An Artificial Intelligence-Based Method and System for Monitoring Crop Pests

Abstract

This invention discloses an artificial intelligence-based method and system for monitoring crop pests, comprising: constructing a crop pest identification model to identify pest types and quantities; acquiring multi-view fused infrared thermal imaging data of crops to identify pest types and quantities in unobstructed areas of the crop monitoring region; identifying thermal anomaly areas to determine the number of pests in obstructed areas of the crop monitoring region; constructing a pest type identification model to determine the pest types in obstructed areas of the crop monitoring region; calibrating pest quantity and type data in the crop monitoring region; constructing a pest prediction model to determine future pest spread trends; evaluating the effectiveness of pest control strategies and adjusting pest control strategies and the lead time for implementing pest control strategies. This invention achieves comprehensive pest monitoring, accurate identification, and scientific control, effectively solving the problems of monitoring blind spots and delayed control, and significantly improving the efficiency and reliability of pest control.

Description

Technical Field

[0001] This invention relates to the field of agricultural management, and in particular to a method and system for monitoring crop pests based on artificial intelligence. Background Technology

[0002] As global agriculture moves towards intelligent and precision farming, pest monitoring and control have become crucial factors influencing crop yield and quality. However, current pest monitoring technologies still face numerous challenges, failing to meet the demands of practical applications. Traditional monitoring methods primarily rely on single-view optical imaging or manual observation. This approach has significant limitations in complex farmland environments, especially in densely planted crops where overlapping leaves and stems create large blind spots, failing to capture pest activity in obscured areas and severely limiting data integrity and accuracy. Furthermore, images acquired from a single viewpoint have limited coverage, making it difficult to comprehensively reflect pest distribution within the monitored area, resulting in missing data for certain regions and impacting the effectiveness of subsequent pest control. Simultaneously, traditional pest monitoring technologies rely heavily on static data analysis, lacking the ability to capture and analyze dynamic pest activity characteristics, failing to reflect key dynamic features such as movement trajectories, crawling speeds, and activity cycles. This static monitoring method struggles to scientifically determine pest spread trends, leading to delayed pest control measures. Furthermore, in large-scale farmland monitoring, the use of multiple devices for data collection can lead to data redundancy and inconsistency, especially in overlapping or adjacent areas where multiple devices collect duplicate pest data, affecting the accuracy of pest quantity statistics. In addition, inconsistencies in coordinate systems between different devices make data integration difficult, and the lack of a unified calibration mechanism increases the complexity of data processing. On the other hand, existing pest control measures are mostly reactive, lacking effective prediction of pest spread trends and failing to provide precise intervention before outbreaks, thus missing the optimal control window. Simultaneously, the lack of standardized evaluation criteria for control measures makes it difficult to quantify the inhibitory effects of control strategies on pest numbers, density, and spread trends, resulting in a failure to optimize and adjust control strategies in a timely manner based on actual pest dynamics, affecting the timeliness and accuracy of pest control. Therefore, there is an urgent need for a new method that can comprehensively and accurately monitor pest data, dynamically analyze pest spread trends, and achieve scientific prediction and control decision-making to improve the efficiency and effectiveness of pest control and promote the development of smart agriculture. Summary of the Invention

[0003] To address the problems existing in the prior art, this invention provides an artificial intelligence-based method and system for monitoring crop pests.

[0004] A first aspect of this invention provides an artificial intelligence-based method for monitoring crop pests, mainly comprising:

[0005] Acquire infrared thermal imaging images of crops, construct a crop pest identification model, identify the types and quantities of pests in the crop monitoring area, and determine the density and activity area of ​​pest distribution;

[0006] By setting up infrared thermal imaging monitoring equipment at different locations in the crop monitoring area, multi-view fused infrared thermal imaging image data of crops is obtained, and the types and quantities of pests in the unobstructed areas of the crop monitoring area are identified.

[0007] Based on continuous time-frame infrared thermal imaging images acquired by thermal imaging equipment, thermal anomaly areas are identified, the density level of pest activity in different thermal anomaly areas is determined, and statistical results of the number of pests in the obscured areas of the crop monitoring area are obtained.

[0008] Based on the shape factor, temperature fluctuation frequency, spatial dynamic characteristics, and periodic characteristics of the thermal anomaly region, a pest type identification model is constructed to determine the pest types in the shaded areas of the crop monitoring area.

[0009] Based on the spatial coordinates of pest locations in the unobstructed and obstructed areas, it is determined whether there are overlapping or adjacent pest locations in the unobstructed and obstructed areas, and the pest quantity and type data in the crop monitoring area are calibrated, and pest control strategies are formulated.

[0010] Based on the data on the number, density and type of pests in the crop monitoring area, a pest prediction model is constructed to determine the future trend of pest spread, and pest control strategies are used to carry out pest control in the crop monitoring area in advance.

[0011] Based on the effectiveness of pest control strategies and the lead time for pest control, the accuracy of pest prediction models in assisting control decisions is assessed, and pest control strategies and the lead time for implementing pest control strategies are adjusted.

[0012] Furthermore, the acquisition of infrared thermal imaging images of crops, the construction of a crop pest identification model, the identification of pest types and quantities in the crop monitoring area, and the determination of pest density and activity areas include:

[0013] Infrared thermal images of crops in the monitored area are acquired using farmland infrared monitoring equipment, and the thermal images and their corresponding time and location information are stored in a crop monitoring database. Based on these images, data annotation tools are used to label the types, quantities, and thermal anomaly locations of pests, resulting in a labeled crop infrared thermal image dataset. This dataset is then augmented using data augmentation methods, including rotation, scaling, translation, and mirroring. Image preprocessing techniques include normalization, denoising, and histogram equalization. A convolutional neural network is used to train a crop pest identification model based on the augmented dataset. Finally, the model identifies the types and quantities of pests in the monitored area, labels their locations, determines their spatial coordinates, and assesses their density and activity areas.

[0014] Furthermore, the method of acquiring multi-view fused infrared thermal imaging image data of crops by using infrared thermal imaging monitoring devices installed at different locations within the crop monitoring area, and identifying the types and quantities of pests in unobstructed areas of the crop monitoring area, includes:

[0015] Infrared thermal imaging monitoring devices are installed at different locations within the crop monitoring area to acquire multi-view infrared thermal imaging image data of the same crop monitoring area from different directions. Based on the original infrared thermal imaging image set acquired from multiple directions, the SIFT algorithm is used to perform geometric correction and coordinate alignment on the images of each viewpoint. Through pixel-by-pixel error calculation and transformation matrix optimization, aligned multi-view thermal imaging image data under the same coordinate system is obtained. The registered multi-view infrared thermal imaging images are fused using an image fusion method to obtain comprehensive thermal map information of infrared thermal imaging data and thermal anomaly areas from different views of the same area, resulting in multi-view fused crop infrared thermal imaging image data. Based on the multi-view fused crop infrared thermal imaging image data, a crop pest identification model is used to identify the type and quantity of pests in the unobstructed areas of the crop monitoring area, mark the location of pests, determine the spatial coordinates of each pest, and judge the density and activity area of ​​pest distribution.

[0016] Furthermore, the step of identifying thermal anomaly areas based on continuous time-frame infrared thermal imaging images acquired by the thermal imaging device, determining the density level of pest activity in different thermal anomaly areas, and obtaining statistical results of the number of pests in the obscured areas of the crop monitoring area includes:

[0017] Based on continuous time-frame infrared thermal imaging images acquired by thermal imaging equipment, noise reduction and enhancement are performed on the original infrared images using preprocessing methods to obtain time-series infrared thermal imaging images. Using a Fast Fourier Transform algorithm, the fluctuation frequency and amplitude of the temperature value of each pixel in the infrared thermal imaging image over time are obtained. Regions with temperature fluctuation frequencies exceeding preset frequency values ​​and temperature fluctuation amplitudes exceeding preset amplitude thresholds are identified, determining the distribution of regions containing abnormal temperature fluctuations and obtaining a time-frequency distribution map of thermal anomaly regions. By extracting the spatial coordinates of temperature extreme points and thermal anomaly regions from the thermal image, and combining threshold segmentation and region growing algorithms, the results are analyzed. The heat map is binarized, and image areas with temperatures above a set first temperature threshold are marked as thermal anomaly regions, obtaining the independent contour and boundary coordinates of each thermal anomaly region. The thermal anomaly regions in the binarized image are clustered according to their density using the K-means clustering algorithm to determine the center point, area size, and density distribution of the thermal anomaly regions, obtaining independent thermal anomaly distribution cluster data. Based on the thermal anomaly distribution cluster data, the density level of pest activity in different thermal anomaly regions is determined by statistically analyzing the size, density, and number of hotspots, obtaining statistical results of the number of pests in the obscured areas of the crop monitoring area.

[0018] Furthermore, based on the shape factor, temperature fluctuation frequency, spatial dynamic characteristics, and periodic characteristics of the thermal anomaly region, a pest type identification model is constructed to determine the pest type in the shaded area of ​​the crop monitoring region, including:

[0019] Based on the extracted thermal anomaly region data, the movement path and speed of the thermal anomaly region in continuous time frames are calculated to obtain the region's movement speed and trajectory data, thus obtaining the spatial dynamic characteristics of pest activity, including crawling speed and movement patterns. Based on the duration of the extracted dynamic thermal signal, the window moving average method is used to identify the time variation length and period of the thermal signal, obtaining the temperature signal duration of each thermal anomaly region, and determining the periodic characteristics of different pest activities, including the shape factor. Among them, P i A represents the perimeter of the potential temperature anomaly region. i The area of ​​potential temperature anomaly regions is determined. Based on the shape factor, temperature fluctuation frequency, spatial dynamic characteristics, and periodic characteristics of the thermal anomaly regions, a dynamic feature vector dataset of heat maps is constructed and saved to the pest activity database. Based on the dynamic feature vector dataset of heat maps labeled with pest types in the pest activity database, a random forest algorithm is used to train the model and construct a pest type identification model. Based on the real-time monitored dynamic feature vectors of heat maps, the pest type identification model is used to determine the pest type of each thermal anomaly region and identify the pest type of the shaded area in the crop monitoring area.

[0020] Furthermore, based on the spatial coordinates of pest locations in the unobstructed and obstructed areas, the method determines whether there are overlapping or adjacent pest locations in the unobstructed and obstructed areas, calibrates the pest quantity and type data in the crop monitoring area, and formulates pest control strategies, including:

[0021] Spatial coordinate mapping is used to map the spatial coordinates of pest locations in both unobstructed and obstructed areas to a unified coordinate system within the same monitoring area. Affine or perspective transformations are employed to calculate the position of each coordinate point in this unified coordinate system, resulting in the mapped pest spatial coordinates and bounding boxes. Based on this mapped pest spatial coordinate data, the Intersection over Union (IoU) method is used to determine whether there are overlapping or adjacent pest locations between the unobstructed and obstructed areas, identifying the overlap in pest distribution, including pest quantity, bounding box position, and spatial relationships. Finally, based on the pest type identification results in both unobstructed and obstructed areas, cosine similarity is used to calculate the dynamic characteristics of the heatmap. The similarity of eigenvector data is used to determine whether the pest types in overlapping or adjacent areas are consistent, and to label the pest types in overlapping or adjacent areas. If the pest types in overlapping or adjacent areas are consistent, the number of pests in the overlapping or adjacent areas is calibrated by deduplication, including removing pest records with fewer pests in overlapping or adjacent areas and only retaining pest records with more pests. The number and type of pests in the crop monitoring area are recorded. Based on the number and type of pests in the crop monitoring area, pest control strategies are formulated, and pest control strategies are implemented in the crop monitoring area by using drones. Pest control strategies include using pesticide spraying, biological control, or physical control methods to prevent pests from spreading in advance.

[0022] Furthermore, based on data on the number, density, and type of pests in the crop monitoring area, a pest prediction model is constructed to determine the future spread trend of pests, and pest control strategies are used to prevent pests from spreading in the crop monitoring area in advance, including:

[0023] Continuously monitor the number, density, and type of pests in the crop monitoring area, use a long short-term memory network to train the model, construct a pest prediction model, predict the number, density, and type of pests in the crop monitoring area within a preset time period, and judge the future pest spread trend; if the predicted pest growth rate in the crop monitoring area within the preset time period is greater than a preset rate threshold, or the density is greater than a preset degree threshold, then pest control strategies are used to carry out pest control in the crop monitoring area in advance.

[0024] Furthermore, the process of judging the accuracy of the pest prediction model in assisting control decisions based on the effectiveness of pest control strategies and the lead time for pest control, and adjusting pest control strategies and the lead time for implementing pest control strategies, includes:

[0025] After obtaining data on changes in pest density, distribution, and type following pest control, the effectiveness of pest control strategies in suppressing pest numbers and spread is assessed by comparing post-control pest data with pre-control predictions. If the effectiveness is lower than a preset effectiveness threshold, the pest control strategy is adjusted. Using pest prediction models and trigger data for pest control strategies, the rationality and lead time of pest control are evaluated by comparing predicted pest growth rates with actual pest reduction rates. The accuracy of the pest prediction model in assisting control decisions is assessed, resulting in evaluation data on the lead time and timeliness of pest control strategies. If the accuracy is lower than a preset accuracy threshold, the lead time for implementing pest control strategies is adjusted.

[0026] A second aspect of this invention provides an artificial intelligence-based crop pest monitoring system, mainly comprising:

[0027] The crop pest analysis module is used to acquire infrared thermal imaging images of crops, build crop pest identification models, identify the types and quantities of pests in the crop monitoring area, and determine the density and activity area of ​​pest distribution.

[0028] The unobstructed area pest identification module is used to acquire multi-view fused infrared thermal imaging image data of crops by using infrared thermal imaging monitoring devices set at different locations in the crop monitoring area, and to identify the type and quantity of pests in the unobstructed areas of the crop monitoring area.

[0029] The module for analyzing the number of pests in the obscured area is used to identify thermal anomaly areas based on continuous time-frame infrared thermal imaging images acquired by thermal imaging equipment, determine the density level of pest activity in different thermal anomaly areas, and obtain statistical results of the number of pests in the obscured areas of the crop monitoring area.

[0030] The occluded area pest type identification module is used to construct a pest type identification model based on the shape factor, temperature fluctuation frequency, spatial dynamic characteristics, and periodic characteristics of the thermal anomaly area, and to determine the pest type in the occluded area of ​​the crop monitoring area.

[0031] The pest data calibration module is used to determine whether there are overlapping or adjacent pest locations in the unobstructed area and the obstructed area based on the spatial coordinates of pest locations in the unobstructed area and the obstructed area, and to calibrate the pest quantity and type data in the crop monitoring area and formulate pest control strategies.

[0032] The pest spread trend prediction module is used to build a pest prediction model based on the number, density and type of pests in the crop monitoring area, to judge the future pest spread trend, and to use pest control strategies to carry out pest control in the crop monitoring area in advance.

[0033] The pest control strategy adjustment module is used to assess the accuracy of pest prediction models in assisting pest control decisions based on the effectiveness of pest control strategies and the lead time for pest control, and to adjust pest control strategies and the lead time for implementing pest control strategies.

[0034] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0035] This invention provides an artificial intelligence-based method and system for monitoring crop pests. By employing multi-view fusion technology, this invention improves the comprehensiveness of monitoring unobstructed areas. Simultaneously, it utilizes continuous time-frame thermal imaging data to identify thermal anomaly areas and analyze the density and statistical quantity of pests in obstructed areas, effectively compensating for monitoring blind spots in obstructed areas. This invention further constructs a pest type identification model by combining the shape factor, temperature fluctuation frequency, dynamic characteristics, and periodic characteristics of pest areas, achieving accurate judgment of pest types within obstructed areas. This invention eliminates data overlap and redundancy through spatial coordinate calibration between unobstructed and obstructed areas, improving the accuracy and consistency of monitoring data. This invention uses pest quantity, density, and type data to construct a pest prediction model, judging pest spread trends. Combined with pest control strategies, it allows for precise early intervention, effectively preventing further pest spread and reducing crop losses. This invention also improves the timeliness and accuracy of pest control by dynamically adjusting control strategies and lead times based on the effectiveness of control strategies and the rationality of control timing. This achieves efficient closed-loop management of pest monitoring, prediction, and control, promoting the intelligent and precise development of smart agriculture management. Through multi-perspective fusion, dynamic thermal anomaly analysis, and pest prediction models, this invention enables comprehensive pest monitoring, accurate identification, and scientific control, effectively solving problems such as monitoring blind spots, data redundancy, and control lag. It significantly improves the efficiency and reliability of pest control, providing technical support for the development of smart agriculture. Attached Figure Description

[0036] Figure 1 This is a flowchart of an artificial intelligence-based crop pest monitoring method according to the present invention;

[0037] Figure 2 This is a flowchart of an artificial intelligence-based crop pest monitoring method according to the present invention;

[0038] Figure 3 This is a schematic diagram of an artificial intelligence-based crop pest monitoring system according to the present invention; Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0040] like Figure 1-2 This embodiment of an artificial intelligence-based crop pest monitoring method may specifically include:

[0041] Step S101: Obtain infrared thermal imaging images of crops, construct a crop pest identification model, identify the types and quantities of pests in the crop monitoring area, and determine the density and activity area of ​​pest distribution.

[0042] Infrared thermal images of crops in the monitored area are acquired using farmland infrared monitoring equipment, and the images, along with their corresponding time and location information, are stored in a crop monitoring database. Based on these images, data annotation tools are used to label the types, quantities, and thermal anomaly locations of pests, resulting in a labeled dataset. The labeled dataset is then augmented using data augmentation methods, including rotation, scaling, translation, and mirroring. Image preprocessing techniques include normalization, denoising, and histogram equalization. A convolutional neural network is used to train a crop pest identification model based on the augmented dataset. Using real-time acquired infrared thermal images of the monitored area, the pest identification model identifies the types and quantities of pests, labels their locations, determines their spatial coordinates, and assesses their density and activity areas.

[0043] For example, infrared thermal imaging images of the crop monitoring area are acquired using farmland infrared monitoring equipment. In a 5-hectare farmland area, the equipment captures a total of 500 infrared thermal images. These images, along with their capture time (June 1, 2020, 08:00-12:00) and corresponding geographic coordinates (X=123.456, Y=456.789), are stored in the crop monitoring database. Based on these infrared thermal images, data annotation tools are used to manually label the pest types (including rice planthoppers and rice leaf rollers), pest numbers (150 rice planthoppers and 50 rice leaf rollers), and the locations of thermal anomalies (X=123.46 to 123.50, Y=456.78 to 456.80), resulting in a labeled crop infrared thermal imaging dataset containing pest annotation information. Data augmentation methods, such as rotating the original images by 90 degrees, scaling them to a factor of 0.8, shifting them horizontally by 10 pixels, and vertically mirroring them, expanded 500 labeled images into 2000 augmented images. Image preprocessing techniques, such as normalization to bring image pixel values ​​to between 0 and 1, noise reduction to remove high-frequency noise, and histogram equalization to improve image brightness distribution, resulted in a data-augmented dataset of crop infrared thermal imaging images with optimized quality. A convolutional neural network was trained on this dataset, and after 20 training rounds, the model achieved a recognition accuracy of 95.2%. When real-time infrared thermal imaging images of the crop monitoring area are acquired, the trained crop pest identification model is used to analyze the images of the area. The pest types are identified as 100 rice planthoppers and 40 rice leaf rollers. The spatial coordinates of each pest are marked. The coordinates of the hot spot center of the rice planthopper are X=123.47, Y=456.79. The density of the pest distribution is also calculated. The average number of pests per square meter is 2.8. It is determined that the rice planthoppers are mainly concentrated in the southern area of ​​field A, while the rice leaf rollers are more dispersed.

[0044] Step S102: By using infrared thermal imaging monitoring devices set at different locations in the crop monitoring area, multi-view fused infrared thermal imaging image data of crops is obtained, and the types and quantities of pests in the unobstructed areas of the crop monitoring area are identified.

[0045] Infrared thermal imaging monitoring devices are installed at different locations within the crop monitoring area to acquire multi-view infrared thermal imaging image data of the same crop monitoring area from different perspectives. Based on the original infrared thermal imaging image set acquired from multiple perspectives, the SIFT algorithm is used to perform geometric correction and coordinate alignment of the images from each perspective. Through pixel-by-pixel error calculation and transformation matrix optimization, aligned multi-view thermal imaging image data under the same coordinate system is obtained. The registered multi-view infrared thermal imaging images are fused using an image fusion method to obtain comprehensive thermal map information of infrared thermal imaging data and thermal anomaly areas from different perspectives within the same area, resulting in multi-view fused crop infrared thermal imaging image data. Based on the multi-view fused crop infrared thermal imaging image data, a crop pest identification model is used to identify the type and quantity of pests in the unobstructed areas of the crop monitoring area, mark the location of pests, determine the spatial coordinates of each pest, and judge the density and activity area of ​​pest distribution.

[0046] For example, infrared monitoring devices were set up in four different locations in the crop monitoring area, such as east, south, west, and north, to acquire infrared thermal imaging image data of the same 2-hectare farmland area. Each device captured 100 infrared images, forming a total of 400 multi-view thermal imaging images. Based on these multi-view image data, the SIFT feature point matching algorithm was used to extract key feature points from the images, such as an average of 500 feature points extracted from each image. Then, through pixel-by-pixel error calculation, the error threshold was set to 0.5 pixels, and geometric correction and coordinate alignment were performed on the images from each view. After transformation matrix optimization, all images were finally aligned to the same coordinate system, resulting in spatially consistent registered multi-view infrared thermal imaging image data. By using weighted average fusion, the aligned images were fused together. Combining temperature data from different perspectives with information on thermal anomalies, a comprehensive heat map of the farmland area was obtained, showing the pest activity areas with hotspot temperatures ranging from 30.5℃ to 33.2℃. Using a crop pest identification model to analyze the fused heat map, the types of pests in the unobstructed areas were identified, such as 150 rice planthoppers and 60 rice leaf rollers. The spatial coordinates of these pests were also marked, such as the center coordinates of the rice planthopper hotspot area being X = 123.47, Y = 456.79. It was also determined that rice planthoppers were mainly distributed in the southwest area of ​​the farmland, with a density of 5 per square meter, while the activity area of ​​rice leaf rollers was relatively dispersed, with an average of 2 per square meter. The peak activity area of ​​rice planthoppers was determined to be in the southwest field.

[0047] Step S103: Based on the continuous time frame infrared thermal imaging images acquired by the thermal imaging device, identify thermal anomaly areas, determine the density level of pest activity in different thermal anomaly areas, and obtain the statistical results of the number of pests in the obscured areas of the crop monitoring area.

[0048] Based on continuous-time-frame infrared thermal imaging images acquired by thermal imaging equipment, noise reduction and enhancement are performed on the original infrared images using preprocessing methods to obtain time-series infrared thermal imaging images. Using a Fast Fourier Transform algorithm, the frequency and amplitude of temperature fluctuations of each pixel in the infrared thermal imaging image over time are obtained. Regions with temperature fluctuation frequencies exceeding preset frequency values ​​and amplitudes exceeding preset amplitude thresholds are identified, determining the distribution of temperature fluctuation anomalies and obtaining a time-frequency distribution map of thermal anomaly regions. By extracting the spatial coordinates of temperature extreme points and thermal anomaly regions from the thermal image, and combining threshold segmentation and region growing algorithms, the thermal image is binarized. Image regions with temperatures higher than a set first temperature threshold are marked as thermal anomaly regions, obtaining the independent contour and boundary coordinates of each thermal anomaly region. The K-means clustering algorithm is used to cluster the thermal anomaly regions in the binarized image according to their density, determining the center point, area size, and density distribution of the thermal anomaly regions, obtaining independent thermal anomaly distribution cluster data. Based on the data of thermal anomaly distribution clusters, the density level of pest activity in different thermal anomaly areas is determined by statistically analyzing the size, density, and number of hotspots of the thermal anomaly areas, and the statistical results of the number of pests in the obscured areas of the crop monitoring area are obtained.

[0049] For example, based on continuous time-frame infrared thermal imaging images acquired by a thermal imaging device, within a 1-hectare crop monitoring area, the device continuously captures 300 frames of infrared thermal imaging images at a rate of 1 frame per second, obtaining temperature data of each pixel changing over time. Image preprocessing methods are used to perform Gaussian filtering for noise reduction and contrast stretching to enhance the original thermal imaging images, removing environmental noise and improving details in hotspot areas. A fast Fourier transform algorithm is applied to the preprocessed time-series thermal image data to analyze the frequency and amplitude of temperature fluctuations of each pixel over time. Areas with temperature fluctuation frequencies greater than a preset value of 0.1 Hz and temperature fluctuation amplitudes greater than 2°C are extracted, identifying the spatial distribution of abnormal temperature fluctuation points and obtaining a time-frequency distribution map of thermal anomaly areas covering the entire monitoring area. Temperature extremes exceeding 32℃ and their corresponding spatial coordinates (X = 100.5, Y = 200.3) are extracted from the heatmap. Using a threshold segmentation method, all pixel regions with temperatures above 30℃ are marked as thermal anomaly regions. A region growing algorithm is then used to expand the marked regions, generating independent contours and boundary coordinates for each thermal anomaly region. For example, the boundary coordinates of a thermal anomaly region range from X = 100.0 to 105.0, and Y = 200.0 to 205.0. K-means clustering is applied to the binarized image data to divide these thermal anomaly regions into three clusters based on their density: cluster 1 has an average density of 10 thermal anomalies per square meter, cluster 2 has 5 thermal anomalies per square meter, and cluster 3 has 2 thermal anomalies per square meter. The center point of each cluster is determined; for example, the center point of cluster 1 has coordinates of X = 102.5, Y = 202.5. The area of ​​each cluster is calculated; for example, the area of ​​cluster 1 is 5 square meters. Based on the clustered thermal anomaly distribution data, the size, density, and number of hotspots of each cluster were statistically analyzed to determine the insect density level in the shaded area. For example, the insect density in cluster 1 was high-risk, with about 3 insects per square meter, for a total of 15 insects. Clusters 2 and 3 were medium- and low-risk areas, with a total of 12 insects. Finally, the statistical result of insects in the entire shaded area was 27.

[0050] Step S104: Based on the shape factor, temperature fluctuation frequency, spatial dynamic characteristics, and periodic characteristics of the thermal anomaly area, construct a pest type identification model to determine the pest type in the shaded area of ​​the crop monitoring area.

[0051] Based on the extracted thermal anomaly region data, the movement path and speed of the thermal anomaly regions in continuous time frames are calculated to obtain the region's movement speed and trajectory data, thus obtaining the spatial dynamic characteristics of pest activity, including crawling speed and movement patterns. Based on the duration of the extracted dynamic thermal signals, a window moving average method is used to identify the temporal variation length and period of the thermal signals, obtaining the temperature signal duration for each thermal anomaly region, and determining the periodic characteristics of different pest activities, including the shape factor. Among them, P i A represents the perimeter of the potential temperature anomaly region. i The area of ​​potential temperature anomaly regions is defined. Based on the shape factor, temperature fluctuation frequency, spatial dynamic characteristics, and periodic characteristics of the thermal anomaly regions, a dynamic feature vector dataset of heat maps is constructed and stored in the pest activity database. Using the dynamic feature vector dataset of heat maps labeled with pest types in the pest activity database, a random forest algorithm is used to train a model to construct a pest type identification model. Based on the real-time monitored dynamic feature vectors of the heat maps, the pest type identification model is used to determine the pest type of each thermal anomaly region, thus identifying the pest type in the obscured areas of the crop monitoring region.

[0052] For example, based on the extracted thermal anomaly area data, in a 2-hectare farmland monitoring area, the movement path and speed of the thermal anomaly areas were calculated through continuous time-frame infrared thermal imaging data analysis. The movement speed of one thermal anomaly area was found to be 0.15 m / s, and a corresponding trajectory was generated. The trajectory shows that its movement direction follows a linear crawling pattern, moving from coordinates X = 120.5, Y = 200.8 to X = 123.0, Y = 203.5. Based on the temporal variation of the dynamic thermal signal, the duration of the thermal signal was calculated using the window moving average method, determining that the temperature signal duration of this thermal anomaly area was 180 seconds, with a periodic characteristic of a temperature fluctuation peak occurring every 40 seconds. Based on the shape factor calculation formula... Calculate the shape factor φ of the thermal anomaly region i , where the perimeter P i =10.5m, area A i =8.7m 2The shape factor of the thermal anomaly region was calculated to be 1.12, indicating that the region has an irregular but nearly circular thermal anomaly shape. Combining the shape factor, temperature fluctuation frequency (0.025 Hz), movement speed, and periodicity characteristics of the thermal anomaly region, these data were used to construct a dynamic feature vector [0.15, 180, 1.12, 0.025], which was then saved to the pest activity database. Using the labeled dynamic feature vector dataset from the pest activity database, a random forest algorithm was employed to train the model, constructing a pest type identification model with a training accuracy of 93.5%. When the dynamic feature vector of the real-time monitored heat map [0.15, 180, 1.12, 0.025] is input into the pest type identification model, the model determines that the heat anomaly area belongs to the activity area of ​​rice planthoppers and identifies the pest type of the heat anomaly area as rice planthoppers. The model marks the location of the pest in the coordinate area X = 120.5 to X = 123.0, Y = 200.8 to Y = 203.5. Therefore, the pest type of the shaded area in the crop monitoring area is rice planthoppers.

[0053] Step S105: Based on the spatial coordinates of pest locations in the unobstructed area and the obstructed area, determine whether there are overlapping or adjacent pest locations in the unobstructed area and the obstructed area, calibrate the pest quantity and type data in the crop monitoring area, and formulate pest control strategies.

[0054] Spatial coordinate mapping is used to map the spatial coordinates of pest locations in both unmasked and masked areas to a unified coordinate system within the same monitoring area. Affine or perspective transformations are employed to calculate the position of each coordinate point in the unified coordinate system, resulting in the mapped pest spatial coordinates and bounding boxes. Based on the mapped pest location spatial coordinate data, the Intersection over Union (IoU) method is used to determine whether there are overlapping or adjacent pest locations between the unmasked and masked areas, identifying the overlap in pest distribution, including pest quantity, bounding box position, and spatial relationships. Based on the pest type identification results between the unmasked and masked areas, cosine similarity is used to calculate the similarity of the dynamic feature vector data in the heatmap, determining whether the pest types in the overlapping or adjacent areas are consistent, and thus assigning pest type labels to the overlapping or adjacent areas. If the pest types are consistent in overlapping or adjacent areas, the pest count in those areas is calibrated using a deduplication method. This involves removing pest records with low pest counts from the overlapping or adjacent areas, retaining only those with high pest counts, and recording the pest count and type data for the monitored crop area. Based on the pest count and type data for the monitored crop area, pest control strategies are developed and implemented using drones. These strategies include using pesticide spraying, biological control, or physical control methods for proactive pest prevention.

[0055] For example, through spatial coordinate mapping, the spatial coordinates of pest locations in the unobstructed area and the obstructed area are mapped to a unified coordinate system of the same monitoring area. For instance, the pest location coordinates in the unobstructed area are (X=120.5, Y=200.3) and (X=122.0, Y=202.5), and the pest coordinates in the obstructed area are (X=120.6, Y=200.4) and (X=122.1, Y=202.6). By using affine transformation, all coordinate points are transformed to the same coordinate system to obtain the unified coordinate points of the pests. Based on the mapped pest location data, the overlap of pest bounding boxes is calculated using the Intersection over Union (IoU) method. If the IoU overlap rate between the bounding boxes of the unmasked and masked areas is 0.75, it is determined that there is overlap in pest distribution between the two areas. The number of pests in the overlapping area is determined to be 7, and the bounding box locations cover coordinates X = 120.4 to 122.2 and Y = 200.2 to 202.8, and the spatial relationship is recorded. Based on the pest type identification results, the pest type in the unmasked area is identified as rice planthopper, and the pest type in the masked area is also identified as rice planthopper. The similarity of the dynamic feature vectors of the two heatmaps is calculated using the cosine similarity method, and the dynamic feature vector similarity reaches 97%, indicating that the pest type in the overlapping area is consistent. Because duplicate pest counts were recorded within overlapping areas, records with higher pest counts were retained after deduplication. For example, in overlapping areas, 5 pests were recorded in the unobstructed area and 4 in the obstructed area. After removing the few overlapping records, the total number of pests in the overlapping area was determined to be 5, and the total number of pests in the crop monitoring area was recorded as 23, with rice planthoppers accounting for the majority. Based on the pest count and type data, a pest control strategy was formulated, and pesticide spraying methods were selected to control the high-density areas of rice planthoppers. A drone was used to uniformly spray pesticides in the rice planthopper activity hotspots at coordinates X = 120.4 to 122.2 and Y = 200.2 to 202.8, completing the pest control task in the crop monitoring area.

[0056] Step S106: Based on the data on the number, density and type of pests in the crop monitoring area, construct a pest prediction model, determine the future pest spread trend, and use pest control strategies to carry out pest control in advance in the crop monitoring area.

[0057] Continuous monitoring of pest quantity, density, and type data in crop monitoring areas is conducted. A long short-term memory network is used to train a pest prediction model, which predicts the pest quantity, density, and type data in the crop monitoring area within a predetermined time period, and assesses future pest spread trends. If the predicted pest growth rate in the crop monitoring area within the predetermined time period exceeds a predetermined rate threshold, or the density exceeds a predetermined density threshold, pest control strategies are implemented to prevent pest infestation in the crop monitoring area in advance.

[0058] For example, by continuously monitoring pest data in a crop monitoring area, in a 3-hectare farmland, an infrared thermal imaging device acquired data on the number, density, and type of pests hourly throughout the day. The recorded number of pests increased from 50 at 8:00 AM to 150 at 2:00 PM, and the density increased from an average of 2 pests per square meter to 5. The main pest types were rice planthoppers and rice leaf rollers. This time-series data was input into a long short-term memory network for model training, and the model was used to predict the pest trend over the next 12 hours. The results showed that the number of pests would increase to 200 within the next 6 hours, with a growth rate of 8.3% per hour, exceeding the preset rate threshold of 5%. Simultaneously, the pest density was projected to reach 6 pests per square meter within the next 6 hours, exceeding the preset density threshold of 4 pests per square meter. Based on the forecast results, it was determined that the pest spread trend was quite serious and might further aggravate the damage. Therefore, pest control strategies were adopted to carry out early prevention and control in the crop monitoring area. Drones were selected to spray pesticides in areas with high pest density, such as coordinates X=120.5 to 125.5 and Y=200.0 to 205.0. Precision control was carried out in the hotspot distribution areas of rice planthoppers and rice leaf rollers to prevent the further spread of pests and reduce the threat of pests to crops.

[0059] Step S107: Based on the effectiveness of pest control strategies and the lead time for pest control, determine the accuracy of the pest prediction model in assisting control decisions, and adjust the pest control strategies and the lead time for implementing pest control strategies.

[0060] After pest control measures are implemented, changes in pest density, distribution, and type are obtained. By comparing post-control pest data with pre-control predictions, the effectiveness of the control strategy in suppressing pest numbers and spread is assessed, thus determining the strategy's effectiveness. If the effectiveness falls below a preset threshold, the control strategy is adjusted. Using pest prediction models and trigger data for control strategies, the rationality and lead time of pest control are evaluated by comparing predicted pest growth rates with actual pest reduction rates. The accuracy of the pest prediction model in assisting control decisions is assessed, yielding evaluation data on the lead time and timeliness of the control strategy. If the accuracy falls below a preset accuracy threshold, the lead time for implementing the control strategy is adjusted.

[0061] For example, in a 2-hectare farmland monitoring area, after implementing a pesticide spraying strategy using drones, the effectiveness of pest control was evaluated. Data on pest density, distribution, and type were obtained after treatment. The number of pests decreased from 200 before treatment to 120, and the average density decreased from 6 per square meter to 3.6. The main pest types remained rice planthoppers and rice leaf rollers. Comparing the post-treatment data with the pre-treatment model predictions, the predicted pest number was reduced to 90, and the density to 3 per square meter. The control strategy was assessed as having a 40% suppression rate, lower than the preset effectiveness threshold of 50%, indicating insufficient effectiveness and requiring adjustments to pesticide dosage or coverage. Simultaneously, using pest prediction model data and data triggered by the control strategy, the predicted pest growth rate of 8.3% per hour was compared with the actual pest reduction rate of 5% per hour to assess the rationality of the timing of control. The results showed that the control was implemented 2 hours in advance, but because the actual reduction rate was lower than expected, the pest spread suppression effect was not optimal. By analyzing the lead time and timeliness of pest control strategies, it was determined that the current pest prediction model has an accuracy of 85% in assisting control decisions, which is lower than the preset accuracy threshold of 90%. Therefore, the timing of control triggering needs to be advanced by 1 hour to ensure that control measures are implemented earlier to improve the control effect. The optimized strategy will be implemented in the next round of pest prediction and control.

[0062] like Figure 3 This embodiment discloses an artificial intelligence-based crop pest monitoring system, which may specifically include:

[0063] The crop pest analysis module is used to acquire infrared thermal imaging images of crops, build crop pest identification models, identify the types and quantities of pests in the crop monitoring area, and determine the density and activity area of ​​pest distribution.

[0064] The unobstructed area pest identification module is used to acquire multi-view fused infrared thermal imaging image data of crops by using infrared thermal imaging monitoring devices set at different locations in the crop monitoring area, and to identify the type and quantity of pests in the unobstructed areas of the crop monitoring area.

[0065] The module for analyzing the number of pests in the obscured area is used to identify thermal anomaly areas based on continuous time-frame infrared thermal imaging images acquired by thermal imaging equipment, determine the density level of pest activity in different thermal anomaly areas, and obtain statistical results of the number of pests in the obscured areas of the crop monitoring area.

[0066] The occluded area pest type identification module is used to construct a pest type identification model based on the shape factor, temperature fluctuation frequency, spatial dynamic characteristics, and periodic characteristics of the thermal anomaly area, and to determine the pest type in the occluded area of ​​the crop monitoring area.

[0067] The pest data calibration module is used to determine whether there are overlapping or adjacent pest locations in the unobstructed and obstructed areas based on the spatial coordinates of pest locations in the unobstructed and obstructed areas, and to calibrate the pest quantity and type data in the crop monitoring area, and to formulate pest control strategies.

[0068] The pest spread trend prediction module is used to construct a pest prediction model based on the number, density, and type of pests in the crop monitoring area, determine the future pest spread trend, and use pest control strategies to carry out pest control in the crop monitoring area in advance.

[0069] The pest control strategy adjustment module is used to assess the accuracy of pest prediction models in assisting pest control decisions based on the effectiveness of pest control strategies and the lead time for pest control, and to adjust pest control strategies and the lead time for implementing pest control strategies.

[0070] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the concept of this application. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A method for monitoring crop pests based on artificial intelligence, characterized in that, The method includes: Acquire infrared thermal imaging images of crops, construct a crop pest identification model, identify the types and quantities of pests in the crop monitoring area, and determine the density and activity area of ​​pest distribution; By setting up infrared thermal imaging monitoring equipment at different locations in the crop monitoring area, multi-view fused infrared thermal imaging image data of crops is obtained, and the types and quantities of pests in the unobstructed areas of the crop monitoring area are identified. Based on continuous time-frame infrared thermal imaging images acquired by thermal imaging equipment, thermal anomaly areas are identified, the density level of pest activity in different thermal anomaly areas is determined, and statistical results of the number of pests in the obscured areas of the crop monitoring area are obtained. Based on the shape factor, temperature fluctuation frequency, spatial dynamic characteristics, and periodic characteristics of the thermal anomaly region, a pest type identification model is constructed to determine the pest types in the shaded areas of the crop monitoring area. Based on the spatial coordinates of pest locations in the unobstructed and obstructed areas, it is determined whether there are overlapping or adjacent pest locations in the unobstructed and obstructed areas, and the pest quantity and type data in the crop monitoring area are calibrated, and pest control strategies are formulated. Based on the data on the number, density and type of pests in the crop monitoring area, a pest prediction model is constructed to determine the future trend of pest spread, and pest control strategies are used to carry out pest control in the crop monitoring area in advance. Based on the effectiveness of pest control strategies and the lead time for pest control, the accuracy of pest prediction models in assisting control decisions is assessed, and pest control strategies and the lead time for implementing pest control strategies are adjusted.

2. The method according to claim 1, wherein, The process of acquiring infrared thermal imaging images of crops, constructing a crop pest identification model, identifying the types and quantities of pests in the crop monitoring area, and determining the density and activity area of ​​pest distribution includes: Infrared thermal images of crops in the monitored area are acquired using farmland infrared monitoring equipment, and the thermal images and their corresponding time and location information are stored in the crop monitoring database. Based on these images, data annotation tools are used to label the types, quantities, and thermal anomaly locations of pests, resulting in a labeled crop infrared thermal image dataset. This dataset is then augmented using data augmentation methods, including rotation, scaling, translation, and mirroring. Image preprocessing techniques include normalization, denoising, and histogram equalization. Based on this augmented dataset, a convolutional neural network is used to train a crop pest identification model. Finally, using the real-time acquired infrared thermal images of the monitored area, the pest identification model identifies the types and quantities of pests, labels their locations, determines their spatial coordinates, and assesses their density and activity areas.

3. The method according to claim 1, wherein, The method involves acquiring multi-view fused infrared thermal imaging image data of crops using infrared thermal imaging monitoring devices positioned at different locations within the crop monitoring area, and identifying the types and quantities of pests in unobstructed areas of the crop monitoring area, including: Infrared thermal imaging monitoring devices are installed at different locations within the crop monitoring area to acquire multi-view infrared thermal imaging image data of the same crop monitoring area from different directions. Based on the original infrared thermal imaging image set acquired from multiple directions, the SIFT algorithm is used to perform geometric correction and coordinate alignment on the images of each viewpoint. Through pixel-by-pixel error calculation and transformation matrix optimization, aligned multi-view thermal imaging image data under the same coordinate system is obtained. The registered multi-view infrared thermal imaging images are fused using an image fusion method to obtain comprehensive thermal map information of infrared thermal imaging data and thermal anomaly areas from different views of the same area, resulting in multi-view fused crop infrared thermal imaging image data. Based on the multi-view fused crop infrared thermal imaging image data, a crop pest identification model is used to identify the type and quantity of pests in the unobstructed areas of the crop monitoring area, mark the location of pests, determine the spatial coordinates of each pest, and judge the density and activity area of ​​pest distribution.

4. The method according to claim 1, wherein, The process involves identifying thermal anomaly areas based on continuous time-frame infrared thermal imaging images acquired by thermal imaging equipment, determining the density level of pest activity within different thermal anomaly areas, and obtaining statistical results of the number of pests in the obscured areas of the crop monitoring region, including: Based on continuous time-frame infrared thermal imaging images acquired by thermal imaging equipment, the original infrared images are denoised and enhanced using preprocessing methods to obtain time-series infrared thermal imaging images. Using a fast Fourier transform algorithm, the fluctuation frequency and amplitude of the temperature value of each pixel in the infrared thermal imaging image over time are obtained. Regions with temperature fluctuation frequencies greater than a preset frequency value and temperature fluctuation amplitudes greater than a preset amplitude threshold are identified, determining the distribution of temperature fluctuation anomaly points and obtaining a time-frequency distribution map of thermal anomaly regions. By extracting the spatial coordinates of temperature extreme points and thermal anomaly regions in the thermal image, combined with threshold segmentation and region growing algorithms, the thermal image is binarized. Image regions with temperatures higher than a set first temperature threshold are marked as thermal anomaly regions, obtaining the independent contour and boundary coordinates of each thermal anomaly region. Using a K-means clustering algorithm, the thermal anomaly regions in the binarized image are clustered according to their density, determining the center point, area size, and density distribution of the thermal anomaly regions, obtaining independent thermal anomaly distribution cluster data. Based on the data of thermal anomaly distribution clusters, the density level of pest activity in different thermal anomaly areas is determined by statistically analyzing the size, density, and number of hotspots of the thermal anomaly areas, and the statistical results of the number of pests in the obscured areas of the crop monitoring area are obtained.

5. The method according to claim 1, wherein, The method involves constructing a pest type identification model based on the shape factor, temperature fluctuation frequency, spatial dynamic characteristics, and periodic characteristics of the thermal anomaly region to determine the pest types in the obscured areas of the crop monitoring area, including: Based on the extracted thermal anomaly region data, the movement path and speed of the thermal anomaly region in continuous time frames are calculated to obtain the region's movement speed and trajectory data, thus obtaining the spatial dynamic characteristics of pest activity, including crawling speed and movement patterns. Based on the duration of the extracted dynamic thermal signal, the window moving average method is used to identify the time variation length and period of the thermal signal, obtaining the temperature signal duration of each thermal anomaly region, and determining the periodic characteristics of different pest activities, including the shape factor. ,in, Perimeter of the potential temperature anomaly region. The area of ​​potential temperature anomaly regions is determined. Based on the shape factor, temperature fluctuation frequency, spatial dynamic characteristics, and periodic characteristics of the thermal anomaly regions, a dynamic feature vector dataset of heat maps is constructed and saved to the pest activity database. Based on the dynamic feature vector dataset of heat maps labeled with pest types in the pest activity database, a random forest algorithm is used to train the model and construct a pest type identification model. Based on the real-time monitored dynamic feature vectors of heat maps, the pest type identification model is used to determine the pest type of each thermal anomaly region and identify the pest type of the shaded area in the crop monitoring area.

6. The method according to claim 5, wherein, The method involves determining whether there are overlapping or adjacent pest locations in the unobstructed and obstructed areas based on the spatial coordinates of pest locations in the unobstructed and obstructed areas, calibrating the pest quantity and type data in the crop monitoring area, and formulating pest control strategies, including: By mapping spatial coordinates, the spatial coordinates of pest locations in the unobstructed area and the obstructed area are mapped to a unified coordinate system in the same monitoring area; By using affine transformation or perspective transformation, the position of each coordinate point in a unified coordinate system is calculated, and the spatial coordinates and bounding box of the pests mapped to the same coordinate system are obtained. Based on the spatial coordinate data of pest locations mapped to the same coordinate system, the IoU (Intersection over Union) method is used to determine whether there are overlapping or adjacent pest locations between the unmasked and masked areas, and to determine the pest distribution overlap status of the overlapping or adjacent areas, including the number of pests, the location of the bounding box, and the spatial relationship. Based on the pest type identification results of unobstructed and obstructed areas, the cosine similarity calculation method is used to calculate the similarity of the dynamic feature vector data of the heat map, determine whether the pest types in the overlapping or adjacent areas are consistent, and determine the pest type labeling of the overlapping or adjacent areas; if the pest types in the overlapping or adjacent areas are consistent, the number of pests in the overlapping or adjacent areas is calibrated by deduplication, including removing pest records with fewer pests in the overlapping or adjacent areas, retaining only the records with more pests, and recording the pest quantity and type data of the crop monitoring area; based on the pest quantity and type data of the crop monitoring area, pest control strategies are formulated, and pest control strategies are implemented in the crop monitoring area by drones. Pest control strategies include using pesticide spraying, biological control, or physical control methods to prevent pests from spreading in advance.

7. The method according to claim 1, wherein, The process involves constructing a pest prediction model based on data on the quantity, density, and type of pests in the crop monitoring area, determining future pest spread trends, and using pest control strategies to proactively control pests in the crop monitoring area, including: Continuously monitor the number, density, and type of pests in the crop monitoring area, use a long short-term memory network to train the model, build a pest prediction model, predict the number, density, and type of pests in the crop monitoring area in the future within a preset time period, and judge the future pest spread trend. If the predicted rate of increase in the number of pests in the crop monitoring area within a preset time period is greater than a preset rate threshold, or the density is greater than a preset density threshold, then pest control strategies will be used to carry out pest control in the crop monitoring area in advance.

8. The method according to claim 1, wherein, The method of assessing the accuracy of pest prediction models in assisting pest control decisions based on the effectiveness of pest control strategies and the lead time for pest control, and adjusting pest control strategies and the lead time for implementing pest control strategies, includes: After obtaining data on changes in pest density, distribution, and type following pest control, the effectiveness of pest control strategies in suppressing pest numbers and spread is assessed by comparing post-control pest data with pre-control predictions. If the effectiveness is lower than a preset effectiveness threshold, the pest control strategy is adjusted. Using pest prediction models and trigger data for pest control strategies, the rationality and lead time of pest control are evaluated by comparing predicted pest growth rates with actual pest reduction rates. The accuracy of the pest prediction model in assisting control decisions is assessed, resulting in evaluation data on the lead time and timeliness of pest control strategies. If the accuracy is lower than a preset accuracy threshold, the lead time for implementing pest control strategies is adjusted.

9. An artificial intelligence-based crop pest monitoring system, comprising the steps of an artificial intelligence-based crop pest monitoring method as described in any one of claims 1-8, characterized in that, The system includes: The crop pest analysis module is used to acquire infrared thermal imaging images of crops, build crop pest identification models, identify the types and quantities of pests in the crop monitoring area, and determine the density and activity area of ​​pest distribution. The unobstructed area pest identification module is used to acquire multi-view fused infrared thermal imaging image data of crops by using infrared thermal imaging monitoring devices set at different locations in the crop monitoring area, and to identify the type and quantity of pests in the unobstructed areas of the crop monitoring area. The module for analyzing the number of pests in the obscured area is used to identify thermal anomaly areas based on continuous time-frame infrared thermal imaging images acquired by thermal imaging equipment, determine the density level of pest activity in different thermal anomaly areas, and obtain statistical results of the number of pests in the obscured areas of the crop monitoring area. The occluded area pest type identification module is used to construct a pest type identification model based on the shape factor, temperature fluctuation frequency, spatial dynamic characteristics, and periodic characteristics of the thermal anomaly area, and to determine the pest type in the occluded area of ​​the crop monitoring area. The pest data calibration module is used to determine whether there are overlapping or adjacent pest locations in the unobstructed area and the obstructed area based on the spatial coordinates of pest locations in the unobstructed area and the obstructed area, and to calibrate the pest quantity and type data in the crop monitoring area and formulate pest control strategies. The pest spread trend prediction module is used to build a pest prediction model based on the number, density and type of pests in the crop monitoring area, to judge the future pest spread trend, and to use pest control strategies to carry out pest control in the crop monitoring area in advance. The pest control strategy adjustment module is used to assess the accuracy of pest prediction models in assisting pest control decisions based on the effectiveness of pest control strategies and the lead time for pest control, and to adjust pest control strategies and the lead time for implementing pest control strategies.

Images