throughout the region. After the health coefficient of the ironwood plants is amplified, there is severe irreversible damage to all ironwood plants of all health levels, and emergency prevention and control need to be initiated.

[0040] This approach first quantifies the proportion of yellowed leaves, damaged branches and trunks, and insect-damaged area using a pest identification model, along with the quantitative level of pest damage. Then, a weighted calculation of the health index and health level of the *Imperata cylindrica* plants is performed, providing a dynamic health coefficient for predicting pest risk levels. Next, plant status data, structured environmental data, and structured data on insect trapping quantities are input into a pest time-series prediction model to deduce pest evolution trends. This progressive analysis method, prioritizing plant status analysis followed by pest trend prediction, achieves a comprehensive assessment of pest risk levels. This provides accurate and reliable data support for pest control in *Imperata cylindrica*, enabling early detection, early warning, and early prevention, reducing manual inspection costs, effectively improving the accuracy and efficiency of pest control, and ensuring the healthy growth of *Imperata cylindrica* communities.

[0041] Secondly, embodiments of this application provide a monitoring and prediction device for pests of Hemiberlesia lataniae. The monitoring and prediction device for pests of Hemiberlesia lataniae can be a personal computer (PC), a laptop computer, a server, or other devices with data processing capabilities.

[0042] Reference Figure 2 , Figure 2 This is a schematic diagram of the hardware structure of the hemlock pest monitoring and prediction device involved in the embodiments of this application. In the embodiments of this application, the hemlock pest monitoring and prediction device may include a processor, a memory, a communication interface, and a communication bus.

[0043] The communication bus can be of any type and is used to interconnect the processor, memory, and communication interface.

[0044] The communication interface includes input / output (I / O) interfaces, physical interfaces, and logical interfaces used for interconnecting internal components of the ironwood pest monitoring and prediction equipment, as well as interfaces used for interconnecting the ironwood pest monitoring and prediction equipment with other devices (such as other computing devices or user equipment). Physical interfaces can be Ethernet interfaces, fiber optic interfaces, ATM interfaces, etc.; user equipment can be displays, keyboards, etc.

[0045] Memory can be various types of storage media, such as random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), flash memory, optical storage, hard disk, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), etc.

[0046] The processor can be a general-purpose processor, which can call the ironwood pest monitoring and prediction program stored in the memory and execute the ironwood pest monitoring and prediction method provided in the embodiments of this application. For example, the general-purpose processor can be a central processing unit (CPU). The method executed when the ironwood pest monitoring and prediction program is called can be referred to the various embodiments of the ironwood pest monitoring and prediction method of this application, and will not be repeated here.

[0047] Those skilled in the art will understand that Figure 2 The hardware structure shown does not constitute a limitation of this application and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0048] Thirdly, embodiments of this application also provide a computer-readable storage medium.

[0049] The computer-readable storage medium of this application stores a hammerwood pest monitoring and prediction program, wherein when the hammerwood pest monitoring and prediction program is executed by a processor, it implements the steps of the hammerwood pest monitoring and prediction method described above.

[0050] The method implemented when the ironwood pest monitoring and prediction program is executed can be referred to in the various embodiments of the ironwood pest monitoring and prediction method of this application, and will not be repeated here.

[0051] It should be noted that the sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0052] The terms "comprising" and "having," and any variations thereof, in the specification, claims, and accompanying drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus. The terms "first," "second," and "third," etc., are used to distinguish different objects, etc., and do not indicate a sequence, nor do they limit "first," "second," and "third" to different types.

[0053] In the description of the embodiments of this application, terms such as "exemplary," "for example," or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "exemplary," "for example," or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary," "for example," or "for instance" is intended to present the relevant concepts in a concrete manner.

[0054] In the description of the embodiments of this application, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. The "and / or" in the text is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. In addition, in the description of the embodiments of this application, "multiple" means two or more.

[0055] In some processes described in the embodiments of this application, multiple operations or steps are included in a specific order. However, it should be understood that these operations or steps may not be executed in the order they appear in the embodiments of this application, or they may be executed in parallel. The sequence number of the operation is only used to distinguish different operations, and the sequence number itself does not represent any execution order. In addition, these processes may include more or fewer operations, and these operations or steps may be executed sequentially or in parallel, and these operations or steps may be combined.

[0056] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device to execute the methods described in the various embodiments of this application.

[0057] The above are merely specific embodiments of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the protection scope of this application. Therefore, the protection scope of this application should be determined by the scope of the claims.

Claims

1. A method for monitoring and predicting pests in ironwood, characterized in that, The method includes the following steps: Collect image data, environmental data, and pest trapping data of the growing area of ​​ironwood; The image data, environmental data, and pest trapping data are standardized and structured to obtain structured image data, structured environmental data, and structured pest trapping data. The image structured data is input into a pre-trained pest identification model to obtain the status data of ironwood plants; Input the plant status data, environmental structured data, and insect trapping quantity structured data of ironwood into a pre-trained insect pest time series prediction model to obtain the predicted insect pest evolution curve. Based on the plant status data of ironwood and the predicted pest evolution curve, the predicted pest risk level is obtained.

2. The method for monitoring and predicting pests in ironwood according to claim 1, characterized in that, The process for acquiring the image structured data includes: Images of ironwood branches and leaves were selected from the image data and retained as valid images; After performing size normalization and pixel normalization on the effective images, normalized image matrix data is obtained; By binding standardized image matrix data, timestamps of image data acquisition, and identifiers of image data acquisition areas, structured image data is obtained.

3. The method for monitoring and predicting pests in ironwood according to claim 1, characterized in that, The pre-trained pest identification model is built on a convolutional neural network, with the objective loss function being to minimize the sum of cross-entropy loss and detection box regression loss. The data on the condition of the ironwood plants include the percentage of withered and yellowed branches and leaves, the percentage of damaged branches and trunks, the percentage of area damaged by pests, and the quantitative level of pest damage.

4. The method for monitoring and predicting pests in ironwood according to claim 3, characterized in that, The environmental data includes air temperature and humidity, soil temperature and humidity, and light intensity; The process for acquiring the structured environmental data includes: The 3σ criterion was used to remove outliers in air temperature and humidity, soil temperature and humidity, and light intensity. After time-series alignment of air temperature and humidity, soil temperature and humidity, and light intensity, missing values ​​are filled in using the average of adjacent time series to complete the preprocessing. The pre-processed air temperature and humidity, soil temperature and humidity, and light intensity were normalized to obtain environmental feature vectors. The environmental feature vectors are sorted according to a fixed field order, and the timestamps of environmental data collection and the identifiers of the environmental data collection areas are bound to obtain structured environmental data.

5. The method for monitoring and predicting pests in ironwood according to claim 4, characterized in that, The process for obtaining the structured data on pest trapping includes: Screen the total number of target pests from the pest trapping data; Based on the capture rate of the pre-calibrated pest trapping equipment and the total number of target pests, determine the stock of target pests in the corresponding collection area of ​​the pest trapping equipment; After binding the timestamps of the data collection on the target pest stock and the amount of pests trapped, and the identifier of the data collection area for the amount of pests trapped, structured data on the amount of pests trapped can be obtained.

6. The method for monitoring and predicting pests in ironwood according to claim 1, characterized in that, The pre-trained insect pest time-series prediction model is built on an LSTM network, with the objective loss function being the minimization of mean square error.

7. The method for monitoring and predicting pests in ironwood according to claim 6, characterized in that, The process of obtaining the predicted pest risk level based on the status data of ironwood plants and the predicted pest evolution curve includes: The health index of ironwood plants was obtained by weighting and summing the percentage of yellowing branches and leaves, the percentage of damaged branches and trunks, the percentage of area damaged by pests, and the quantitative level of pest damage. Add one to the health index of the ironwood plant to get the health coefficient of the ironwood plant. Use the product of the health index of the ironwood plant and the health coefficient of the ironwood plant as the comprehensive pest risk index. The temporal growth rate and stock of target pests are quantified based on the predicted pest evolution curve. The environmental suitability coefficient is obtained by weighted summing of the normalized values ​​of pretreated air temperature and humidity, soil temperature and humidity, and light intensity. The comprehensive pest outbreak index is obtained by weighted summation of the health coefficient of ironwood plants, the temporal growth rate of target pests, the stock of target pests, and the environmental suitability coefficient. Based on the comprehensive pest risk index and the comprehensive pest outbreak index, the predicted pest risk level is obtained.

8. The method for monitoring and predicting pests in ironwood according to claim 7, characterized in that, The logic for determining the predicted pest risk level based on the comprehensive pest risk index and the comprehensive pest outbreak index includes: When the comprehensive pest risk index is 0 ≤ pest risk index < 0.6 and the comprehensive pest outbreak index is < 0.3, it is judged as no risk level; When the comprehensive pest risk index is between 0.6 and 1.2 and the comprehensive pest outbreak index is between 0.3 and 0.6, it is judged to be at a mild risk level. When the comprehensive pest risk index is between 1.2 and 1.6 and the comprehensive pest outbreak index is between 0.6 and 0.8, it is classified as a medium risk level. When the comprehensive pest risk index is ≥1.6 and the comprehensive pest outbreak index is ≥0.8, it is judged as a severe risk level.

9. A monitoring and prediction device for pests in ironwood, characterized in that, The ironwood pest monitoring and prediction device includes a processor, a memory, and an ironwood pest monitoring and prediction program stored in the memory and executable by the processor, wherein when the ironwood pest monitoring and prediction program is executed by the processor, it implements the steps of the ironwood pest monitoring and prediction method as described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a pest monitoring and prediction program for ironwood, wherein when the ironwood pest monitoring and prediction program is executed, it implements the steps of the ironwood pest monitoring and prediction method as described in any one of claims 1 to 8.

Images