Plant disease small sample classification method and system based on feature space calibration

Abstract

The application provides a plant disease small sample classification method and system based on feature space calibration, and belongs to the field of data processing. To solve the problem that the existing deep learning algorithm needs a large number of annotated data sets when classifying images, but the incidence of some plant diseases is low, the image data is small, and the deep learning algorithm is limited in large-scale application in plant disease identification. The FAS score is used to score and analyze each feature output layer of the Swin-Transformer V2 network after removing the full connection layer, the layer with the highest score is selected as the feature vector, the Swin-Transformer V2F6 network is obtained, and then the plant disease feature space calibration is carried out. The accuracy of the small sample disease classification can be effectively improved, and compared with the use of a large number of annotated data sets, the accuracy of identification can be improved on the basis of a small amount of data sets.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method and system for classifying small samples of plant diseases based on feature space calibration. Background Art

[0002] Against the backdrop of rapid advances in hardware and information technology, deep learning has gradually become a mainstream method in the field of image recognition. Conventional deep learning image classification processes require a large number of images as input for model training. Models trained using appropriate methods can effectively extract image features and demonstrate excellent performance during testing. With the development of the Internet of Things (IoT), an increasing number of end devices are capturing massive amounts of images. With the use of high-definition mobile phone cameras, ordinary people can now capture real-world images of crop diseases worldwide. Therefore, low-cost, universal, automatic, and accurate plant disease classification models are beginning to replace manual judgment of crop diseases.

[0003] Currently, the best image classification results are mostly achieved using deep learning algorithms based on convolutional neural networks (CNNs). Extensive research has demonstrated the feasibility of using deep learning techniques for plant disease classification. However, one limitation of these algorithms is the requirement for large datasets of annotated images for training. However, in practical applications, the temporal and spatial irregularities of some plant diseases result in low disease probabilities and limited image data. This directly leads to low recognition rates for these rare diseases, significantly limiting the large-scale application of deep learning techniques for plant disease identification. Summary of the Invention

[0004] The technical problems to be solved by the present invention are:

[0005] In order to solve the problem that existing deep learning algorithms require a large amount of annotated data sets for image classification, but the incidence of some plant diseases is low and the image data is scarce, which limits the large-scale application of deep learning algorithms in plant disease identification.

[0006] The present invention is to solve the above technical problems using the following technical solutions:

[0007] The present invention provides a small sample classification method for plant diseases based on feature space calibration, comprising the following steps:

[0008] S100, based on the Swin-Transformer V2 network, after removing the fully connected layer, sequentially input multiple feature vectors and run multiple plant disease datasets, and score each feature output layer in the Swin-Transformer V2 network using a feature layer evaluation index, wherein the feature evaluation index is the feature adaptation score FAS;

[0009] S200, select the output result FAS with the highest score in step S100 as the feature vector, delete the Stage4 block before the last Patch Merging layer in the Swin-Transformer V2 network, and obtain the Swin-Transformer V2 F6 network;

[0010] S300. Calibrate the plant disease feature space based on the network structure Swin-Transformer V2 F6 obtained in step S100 and the Swin-Transformer V2 network designed in step S200.

[0011] Furthermore, in step S100, the plant disease dataset includes the PlantVillage dataset, the PlantDoc dataset, the Plant Real-World dataset, and the Plant&Pest dataset, and the FAS score is performed after seven feature vectors are input into the above four datasets.

[0012] Furthermore, the output of the 6th layer after FAS scoring is used as the feature vector for small sample analysis.

[0013] Furthermore, in step S300, it specifically includes:

[0014] S310, the query set Q and the support set S are calculated by Swin-Transformer V2 F6 to obtain the feature vector Q features 、S features , the training set P of the PlantVillage dataset is transformed into a feature vector set P through SwinTransformerV2F6 features , and as a benchmark, the eigenvector S features Perform calibration;

[0015] S320, for category i, the set of all characteristic vectors of PlantVillage is P features The support set for this class in turn Find the L2 distance d i , sort all distances from small to large to get Take the first m vectors with the smallest distance after sorting Calculate its center of mass

[0016] Determine all the centroids c used for calibration i Is the calculation completed? If it is completed, continue to the next step. If not, repeat this step until all are completed.

[0017] S330, the centroids of all categories calculated in step S320 are For class i, first calculate the class centroid c i and The L2 distance of and sorting it from small to large, we get Take the two distances with the smallest distance after sorting and The corresponding centroids are c a and c b , select a point T on the line connecting the two centroids; Move γ units along the direction of itself and T to complete the support set The calibrated support set is obtained

[0018] Repeat the above steps until all support set features are completed Calibration;

[0019] S340, calculate all calibrated support sets and the query set feature vector q features The distance d i , output the label corresponding to the query set q:

[0020] Furthermore, for 5-way-5-shot and 5-way-10-shot tasks, as the average of the support set features.

[0021] Furthermore, the location of point T is determined by the hyperparameter λ.

[0022] Furthermore, the calibrated support set is

[0023] A plant disease small sample classification system based on feature space calibration has a program module corresponding to the above steps, and executes the steps in the plant disease small sample classification method based on feature space calibration when running.

[0024] A computer-readable storage medium stores a computer program configured to implement the steps of a plant disease small sample classification method based on feature space calibration when called by a processor.

[0025] Compared with the prior art, the present invention has the following beneficial effects:

[0026] The present invention provides a small sample plant disease classification method and system based on feature space calibration. The method uses the FAS score to score and analyze each feature output layer of the Swin-Transformer V2 network, selects the layer with the highest score as the feature vector of the present invention, deletes the Stage 4 block of the previous layer of the last Patch Merging in the existing Swin-Transformer V2 network, and then performs plant disease feature space calibration. After comparative analysis with the existing algorithm, it can be seen that the feature extraction network structure Swin-Transformer V2 is used to extract the disease sample. Compared with existing network structures, F6 does not require training parameters and can be directly obtained through a reasonable FAS. The present invention can significantly improve the performance of small-sample disease classification, especially in 5-way-1-shot, 5-way-5-shot, and 5-way-10-shot tasks. Its accuracy can be effectively improved. Compared with the need to use a large number of annotated datasets, this method can improve recognition accuracy based on a small number of datasets. In particular, when each sample has only 10 images (5W10S), the performance on PlantVillage and FSPlantPest can reach the performance level of conventional supervised classification tasks. This is an important discovery for small-sample plant disease tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 This is a score graph of the FAS indicator at each layer F1-F7 in an embodiment of the present invention;

[0028] Figure 2 This is a network structure diagram of Swin-Transformer V2 F6 in an embodiment of the present invention;

[0029] Figure 3 This is a comparison chart of the classification accuracy of Swin-Transformer V2 F6 and Swin-Transformer V2 F7 in 5-way-1-shot, 5-way-5-shot, and 5-way-10-shot tasks, respectively, according to an embodiment of the present invention.

[0030] Figure 4 Flowchart of a small sample classification method for plant diseases based on feature space calibration in an embodiment of the present invention;

[0031] Figure 5 1 is a diagram illustrating the implementation process of a small sample classification method for plant diseases based on feature space calibration in an embodiment of the present invention;

[0032] Figure 6This is a performance comparison chart of the method in the embodiment of the present invention and the existing algorithm on the PlantVillage dataset;

[0033] Figure 7 This is a performance comparison chart of the method in the embodiment of the present invention and the existing algorithm on the PlantDoc dataset;

[0034] Figure 8 This is a performance comparison chart of the method in the embodiment of the present invention and the existing algorithm on the Plant Real-World dataset;

[0035] Figure 9 This is a performance comparison chart of the method in the embodiment of the present invention and the existing algorithm on the Plant&Pest dataset. DETAILED DESCRIPTION

[0036] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0037] Specific implementation plan 1: Combined Figures 1 to 5 As shown, the present invention provides a small sample classification method for plant diseases based on feature space calibration, comprising the following steps:

[0038] S100, based on the Swin-Transformer V2 network, each feature output layer in the Swin-Transformer V2 network is scored using the feature layer evaluation index. The feature evaluation index is the feature adaptation score FAS. In order to ensure the accuracy of the feature adaptation score FAS, currently available and commonly used datasets including the PlantVillage dataset, the PlantDoc dataset, the Plant Real-World dataset, and the Plant&Pest dataset are selected for analysis. The analysis results are as follows Figure 1 As shown,

[0039] We assume that a network structure has a total of L layers, and we use φ l (x) represents the feature representation of the input image x in the lth layer of the network. The jth sample xij of category i in the training set is sent into the network and after L layers of forward inference, the image x is obtained. ij After each layer of features is flattened to obtain a one-dimensional feature vector φ l (x ij ), where l = 1, 2, ..., L;

[0040] If we use the ImageNet pre-trained model to directly calculate the feature vector, in order to adapt to different tasks, we need to remove the last fully connected layer of the network and use the feature vector φ of the second-to-last layer, which is the L-1 layer. L-1 (x) to compare with the prototype; The first goal of the present invention is to explore the L-1 layer feature vector φ that has been considered to be most suitable for comparison with the prototype in the FSL task. L-1 (x) Whether it is also suitable for small sample tasks of crops;

[0041] To evaluate the feature vector φ of layer l l (x) Feature Adaptation Score (FAS) in the small sample task of plant diseases l , we first define three parameters, namely, intra-class variance Between-class variance and the average inter-class distance In order to derive these three parameters, we first need to calculate the mean μ of the feature vector of the sample of category i at layer l. i,l And the feature vector mean μl of all categories in the lth layer is expressed as follows:

[0042]

[0043] Among them, N represents the total number of categories; K represents the total number of samples in each category;

[0044] Based on the above discussion, we derive expressions for three key parameters:

[0045]

[0046] Assume that the eigenvector φ l (x) is subject to multidimensional normal distribution, then μ i,l and μ l It also obeys the multidimensional normal distribution, so σ2 obeys the non-central chi-square distribution; since the dimension directly affects the mean and variance of the non-central chi-square distribution, if you want to compare, you need to transform the eigenvector φ of each layer l (x) scales it according to the dimension so that the feature vectors of all layers are compared on the same scale; finally, after the dimension scaling, we can define FAS as follows l :

[0047]

[0048] Ideal feature vector φ for calculating the distance from the prototype l(x), should first have low intra-class variance, high inter-class variance and high inter-class distance, so as to ensure that the model has better consistency for samples in the same class, and has higher discrimination for samples of different classes; the more suitable feature vector φ l (x) corresponding to the FAS score should be higher, formula (6) can meet this requirement;

[0049] We use Swin-Transformer V2-T as the feature extractor of the crop disease few-shot task, according to the official source code, remove the full connection layer after Swin-Transformer V2-T, in the inference process will output 7 features in turn, we use the pre-training weight of ImageNet-1K, and the backbone network of Swin-Transformer V2-T structure, run the dataset PlantVillage, FSPlantPest, PlantDoc and PlantReal-World in turn, get 7 feature vectors φ 1～7 (x) of 4 crop disease datasets; Figure 1

[0050] From the table in Figure 1 , we can see that the 6th layer feature vector of Swin-Transformer V2-T far exceeds the scores of all other layers, indicating that the 6th layer is more suitable for crop disease few-shot task than other layers. At this time, we use the 6th layer feature vector F6 and the most commonly used 7th layer feature vector F7 to continue the comparison experiment on different datasets to test the accuracy of few-shot task. The results are shown in Figure 3

[0051] From the experimental results, we can see that the highest scoring F6 has greatly improved the accuracy of the few-shot task compared to the commonly used F7, so whether it is through FAS or the test set accuracy of the few-shot task, we can determine that F6 is more suitable for plant disease few-shot task than F7, which is the first important discovery in the field of few-shot; based on the above experimental results, we can get the most suitable feature extraction network structure Swin-Transformer V2 F6 for plant disease few-shot task, as shown in Figure 2

[0052] ​​​Compared to previous architectures proposed for small-sample tasks, the Swin-Transformer V2 F6 architecture boasts no training parameters and is directly found through a reasonable FAS. This architecture, directly inspired by Swin-Transformer V2, utilizes official pre-trained weights. Furthermore, with only 10 samples per sample (5W10S), its performance on PlantVillage and FSPlantPest matches that of conventional supervised classification tasks. This is a significant discovery for small-sample plant disease tasks.

[0053] S200, combined Figure 2 As shown, unlike the existing Swin-Transformer V2 network, the present invention deletes the Stage 4 block before the last Patch Merging in the Swin-Transformer V2 network;

[0054] It should be noted that Swin Transformer is an innovative neural network architecture designed for processing visual tasks. It extracts and utilizes image features by subdividing the input image into small non-overlapping blocks (called "tokens") and performing a series of complex transformations on these blocks. It should be noted that the Swin Transformer architecture can be found in

[0055] Liu, Z., et al. Swin transformer: Hierarchical vision transformer using shifted windows. in Proceedings of the IEEE / CVF international conference on computer vision. 2021.

[0056] At this point, the main components of Swin Transformer and their functions are as follows:

[0057] (1) Image block segmentation and feature embedding: Swin Transformer first divides the input image into non-overlapping blocks of fixed size, which are regarded as independent "tokens";

[0058] Each image block is processed using a 4×4 convolution kernel, generating a feature dimension of 48 (4×4×3); these features are then transformed into a higher-dimensional feature space, denoted as C, through a linear embedding layer;

[0059] (2) Structure and function of Swin Transformer Block: The first stage of the network consists of two consecutive

[0060] Swin Transformer Block, each block processes tokens while maintaining the total number of tokens (H / 4×W / 4);

[0061] The first Swin Transformer Block starts with a layer normalization (LN), followed by a multi-head self-attention mechanism (MSA), followed by another layer of LN and two layers of multi-layer perceptrons (MLPs) with GELU nonlinear functions;

[0062] In the second Swin Transformer Block, Shifted Window MSA (SW-MSA) replaces MSA, while other parts remain unchanged. This design optimizes context capture by adjusting the window position. The linear embedding layer and two consecutive Swin Transformer Blocks are marked as "Step 1";

[0063] (3) Hierarchical structure and feature integration: Swin Transformer adopts a hierarchical structure and gradually reduces the number of tokens through the "patchmerging" layer; in this process, each group of 2×2 adjacent patches is merged, and a linear transformation is applied to the merged 4C-dimensional features;

[0064] This operation effectively reduces the number of tokens by a quarter and doubles the feature dimension to 2C, thus achieving a 2-fold reduction in spatial resolution. The following Swin Transformer blocks further refine these features to maintain a resolution of H / 8 × W / 8.

[0065] The “patch merging” and subsequent two consecutive Swin Transformer Blocks are marked as “Step 2”. This process is repeated twice, marked as “Stage 3” and “Stage 4”, to achieve output resolutions of H / 16×W / 16 and H / 32×W / 32 respectively.

[0066] Swin Transformer V2 is a state-of-the-art visual Transformer architecture. It builds on the original Swin Transformer with several key improvements designed to enhance the model's processing power and scalability across window resolutions. These improvements make Swin Transformer V2 more efficient and flexible for a wide range of vision tasks, making it particularly well-suited for applications requiring advanced feature integration, such as image smoothing, object detection, and semantic segmentation.

[0067] At this time, the core improvements of Swin Transformer V2 (compared to V1) include: (1) Post-normalization layer: In the V2 version, the post-normalization layer is introduced to replace the pre-normalization layer of the previous version; this new configuration helps the model to be more stable when expanding its capacity and simplifies the training process; (2) Scaled Cosine Attention Mechanism: The scaled cosine attention mechanism introduced in the V2 version replaces the traditional dot-product attention; scaled cosine attention improves the flexibility and efficiency of the model in processing features of different scales by adjusting the scaling factor of the attention mechanism; (3) Log-Spaced Continuous Relative Position Bias: V2 adopts a new position bias method, which expresses the relative position relationship in a logarithmic spaced continuous form. This method has better transferability between different window resolutions and improves the accuracy of the model when processing large-scale inputs; these innovative features make Swin Transformer While preserving local information, V2 can also effectively capture global context, which is particularly critical for image processing tasks. In image smoothing applications, these characteristics of the Swin Transformer V2 can achieve accurate processing of high-frequency and low-frequency information in the image, optimizing the overall visual effect of the image.

[0068] S300, combined Figure 4 and Figure 5 As shown, the plant disease feature space calibration is performed based on the network structure Swin-Transformer V2F6 obtained in step S100 and the Swin-Transformer V2 network designed in step S200, specifically including:

[0069] S310, the query set Q and the support set S are calculated by Swin-Transformer V2 F6 to obtain the feature vector Q features 、S features, the training set P of the PlantVillage dataset is transformed into a feature vector set P through SwinTransformerV2F6 features , and as a benchmark, the eigenvector S features Perform calibration;

[0070] S320, for category i, the set of all characteristic vectors of PlantVillage is P features The support set for this class in turn Find the L2 distance d i , sort all distances from small to large to get Take the first m vectors with the smallest distance after sorting Calculate its center of mass

[0071] Determine all the centroids c used for calibration i Is the calculation completed? If it is completed, continue to the next step. If not, repeat this step until all are completed.

[0072] S330, we will calculate the centroid of all categories based on this method For class i, we first calculate the class centroid c i and The L2 distance is obtained by sorting it from small to large. Take the two distances with the smallest distance after sorting The corresponding center of mass c a 、c b , take a point T on the line connecting the two centroids, and its position is determined by the hyperparameter λ; at this time, Move γ units along the direction of itself and T to complete the support set The calibrated support set is obtained

[0073] Repeat the above steps until all support set features are completed The calibration is completed;

[0074] S340, finally calculate all calibrated support sets and the query set feature vector q features The distance d i , we can output the label corresponding to the query set q: For 5-way-5-shot and 5-way-10-shot tasks, just The average value of the support set features can be used.

[0075] Specific implementation scheme 2: The present invention provides a plant disease small sample classification system based on feature space calibration, which has a program module corresponding to the above steps and executes the steps in the above plant disease small sample classification method based on feature space calibration during operation.

[0076] The other combinations and connection relationships of this embodiment are the same as those of the first embodiment.

[0077] Specific implementation scheme three: The present invention provides a computer-readable storage medium, which stores a computer program. The computer program is configured to implement the steps of a small sample classification method for plant diseases based on feature space calibration when called by a processor.

[0078] The other combinations and connection relationships of this embodiment are the same as those of the first embodiment.

[0079] Comparative experiment

[0080] Hardware: The experiments were conducted on a computer equipped with an Intel i7 processor and 16GB of RAM.

[0081] Software environment: Use Python 3.9, relying on NumPy and PyTorch libraries.

[0082] Dataset: Public image classification datasets used include the PlantVillage dataset, the PlantDoc dataset, the Plant Real-World dataset, and the Plant&Pest dataset.

[0083] Experimental process: After the software environment has been installed, follow the process Figure 2 By executing the algorithm, the effects mentioned in the invention can be achieved.

[0084] Application scenarios: This algorithm is suitable for plant disease classification and computer image processing scenarios.

[0085] The Swin-Transformer V2 F6 is used as a fixed feature extractor, and the feature space calibration algorithm is used for feature space calibration of the features extracted by the structure. Meanwhile, the performance results brought by the joint action of Swin-Transformer V2 F6 and the feature space calibration algorithm are compared with the methods in the current Plant Disease Few Shot (PDFS) field. The evaluation index for comparison is the standard evaluation index in small sample learning: the classification accuracy of 5-way-1-shot, 5-way-5-shot and 5-way-10-shot 3 tasks. We use various methods in the PDFS task in the past 3 years for comparison, and the comparison results of various methods on 4 data sets are shown in Figures 6 to 9 From the results, it is not difficult to see that the classification accuracy of the Swin-Transformer V2 F6 feature space calibration structure on the PDFS task is better than that of the new methods in the past few years.

[0086] In Figures 6 to 9 , the source of the method used is:

[0087] The source of Li er al. (2021) is Y. Li and J. Yang, “Meta-learning baselines and database for few-shot classification in agriculture,” Comput. Electron. Agric., vol. 182, p. 106055, 2021. [Online]. Available: https: / / api.semanticscholar.org / CorpusID:232342491.

[0088] The source of Yang et al. (2022) is J. Yang, Y. shuang Yang, Y. Li, S. Xiao, and S. Ercis, li, “Image information contribution evaluation for plant diseases classification via inter-class similarity,” Sustainability, 2022. [Online]. Available: https: / / api.semanticscholar.org / CorpusID:252060364.

[0089] Huang et al. (2022) is available at Y. Huang, F. Chang, Y. Tao, Y. Zhao, L. Ma, and H. Su, “Few-shot learning based on attn-cutmix and task-adaptive transformer for the recognition of cotton growth state,” Comput. Electron. Agric., vol. 202, p. 107406, 2022. [Online]. Available: https: / / api.semanticscholar.org / CorpusID: 252708627.

[0090] Zhang et al. (2021) is available at D. Zhang, F. Pan, Q. Diao, X. Feng, W. Li, and J. Wang, “Seeding crop detection framework using prototypical network method in uav images,” Agriculture, 2021. [Online]. Available: https: / / api.semanticscholar.org / CorpusID: 245562469.

[0091] Rezaei et al. (2024) is available at M. Rezaei, D. Diepeveen, H. Laga, M. G. K. Jones, and F. Sohel, “Plant disease recognition in a low data scenario using few-shot learning,” Comput. Electron. Agric., vol. 219, p. 108812, 2024. [Online]. Available: https: / / api.semanticscholar.org / CorpusID: 268242967.

[0092] Although the present disclosure has been disclosed as above, the protection scope of the present disclosure is not limited to this. Various changes and modifications can be made by those skilled in the art without departing from the spirit and scope of the present disclosure, and these changes and modifications shall fall within the protection scope of the present disclosure.

Claims

1. A small sample classification method for plant diseases based on feature space calibration, characterized in that: The following steps are involved: S100, based on the Swin-Transformer V2 network, after removing the fully connected layer, sequentially input multiple feature vectors and run multiple plant disease datasets, and score each feature output layer in the Swin-Transformer V2 network using a feature layer evaluation index, wherein the feature layer evaluation index is the feature adaptation score FAS; No. The feature adaptation score FAS of the layer is : ； in, is the between-class variance; is the intra-class variance; is the average inter-class distance; l = 1, 2, ..., 𝐿; S200, select the output result FAS with the highest score in step S100 as the feature vector, delete the Stage4 block before the last Patch Merging layer in the Swin-TransformerV2 network, and obtain the Swin-Transformer V2F6 network; S300, performing plant disease feature space calibration based on the network structure Swin-Transformer V2 F6 obtained in step S200 and the Swin-Transformer V2 network designed in step S100; Specifically include: S310, query set With support set The feature vector is calculated by Swin-Transformer V2 F6 , the training set of the PlantVillage dataset After SwinTransformerV2F6, the feature vector set is obtained , and as a benchmark, the eigenvector Perform calibration; S320, for categories , the entire set of PlantVillage's feature vectors The support set for this class in turn Find the L2 distance , sort all distances from small to large to get , take the first one with the smallest distance after sorting vectors , calculate its center of mass ; Determine all the centroids used for calibration Is the calculation completed? If it is completed, continue to the next step. If not, repeat this step until all are completed. S330, the centroids of all categories calculated in step S320 are Calibration, for category , first calculate the category centroid and The L2 distance of and sorting it from small to large, we get , take the two distances with the smallest distance after sorting and The corresponding centroids are and , select a point T on the line connecting the two centroids; Move along the direction that is collinear with T units, complete the support set The calibrated support set is obtained ; Repeat the above steps until all support set features are completed calibration; S340, calculate all calibrated support sets and the query set feature vector distance , the output is the query set Corresponding tags: .

2. The method for plant disease small sample classification based on feature space calibration according to claim 1, characterized in that: In step S100 , the plant disease datasets include the PlantVillage dataset, the PlantDoc dataset, the PlantReal-World dataset, and the Plant&Pest dataset. Seven feature vectors are input into the above four datasets and then FAS scoring is performed.

3. The method for plant disease small sample classification based on feature space calibration according to claim 2, characterized in that: The output of the 6th layer after FAS scoring is used as the feature vector for small sample analysis.

4. The method for plant disease small sample classification based on feature space calibration according to claim 3, characterized in that: For 5-way-5-shot and 5-way-10-shot tasks, as the average of the support set features.

5. The method for plant disease small sample classification based on feature space calibration according to claim 4, characterized in that: The T point is determined by the hyperparameter Determine its location.

6. The method for plant disease small sample classification based on feature space calibration according to claim 5, characterized in that: The calibrated support set is .

7. A plant disease small sample classification system based on feature space calibration, characterized by: The system has a program module corresponding to the steps of any one of claims 1 to 6, and executes the steps of the plant disease small sample classification method based on feature space calibration when running.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is configured to implement the steps of the plant disease small sample classification method based on feature space calibration according to any one of claims 1 to 6 when called by a processor.

Images