Poplar pest image recognition method and system based on deep learning

Abstract

The application discloses a poplar pest image recognition method and system based on deep learning, comprising the following steps: acquiring poplar pest field knowledge to perform knowledge distillation enhancement and construct a poplar pest knowledge graph, determining a poplar variety, constructing a poplar pest ecological deep recognition model, inputting a to-be-recognized poplar pest ecological picture into the poplar pest ecological deep recognition model to obtain first poplar pest ecological characteristics, processing a multi-waveband pest picture to obtain secondary pest characteristics, performing feature cross to obtain second poplar pest ecological characteristics, determining space-time characteristics, inputting the space-time characteristics, the poplar variety and the second poplar pest ecological characteristics into the poplar pest knowledge graph to obtain a poplar pest prediction result. The method can not only improve the efficiency and accuracy of poplar pest image recognition, but also has good interpretability and can be directly applied to a poplar pest image recognition system.

Description

Technical Field

[0001] This invention relates to the field of image recognition technology, and in particular to a method and system for recognizing poplar pests based on deep learning. Background Technology

[0002] Poplar trees, as an important fast-growing tree species for sustainable forestry development, are widely used in timber production and ecological protection. Their healthy growth is directly related to the stability of the forestry economy and the balance of the ecological environment. However, frequent outbreaks of poplar pests often lead to stunted tree growth, decreased timber quality, and even large-scale mortality, causing huge losses to the forestry industry. Therefore, quickly and accurately identifying the types and severity of poplar pests is crucial for taking timely and targeted control measures and reducing disaster losses.

[0003] Existing methods for identifying poplar pests mainly rely on single visible light RGB images and general deep models, which have the following drawbacks: First, the information dimension is limited, lacking multi-band data such as near-infrared and ultraviolet light, leading to high false negative rates in early egg masses and larval stages; second, prior knowledge such as poplar varieties, symbiotic organisms, diseases, and spatiotemporal distribution is not systematically incorporated, resulting in poor model transferability in complex field scenarios; finally, the lack of knowledge distillation and graph reasoning mechanisms makes it difficult to distinguish between "similar pests" or "rare regional types" with fine-grained precision. Therefore, this invention proposes a deep learning-based image recognition method and system for poplar pests. It constructs a poplar pest knowledge graph by enhancing domain knowledge through knowledge distillation, integrating the ecological characteristics of pests, symbiotic organisms, and poplar diseases, and combining multi-band processing and spatiotemporal alignment techniques to effectively overcome the shortcomings of traditional technologies. This method not only overcomes the limitations of traditional identification technologies but also provides strong technical support for early warning and scientific control of forestry pests, which has significant practical implications for improving the intelligent level of forestry management. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for recognizing poplar pest images based on deep learning.

[0005] To achieve the above objectives, the present invention is implemented according to the following technical solution:

[0006] This invention includes the following steps:

[0007] Knowledge distillation and enhancement are performed on knowledge of poplar pests to construct a knowledge graph of poplar pests.

[0008] Poplar tree images are obtained and compared with a standard poplar tree image library to determine poplar tree species. Annotated ecological images of poplar tree pests are obtained to construct a deep ecological identification model of poplar tree pests. The ecological images of poplar tree pests include images of pests, images of symbiotic organisms, and images of poplar tree diseases.

[0009] The ecological images of poplar pests to be identified are input into the deep ecological identification model of poplar pests to obtain the first ecological characteristics of poplar pests; the first ecological characteristics of poplar pests include the characteristics of major pests, the characteristics of symbiotic organisms and the characteristics of poplar diseases.

[0010] The band difference significance map algorithm is used to process the multi-band pest map to obtain secondary pest features. The secondary pest features are then combined with the first poplar pest ecological features to obtain the second poplar pest ecological features.

[0011] The spatiotemporal features of the ecological images of poplar pests to be identified are determined by spatiotemporal alignment with the shooting data. The spatiotemporal features, the poplar species, and the second ecological features of poplar pests are then input into the poplar pest knowledge graph to obtain the poplar pest prediction results.

[0012] Furthermore, the method for constructing a knowledge graph of poplar pests includes:

[0013] A structured knowledge base is constructed by acquiring knowledge about poplar pests from expert experience, literature, image annotation, and text databases. The knowledge about poplar pests includes the morphological characteristics of pests, the characteristics of pest growth stages, the characteristics of symbiotic organisms, the characteristics of poplar diseases, and the characteristics of the growth environment.

[0014] A structured knowledge base is used to train a teacher model. A knowledge graph alignment loss function is used to establish the relationship between the features learned by the teacher model and the knowledge graph. The output of the teacher model is then transferred to the student model for knowledge distillation. The knowledge distillation includes soft label distillation, feature layer distillation, and knowledge graph embedding distillation. The soft label distillation uses the soft labels output by the teacher model to guide the student model. The feature layer distillation constructs a mapping relationship between the intermediate feature maps of the teacher model and the intermediate feature maps of the student model. The knowledge graph embedding distillation constructs a mapping relationship between the knowledge graph embedding vectors of the teacher model and the feature maps of the student model.

[0015] The expression for the knowledge graph alignment loss function is as follows:

[0016]

[0017]

[0018] in For knowledge graph alignment loss function, This is a domain knowledge weight matrix, which is related to the feature weights corresponding to the domain knowledge of poplar pests. KL divergence is used to measure the difference in probability distributions between the teacher model and the student model. The softened probability distribution for the teacher model. The softened probability distribution for the student model. The original prediction vector output by the teacher model. The original prediction vector output by the student model. For temperature hyperparameters, This is the feature map of the intermediate layer of the teacher model. The feature map of the corresponding layer in the student model;

[0019] A knowledge graph of poplar pests is constructed based on knowledge distillation-enhanced domain knowledge of poplar pests.

[0020] Feature information of newly labeled poplar pest ecological images is extracted and randomly formed into multiple candidate triples. The similarity between the corresponding candidate triples and the existing triples in the knowledge graph is calculated. When the similarity is higher than the similarity threshold, the weight of the corresponding relation is updated. When there is no corresponding triple, it is added as a new relation.

[0021] A pre-trained relation extraction model is used to process new text and extract multiple candidate triples. Entity linking technology is used to link the candidate triples to nodes in the knowledge graph. The knowledge graph is updated for nodes with confidence scores greater than a confidence threshold. The weights of neighboring edges are adjusted based on the node confidence scores. The expression is as follows:

[0022]

[0023] in for Time Node To neighboring nodes edge weights, For nodes The set of neighboring nodes, , Node confidence, , For relation type weights, Self-enhancing coefficient

[0024] Furthermore, the method for determining poplar varieties includes:

[0025] Color correction and distortion correction are performed on poplar images to obtain corrected poplar images. The corrected poplar images are then segmented into poplar trunk images, poplar leaf images, and poplar branch images according to their parts. Image features of the poplar trunk images, poplar leaf images, and poplar branch images are extracted respectively to obtain trunk features, leaf features, and branch features.

[0026] The cosine similarity between the trunk features, leaf features, and branch features and the corresponding features of the same poplar species in the standard poplar image library is used to obtain the part similarity. The seasonal weight of the part is determined according to the identification season. The species similarity is calculated by weighting the part similarity and the seasonal weight of the part. The seasonal weight of the part is a fixed value set in advance and is set according to the growth pattern of poplar leaves.

[0027] The standard poplar variety with the highest similarity value is selected as the poplar variety to be identified.

[0028] Furthermore, the method for obtaining the ecological characteristics of the first poplar pest includes:

[0029] A deep ecological identification model for poplar pests was constructed using labeled pest images, symbiotic organism images, and poplar disease images. The deep ecological identification model for poplar pests includes an input layer, a biological layer, a plant layer, an enhancement layer, a fully connected layer, and an output layer. The biological layer and the plant layer are designed in parallel.

[0030] The biological layer uses ResNet to extract pest and symbiotic image features. The initial pest and symbiotic image features are obtained by processing the image features through the biological attention mechanism of the biological channel. The expression is:

[0031]

[0032]

[0033] in Features attention, Feature tensors including pest images Feature tensors of symbiotic organism images , For biological characteristics Nonlinear mapping, Learnable biometric convolution kernels;

[0034] The plant layer uses UNet to extract multi-scale disease features from poplar disease images. The plant attention mechanism in the plant channel is used to process these multi-scale disease features to obtain the extracted poplar disease features. The expression is:

[0035]

[0036] in Multi-scale disease characteristics attention, Features Multi-scale disease perception branch, multi-scale This includes average pooling, max pooling, dilated convolution, and Gabor filtering. for Scale-channel convolution kernel;

[0037] The enhancement layer embeds a pest prototype memory library. It processes the initial pest extraction features and initial symbiotic extraction features using a prototype contrast loss function to increase the distance between approximate pest classes, thereby obtaining the pest extraction features and symbiotic extraction features. The pest prototype memory library stores cross-scale prototype vectors for each pest class, including microscopic, mesoscopic, and macroscopic prototype vectors. The prototype contrast loss function expression is as follows:

[0038]

[0039]

[0040] in For prototype comparison loss function, For the current scale Feature similarity score, For traversal scale Feature similarity score, For temperature parameters, For the current sample at scale Feature representation, For the same category In scale prototype vector, For different categories In scale prototype vector, For boundary margin;

[0041] The fully connected layer connects the plant layer and the reinforcement layer respectively, performs feature fusion and compression on the extracted features of poplar diseases, pests and symbiotic organisms, and performs fully connected feature decoding to output the main pest features, symbiotic organism features and poplar disease features;

[0042] The ecological images of poplar pests to be identified are input into the poplar pest ecological depth recognition model to obtain the first ecological features of poplar pests.

[0043] Furthermore, the method for obtaining the ecological characteristics of the second poplar pest includes:

[0044] Near-infrared, ultraviolet, and visible light images of pests are obtained and spatiotemporally aligned. The aligned images are then subjected to band difference significance processing to obtain a band difference significance map, expressed as:

[0045]

[0046] in This is a graph showing the significance of band differences. To enhance the coefficient, This is a near-infrared image of pests. Image of ultraviolet pests. Image of pests that can be seen in the light. It is the numerical stability constant. This is a picture of the original pest. For the first Stage larval characteristic weight parameters, For the first Stage larval characteristics of Gaussian centers For the first Gaussian standard deviation of larval stage characteristics;

[0047] Feature extraction is performed on the band difference significance map to obtain secondary pest features. The secondary pest features are then cross-referenced with the primary pest features in the first poplar pest ecological features to obtain pest features. The second poplar pest ecological features are composed of pest features, symbiotic organism features, and poplar disease features. The secondary pest features include growth stage, number of body segments, body color transmittance, and movement trajectory entropy.

[0048] Furthermore, the method for obtaining poplar pest prediction results includes:

[0049] By aligning the ecological images of poplar pests with the shooting data in time and space, the shooting time and shooting area are obtained. The climate environment is determined based on the shooting time and shooting area. The spatiotemporal characteristics are composed of the climate environment, shooting time and shooting area.

[0050] The spatiotemporal characteristics, poplar species, and ecological characteristics of second poplar pests are input into the poplar pest knowledge graph to obtain poplar pest prediction results; the poplar pest prediction results include pest species, pest growth stage ratio, and prediction probability.

[0051] Secondly, a deep learning-based image recognition system for poplar pests includes:

[0052] Knowledge graph module: used to acquire knowledge in the field of poplar pests, perform knowledge distillation and enhancement, and construct a knowledge graph of poplar pests; used to input the spatiotemporal features, the poplar varieties and the second ecological features of poplar pests into the knowledge graph of poplar pests to obtain poplar pest prediction results;

[0053] Tree species identification module: used to acquire poplar images and compare them with a standard poplar image library to determine the poplar species;

[0054] The deep recognition module is used to acquire labeled ecological images of poplar pests and construct a deep ecological recognition model for poplar pests; it is also used to input the ecological images of poplar pests to be identified into the deep ecological recognition model for poplar pests to obtain the first ecological features of poplar pests.

[0055] Feature enhancement module: used to process multi-band pest images using the band difference significance map algorithm to obtain secondary pest features, and to perform feature cross-referencing between the secondary pest features and the first poplar pest ecological features to obtain the second poplar pest ecological features; used to perform spatiotemporal alignment between the poplar pest ecological image to be identified and the shooting data to determine spatiotemporal features.

[0056] The beneficial effects of this invention are:

[0057] This invention relates to a method and system for recognizing poplar pests based on deep learning. Compared with existing technologies, this invention has the following technical advantages:

[0058] This invention enhances data preprocessing capabilities and improves model adaptability in poplar pest image recognition through knowledge distillation enhancement, knowledge graph construction, image comparison, model building, and band difference significance processing. This improves the efficiency and accuracy of poplar pest image recognition, significantly saves resources, increases work efficiency, and enables the identification of poplar pests. It provides strong technical support for early warning and scientific control of forest pests, and can adapt to the recognition needs of different poplar pest image recognition systems and users, demonstrating a certain degree of universality. Attached Figure Description

[0059] Figure 1 This is a flowchart illustrating the steps of the deep learning-based poplar pest image recognition method of the present invention. Detailed Implementation

[0060] The present invention will be further described below through specific embodiments. The illustrative embodiments and descriptions herein are used to explain the present invention, but are not intended to limit the present invention.

[0061] The present invention provides a deep learning-based image recognition method and system for poplar pests, comprising the following steps:

[0062] like Figure 1 As shown, this embodiment includes the following steps:

[0063] Knowledge distillation and enhancement are performed on knowledge of poplar pests to construct a knowledge graph of poplar pests.

[0064] Poplar tree images are obtained and compared with a standard poplar tree image library to determine poplar tree species. Annotated ecological images of poplar tree pests are obtained to construct a deep ecological identification model of poplar tree pests. The ecological images of poplar tree pests include images of pests, images of symbiotic organisms, and images of poplar tree diseases.

[0065] The ecological images of poplar pests to be identified are input into the deep ecological identification model of poplar pests to obtain the first ecological characteristics of poplar pests; the first ecological characteristics of poplar pests include the characteristics of major pests, the characteristics of symbiotic organisms and the characteristics of poplar diseases.

[0066] The band difference significance map algorithm is used to process the multi-band pest map to obtain secondary pest features. The secondary pest features are then combined with the first poplar pest ecological features to obtain the second poplar pest ecological features.

[0067] The spatiotemporal features of the ecological images of poplar pests to be identified are determined by spatiotemporal alignment with the shooting data. The spatiotemporal features, the poplar species, and the second ecological features of poplar pests are then input into the poplar pest knowledge graph to obtain the poplar pest prediction results.

[0068] In this embodiment, the method for constructing a knowledge graph of poplar pests includes:

[0069] A structured knowledge base is constructed by acquiring knowledge about poplar pests from expert experience, literature, image annotation, and text databases. The knowledge about poplar pests includes the morphological characteristics of pests, the characteristics of pest growth stages, the characteristics of symbiotic organisms, the characteristics of poplar diseases, and the characteristics of the growth environment.

[0070] A structured knowledge base is used to train a teacher model. A knowledge graph alignment loss function is used to establish the relationship between the features learned by the teacher model and the knowledge graph. The output of the teacher model is then transferred to the student model for knowledge distillation. The knowledge distillation includes soft label distillation, feature layer distillation, and knowledge graph embedding distillation. The soft label distillation uses the soft labels output by the teacher model to guide the student model. The feature layer distillation constructs a mapping relationship between the intermediate feature maps of the teacher model and the intermediate feature maps of the student model. The knowledge graph embedding distillation constructs a mapping relationship between the knowledge graph embedding vectors of the teacher model and the feature maps of the student model.

[0071] The expression for the knowledge graph alignment loss function is as follows:

[0072]

[0073]

[0074] in For knowledge graph alignment loss function, This is a domain knowledge weight matrix, which is related to the feature weights corresponding to the domain knowledge of poplar pests. KL divergence is used to measure the difference in probability distributions between the teacher model and the student model. The softened probability distribution for the teacher model. The softened probability distribution for the student model. The original prediction vector output by the teacher model. The original prediction vector output by the student model. For temperature hyperparameters, This is the feature map of the intermediate layer of the teacher model. The feature map of the corresponding layer in the student model;

[0075] A knowledge graph of poplar pests is constructed based on knowledge distillation-enhanced domain knowledge of poplar pests.

[0076] Feature information of newly labeled poplar pest ecological images is extracted and randomly formed into multiple candidate triples. The similarity between the corresponding candidate triples and the existing triples in the knowledge graph is calculated. When the similarity is higher than the similarity threshold, the weight of the corresponding relation is updated. When there is no corresponding triple, it is added as a new relation.

[0077] A pre-trained relation extraction model is used to process new text and extract multiple candidate triples. Entity linking technology is used to link the candidate triples to nodes in the knowledge graph. The knowledge graph is updated for nodes with confidence scores greater than a confidence threshold. The weights of neighboring edges are adjusted based on the node confidence scores. The expression is as follows:

[0078]

[0079] in for Time Node To neighboring nodes edge weights, For nodes The set of neighboring nodes, , Node confidence, , For relation type weights, This is the self-enhancing coefficient;

[0080] In the actual assessment, knowledge of poplar pests includes morphological characteristics (appearance / texture / size / color / posture / movement), growth stage characteristics (morphological evolution of eggs / larvae / pupa / adults), symbiotic characteristics (visual identifiers of specific parasites and pathogens / symbiotic types / spatial co-location / size ratio / interaction frequency), poplar disease characteristics (affected parts / disease texture / disease shape / leaf fading pattern / disease response), and growth environment characteristics (parasitic poplar varieties / geographical distribution / seasonal adaptability). The corresponding feature weights for knowledge of poplar pests are 0.35 / 0.25 / 0.15 / 0.15 / 0.10, respectively. Knowledge distillation is then performed on the knowledge of poplar pests.

[0081] The knowledge graph of poplar pests includes nodes of the following types: pests, growth stages, symbiotic organisms, disease manifestations, poplar varieties, geographical regions, and climate environments. The edge relationships include: pest-symbiotic organism edge, pest-poplar variety edge, pest-disease manifestation edge, pest-growth stage edge, pest-geographical region edge, pest-climate environment edge, disease manifestation-poplar variety edge, geographical region-climate environment edge, geographical region-poplar variety edge, and poplar variety-climate environment edge.

[0082] Geographic information, temporal information, growth stage information, symbiotic information, and disease manifestation information are extracted from newly labeled poplar pest ecological images. Multiple candidate triples are randomly formed, with a similarity threshold of 0.85. The similarity of the triples is calculated, and the relationship is updated to obtain the relationship weight.

[0083] The pre-trained relation extraction model architecture is BERT-EP+GRN+BiLSTM. The BERT-EP base model is BERT-base-Chinese. The GRN gated relation network controls the text information input through relation gating vectors and ReLU activation function. The BiLSTM bidirectional LSTM decoding layer captures long-distance dependency relations and outputs candidate triples and probability distributions.

[0084] In this embodiment, the method for determining the poplar variety includes:

[0085] Color correction and distortion correction are performed on poplar images to obtain corrected poplar images. The corrected poplar images are then segmented into poplar trunk images, poplar leaf images, and poplar branch images according to their parts. Image features of the poplar trunk images, poplar leaf images, and poplar branch images are extracted respectively to obtain trunk features, leaf features, and branch features.

[0086] The cosine similarity between the trunk features, leaf features, and branch features and the corresponding features of the same poplar species in the standard poplar image library is used to obtain the part similarity. The seasonal weight of the part is determined according to the identification season. The species similarity is calculated by weighting the part similarity and the seasonal weight of the part. The seasonal weight of the part is a fixed value set in advance and is set according to the growth pattern of poplar leaves.

[0087] The standard poplar variety with the highest variety similarity value is selected as the poplar variety to be identified;

[0088] In the actual evaluation, Retinex theory was used for color correction of the poplar image, and an improved Brown-Conrady model was used for distortion correction to obtain the corrected poplar image.

[0089] Mask R-CNN was used to segment and correct poplar images to obtain poplar trunk images. The shape, texture and size of the poplar trunk images were extracted using Hu moment invariant features, LBP-TOP and minimum bounding rectangle respectively to form trunk features.

[0090] Poplar leaf images were obtained by dividing and correcting poplar images using SLIC superpixel clustering. The color, texture, shape, and density of the poplar leaf images were extracted using HSV histogram, GLCM co-occurrence matrix, Zernike moments, and leaf count / region area, respectively.

[0091] U-Net was used to divide and correct poplar images to obtain poplar branch images. The shape and size of the poplar branch images were extracted to form branch features by using the skeleton fractal dimension and the histogram of branch diameter distribution.

[0092] Reference Table 1 for fixed values ​​of seasonal weights for different parts:

[0093] Table 1 Reference Table of Seasonal Weights for Geographic Locations

[0094]

[0095] Taking the identification of the poplar species to be identified as an example, the similarity of the parts of the identified poplar species with three candidate poplar species was determined to be: European Poplar (trunk similarity 0.92 / leaf similarity 0.88 / branch similarity 0.85), White Poplar (trunk similarity 0.78 / leaf similarity 0.91 / branch similarity 0.82), and Xinjiang Poplar (trunk similarity 0.85 / leaf similarity 0.76 / branch similarity 0.88). Taking the identification season (autumn), the seasonal weights of the parts were determined by looking up the table to be 0.4, 0.4, and 0.2, respectively. The weighted similarity of the species was calculated to be 0.89, 0.84, and 0.81, respectively. The poplar species to be identified was determined to be European Poplar No. 107.

[0096] In this embodiment, the method for obtaining the ecological characteristics of the first poplar pest includes:

[0097] A deep ecological identification model for poplar pests was constructed using labeled pest images, symbiotic organism images, and poplar disease images. The deep ecological identification model for poplar pests includes an input layer, a biological layer, a plant layer, an enhancement layer, a fully connected layer, and an output layer. The biological layer and the plant layer are designed in parallel.

[0098] The biological layer uses ResNet to extract pest and symbiotic image features. The initial pest and symbiotic image features are obtained by processing the image features through the biological attention mechanism of the biological channel. The expression is:

[0099]

[0100]

[0101] in Features attention, Feature tensors including pest images Feature tensors of symbiotic organism images , For biological characteristics Nonlinear mapping, Learnable biometric convolution kernels;

[0102] The plant layer uses UNet to extract multi-scale disease features from poplar disease images. The plant attention mechanism in the plant channel is used to process these multi-scale disease features to obtain the extracted poplar disease features. The expression is:

[0103]

[0104] in Multi-scale disease characteristics attention, Features Multi-scale disease perception branch, multi-scale This includes average pooling, max pooling, dilated convolution, and Gabor filtering. for Scale-channel convolution kernel;

[0105] The enhancement layer embeds a pest prototype memory library. It processes the initial pest extraction features and initial symbiotic extraction features using a prototype contrast loss function to increase the distance between approximate pest classes, thereby obtaining the pest extraction features and symbiotic extraction features. The pest prototype memory library stores cross-scale prototype vectors for each pest class, including microscopic, mesoscopic, and macroscopic prototype vectors. The prototype contrast loss function expression is as follows:

[0106]

[0107]

[0108] in For prototype comparison loss function, For the current scale Feature similarity score, For traversal scale Feature similarity score, For temperature parameters, For the current sample at scale Feature representation, For the same category In scale prototype vector, For different categories In scale prototype vector, For boundary margin;

[0109] The fully connected layer connects the plant layer and the reinforcement layer respectively, performs feature fusion and compression on the extracted features of poplar diseases, pests and symbiotic organisms, and performs fully connected feature decoding to output the main pest features, symbiotic organism features and poplar disease features;

[0110] Input the ecological images of poplar pests to be identified into the deep ecological identification model of poplar pests to obtain the first ecological features of poplar pests.

[0111] In actual evaluation, in the enhancement layer, micro-prototype vectors (texture / lines) are extracted using local binary mode + HOG, meso-prototype vectors (shape / color) are extracted using ResNet 34-block 3-feature extraction, and macro-prototype vectors (quantity / spatial distribution) are extracted using Vision Transformer global feature extraction. These elements together form a pest prototype memory database, including temperature parameters. Boundary margin ;

[0112] In the fully connected layer, poplar disease extraction features are decoded through the disease feature branch to obtain poplar disease features (affected location / texture roughness / color outliers / lesion shape / spread rate), pest extraction features are decoded through the pest feature branch to obtain main pest features (shape and size / color histogram / line direction / texture entropy / quantity / spatial density), and symbiotic organism extraction features are decoded through the symbiotic organism feature branch to obtain symbiotic organism features (symbiotic type / parasitic intensity / spatial co-location / size ratio / movement trend).

[0113] The poplar pest ecological depth identification model adjusts the identification accuracy through a hybrid loss function consisting of cross-entropy loss, prototype comparison loss, and auxiliary segmentation loss.

[0114] In this embodiment, the method for obtaining the ecological characteristics of the second poplar pest includes:

[0115] Near-infrared, ultraviolet, and visible light images of pests are obtained and spatiotemporally aligned. The aligned images are then subjected to band difference significance processing to obtain a band difference significance map, expressed as:

[0116]

[0117] in This is a graph showing the significance of band differences. To enhance the coefficient, This is a near-infrared image of pests. Image of ultraviolet pests. Image of pests that can be seen in the light. It is the numerical stability constant. This is a picture of the original pest. For the first Stage larval characteristic weight parameters, For the first Stage larval characteristics of Gaussian centers For the first Gaussian standard deviation of larval stage characteristics;

[0118] Feature extraction is performed on the band difference significance map to obtain secondary pest features. The secondary pest features are then cross-referenced with the primary pest features in the first poplar pest ecological features to obtain pest features. The second poplar pest ecological features are composed of pest features, symbiotic organism features, and poplar disease features. The secondary pest features include growth stage, number of body segments, body color transmittance, and movement trajectory entropy.

[0119] In actual evaluation, the enhancement coefficient is taken. The numerical stability constant is 2.5. 10 -7 The following methods were used to determine the growth stages: band difference significance was obtained by performing band difference significance analysis; morphological opening operation and area analysis were used to process the band difference significance map to determine the number of segments; ridge line detection and local maxima processing were used to process the band difference significance map to determine the body color transmittance; and temporal frame difference entropy processing was used to process the band difference significance map to determine the motion trajectory entropy. The criteria for determining each growth stage are shown in Table 2.

[0120] Table 2 Reference Table for Pest Growth Stages

[0121]

[0122] The specific method for obtaining pest characteristics by feature cross-referencing secondary pest characteristics with primary pest characteristics from the first set of poplar pest ecological characteristics is as follows: Common features (growth stage, body color, texture, movement characteristics) of secondary and primary pest characteristics are weighted and fused; dissimilar features of secondary and primary pest characteristics are spliced ​​together. The weighting weights for primary and secondary pest characteristics are as follows: , ,and , , As a regulating factor, Confidence level for secondary features;

[0123] Pest characteristics are obtained by cross-referencing features, including basic characteristics (shape, size / color), texture characteristics (lines / texture), behavioral characteristics (quantity / density), developmental characteristics (growth stage / segment), physiological characteristics (body color translucency), and movement characteristics (trajectory entropy). The second poplar pest ecological characteristics are composed of pest characteristics, symbiotic organism characteristics, and poplar disease characteristics.

[0124] In this embodiment, the method for obtaining poplar pest prediction results includes:

[0125] By aligning the ecological images of poplar pests with the shooting data in time and space, the shooting time and shooting area are obtained. The climate environment is determined based on the shooting time and shooting area. The spatiotemporal characteristics are composed of the climate environment, shooting time and shooting area.

[0126] The spatiotemporal characteristics, poplar species, and ecological characteristics of second poplar pests are input into the poplar pest knowledge graph to obtain poplar pest prediction results; the poplar pest prediction results include pest species, pest growth stage ratio, and prediction probability.

[0127] In practical assessment, taking the image recognition of poplar pests in a certain location as an example, the EXIF ​​metadata of the poplar pest ecological images was extracted and the shooting season (July 15, summer) was determined based on astronomical algorithms. The GPS coordinates of the poplar pest ecological images were extracted and the ecological region (North China Plain deciduous forest area) was determined based on the H3 geographic grid ecological zoning. Climate retrieval was performed based on the shooting season and ecological region. Based on the climate retrieval results (average daily temperature 28.5℃, relative humidity 65%, rainfall 15.2mm, sunshine duration 8.2h), the spatiotemporal features were composed of the shooting season, ecological region, and climate retrieval results.

[0128] The spatiotemporal characteristics and corresponding poplar varieties (European and American poplar 107), along with the ecological characteristics of second poplar pests, are input into the poplar pest knowledge graph to obtain poplar pest prediction candidates: Asian longhorn beetle / 0.92, mulberry longhorn beetle / 0.41, and poplar stem weevil / 0.63. The Asian longhorn beetle, which has the highest prediction probability, is selected as the prediction result, and the corresponding pest growth stage ratio and prediction probability are output. Based on the poplar pest prediction results, poplar varieties, spatiotemporal characteristics, and second poplar pest ecological characteristics, pest control methods are determined.

[0129] Secondly, a deep learning-based image recognition system for poplar pests includes:

[0130] Knowledge graph module: used to acquire knowledge in the field of poplar pests, perform knowledge distillation and enhancement, and construct a knowledge graph of poplar pests; used to input the spatiotemporal features, the poplar varieties and the second ecological features of poplar pests into the knowledge graph of poplar pests to obtain poplar pest prediction results;

[0131] Tree species identification module: used to acquire poplar images and compare them with a standard poplar image library to determine the poplar species;

[0132] The deep recognition module is used to acquire labeled ecological images of poplar pests and construct a deep ecological recognition model for poplar pests; it is also used to input the ecological images of poplar pests to be identified into the deep ecological recognition model for poplar pests to obtain the first ecological features of poplar pests.

[0133] Feature enhancement module: used to process multi-band pest images using the band difference significance map algorithm to obtain secondary pest features, and to perform feature cross-referencing between the secondary pest features and the first poplar pest ecological features to obtain the second poplar pest ecological features; used to perform spatiotemporal alignment between the poplar pest ecological image to be identified and the shooting data to determine spatiotemporal features.

[0134] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A deep learning-based image recognition method for poplar pests, characterized in that, Includes the following steps: S1. Acquire knowledge in the field of poplar pests, perform knowledge distillation and enhancement, and construct a knowledge graph of poplar pests; S2. Obtain poplar images and compare them with a standard poplar image library to determine the poplar species. Obtain labeled poplar pest ecological images to construct a deep ecological identification model for poplar pests. The poplar pest ecological images include pest images, symbiotic organism images, and poplar disease images. S3. Input the ecological image of poplar pests to be identified into the deep ecological identification model of poplar pests to obtain the first ecological characteristics of poplar pests; the first ecological characteristics of poplar pests include the characteristics of major pests, the characteristics of symbiotic organisms and the characteristics of poplar diseases; S4. Use the band difference significance map algorithm to process the multi-band pest map to obtain secondary pest features, and perform feature cross-interaction between the secondary pest features and the first poplar pest ecological features to obtain the second poplar pest ecological features. S5. The ecological images of poplar pests to be identified are spatiotemporally aligned with the shooting data to determine the spatiotemporal features. The spatiotemporal features, the poplar species, and the second ecological features of poplar pests are input into the poplar pest knowledge graph to obtain the poplar pest prediction results.

2. The method for recognizing poplar pest images based on deep learning according to claim 1, characterized in that, The method for constructing a knowledge graph of poplar pests includes: A structured knowledge base is constructed by acquiring knowledge about poplar pests from expert experience, literature, image annotation, and text databases. The knowledge about poplar pests includes the morphological characteristics of pests, the characteristics of pest growth stages, the characteristics of symbiotic organisms, the characteristics of poplar diseases, and the characteristics of the growth environment. A structured knowledge base is used to train a teacher model. A knowledge graph alignment loss function is used to establish the relationship between the features learned by the teacher model and the knowledge graph. The output of the teacher model is then transferred to the student model for knowledge distillation. The knowledge distillation includes soft label distillation, feature layer distillation, and knowledge graph embedding distillation. The soft label distillation uses the soft labels output by the teacher model to guide the student model. The feature layer distillation constructs a mapping relationship between the intermediate feature maps of the teacher model and the intermediate feature maps of the student model. The knowledge graph embedding distillation constructs a mapping relationship between the knowledge graph embedding vectors of the teacher model and the feature maps of the student model. The expression for the knowledge graph alignment loss function is as follows: in For knowledge graph alignment loss function, This is a domain knowledge weight matrix, which is related to the feature weights corresponding to the domain knowledge of poplar pests. KL divergence is used to measure the difference in probability distributions between the teacher model and the student model. The softened probability distribution for the teacher model. The softened probability distribution for the student model. The original prediction vector output by the teacher model. The original prediction vector output by the student model. For temperature hyperparameters, This is the feature map of the intermediate layer of the teacher model. The feature map of the corresponding layer in the student model; A knowledge graph of poplar pests is constructed based on knowledge distillation-enhanced domain knowledge of poplar pests. Feature information of newly labeled poplar pest ecological images is extracted and randomly formed into multiple candidate triples. The similarity between the corresponding candidate triples and the existing triples in the knowledge graph is calculated. When the similarity is higher than the similarity threshold, the weight of the corresponding relation is updated. When there is no corresponding triple, it is added as a new relation. A pre-trained relation extraction model is used to process new text and extract multiple candidate triples. Entity linking technology is used to link the candidate triples to nodes in the knowledge graph. The knowledge graph is updated for nodes with confidence scores greater than a confidence threshold. The weights of neighboring edges are adjusted based on the node confidence scores. The expression is as follows: in for Time Node To neighboring nodes edge weights, For nodes The set of neighboring nodes, , Node confidence, , For relation type weights, This is the self-enhancing coefficient.

3. The method for recognizing poplar pest images based on deep learning according to claim 1, characterized in that, The method for determining poplar varieties includes: Color correction and distortion correction are performed on poplar images to obtain corrected poplar images. The corrected poplar images are then segmented into poplar trunk images, poplar leaf images, and poplar branch images according to their parts. Image features of the poplar trunk images, poplar leaf images, and poplar branch images are extracted respectively to obtain trunk features, leaf features, and branch features. The cosine similarity between the trunk features, leaf features, and branch features and the corresponding features of the same poplar species in the standard poplar image library is used to obtain the part similarity. The seasonal weight of the part is determined according to the identification season. The species similarity is calculated by weighting the part similarity and the seasonal weight of the part. The seasonal weight of the part is a fixed value set in advance and is set according to the growth pattern of poplar leaves. The standard poplar variety with the highest similarity value is selected as the poplar variety to be identified.

4. The poplar pest image recognition method based on deep learning according to claim 1, characterized in that, The method for obtaining the ecological characteristics of the first poplar pest includes: A deep ecological identification model for poplar pests was constructed using labeled pest images, symbiotic organism images, and poplar disease images. The deep ecological identification model for poplar pests includes an input layer, a biological layer, a plant layer, an enhancement layer, a fully connected layer, and an output layer. The biological layer and the plant layer are designed in parallel. The biological layer uses ResNet to extract pest and symbiotic image features. The initial pest and symbiotic image features are obtained by processing the image features through the biological attention mechanism of the biological channel. The expression is: in Features attention, Feature tensors including pest images Feature tensors of symbiotic organism images , For biological characteristics Nonlinear mapping, Learnable biometric convolution kernels; The plant layer uses UNet to extract multi-scale disease features from poplar disease images. The plant attention mechanism in the plant channel is used to process these multi-scale disease features to obtain the extracted poplar disease features. The expression is: in Multi-scale disease characteristics attention, Features Multi-scale disease perception branch, multi-scale This includes average pooling, max pooling, dilated convolution, and Gabor filtering. for Scale-channel convolution kernel; The enhancement layer embeds a pest prototype memory library. It processes the initial pest extraction features and initial symbiotic extraction features using a prototype contrast loss function to increase the distance between approximate pest classes, thereby obtaining the pest extraction features and symbiotic extraction features. The pest prototype memory library stores cross-scale prototype vectors for each pest class, including microscopic, mesoscopic, and macroscopic prototype vectors. The prototype contrast loss function expression is as follows: in For prototype comparison loss function, For the current scale Feature similarity score, For traversal scale Feature similarity score, For temperature parameters, For the current sample at scale Feature representation, For the same category In scale prototype vector, For different categories In scale prototype vector, For boundary margin; The fully connected layer connects the plant layer and the reinforcement layer respectively, performs feature fusion and compression on the extracted features of poplar diseases, pests and symbiotic organisms, and performs fully connected feature decoding to output the main pest features, symbiotic organism features and poplar disease features; The ecological images of poplar pests to be identified are input into the poplar pest ecological depth recognition model to obtain the first ecological features of poplar pests.

5. The method for recognizing poplar pest images based on deep learning according to claim 1, characterized in that, The method for obtaining the ecological characteristics of the second poplar pest includes: Near-infrared, ultraviolet, and visible light images of pests are obtained and spatiotemporally aligned. The aligned images are then subjected to band difference significance processing to obtain a band difference significance map, expressed as: in This is a graph showing the significance of band differences. To enhance the coefficient, This is a near-infrared image of pests. Image of ultraviolet pests. Image of pests that can be seen in the light. It is the numerical stability constant. This is a picture of the original pest. For the first Stage larval characteristic weight parameters, For the first Stage larval characteristics of Gaussian centers For the first Gaussian standard deviation of larval stage characteristics; Feature extraction is performed on the band difference significance map to obtain secondary pest features. The secondary pest features are then cross-referenced with the primary pest features in the first poplar pest ecological features to obtain pest features. The second poplar pest ecological features are composed of pest features, symbiotic organism features, and poplar disease features. The secondary pest features include growth stage, number of body segments, body color transmittance, and movement trajectory entropy.

6. The method for recognizing poplar pest images based on deep learning according to claim 1, characterized in that, The method for obtaining poplar pest prediction results includes: By aligning the ecological images of poplar pests with the shooting data in time and space, the shooting time and shooting area are obtained. The climate environment is determined based on the shooting time and shooting area. The spatiotemporal characteristics are composed of the climate environment, shooting time and shooting area. The spatiotemporal characteristics, poplar species, and ecological characteristics of second poplar pests are input into the poplar pest knowledge graph to obtain poplar pest prediction results; the poplar pest prediction results include pest species, pest growth stage ratio, and prediction probability.

7. A deep learning-based poplar pest image recognition system, used to perform the method according to any one of claims 1-6, characterized in that, include: Knowledge Graph Module: Used to acquire knowledge in the field of poplar pests, perform knowledge distillation and enhancement, and construct a knowledge graph of poplar pests; This is used to input the spatiotemporal features, the poplar species, and the second poplar pest ecological features into the poplar pest knowledge graph to obtain poplar pest prediction results; Tree species identification module: used to acquire poplar images and compare them with a standard poplar image library to determine the poplar species; The deep recognition module is used to acquire labeled ecological images of poplar pests and construct a deep ecological recognition model for poplar pests; it is also used to input the ecological images of poplar pests to be identified into the deep ecological recognition model for poplar pests to obtain the first ecological features of poplar pests. Feature enhancement module: used to process multi-band pest images using the band difference significance map algorithm to obtain secondary pest features, and to perform feature cross-referencing between the secondary pest features and the first poplar pest ecological features to obtain the second poplar pest ecological features; used to perform spatiotemporal alignment between the poplar pest ecological image to be identified and the shooting data to determine spatiotemporal features.

Images