A crop disease identification method, system, device and storage medium
By introducing pyramid convolution and attention mechanisms into crop disease identification, the problems of time consumption and large computational load in traditional methods are solved, and efficient and accurate disease identification is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SOUTH CHINA UNIV OF TECH
- Filing Date
- 2023-01-18
- Publication Date
- 2026-05-29
AI Technical Summary
Traditional methods for identifying crop diseases are time-consuming and rely on manual intervention, making them difficult to train. Furthermore, deep learning models require a large amount of computation and are difficult to apply in real-world scenarios.
A crop disease identification method based on pyramid convolution and attention mechanism is adopted. The 3×3 convolution is replaced by the CSP ResNet50 backbone network. The pyramid convolution and attention mechanism modules are combined to enhance multi-scale information extraction and feature attention, and reduce the amount of computation.
It improves the accuracy and efficiency of crop disease identification, reduces the computational burden, and achieves more efficient disease identification.
Smart Images
Figure CN115965874B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of crop disease image recognition technology, and in particular to a crop disease recognition method, system, device and storage medium. Background Technology
[0002] In traditional agricultural production practices, due to the loss of grassroots agricultural technicians and farmers' lack of relevant professional knowledge, farmers mainly rely on experience to observe crop diseases with the naked eye and then make subjective judgments. However, this method is not very effective and cannot make accurate judgments. It is easy to cause the abuse of chemical pesticides, which seriously affects the quality and output of crops, causing economic losses and unnecessary environmental pollution.
[0003] Traditional crop disease image recognition mainly consists of four parts. First, image preprocessing, including image scaling, noise reduction, and histogram equalization; next, image segmentation to remove information irrelevant to lesion identification; then, feature extraction, using a series of algorithms to extract texture, color, and shape features from the crop disease image; and finally, disease classification using machine learning algorithms such as support vector machines, K-nearest neighbors, Bayesian classification, and K-means. However, this method still has many drawbacks: firstly, image preprocessing is very time-consuming; secondly, the extraction of crop disease features largely relies on manual intervention, which cannot guarantee the accuracy of disease features; and finally, traditional machine learning algorithms require large amounts of data and are difficult to train. These limitations directly prevent traditional methods from being directly applicable to real-world crop disease recognition scenarios.
[0004] In recent years, due to the rapid development of deep learning technology, it has made great progress and been widely used in the field of crop disease identification. Among them are networks that use attention mechanisms and integrate multi-scale information, but they all introduce additional computational costs to varying degrees, increasing the number of model parameters and making it difficult for the network to converge during training.
[0005] In conclusion, it is necessary to propose a new crop disease identification network that fully utilizes features across multiple scales and integrates contextual information from different scales to improve the identification accuracy of the network model. Summary of the Invention
[0006] In order to at least partially solve one of the technical problems existing in the prior art, the present invention aims to provide a method, system, device and storage medium for identifying crop diseases based on pyramid convolution and attention mechanisms.
[0007] The technical solution adopted in this invention is:
[0008] A method for identifying crop diseases includes the following steps:
[0009] Obtain a crop disease dataset, preprocess and augment the dataset to obtain a preprocessed dataset;
[0010] A disease identification model is constructed, and the obtained dataset is used to train the disease identification model. The trained model is then used for crop disease identification.
[0011] The disease identification model uses CSP ResNet50 as its backbone network. The residual blocks in the backbone network are replaced with pyramidal convolutions instead of 3×3 convolutions. This increases the receptive field of the convolutional kernels without increasing computational cost or model size, improving the network model's ability to extract multi-scale information. It also integrates feature information from receptive fields of different sizes to enhance the overall performance of the model. Furthermore, an attention mechanism module is added to the residual blocks of the backbone network to establish dependencies between feature map channels and correlations between pixels. This allows the model to focus more on important disease features and weaken irrelevant interference factors, thereby improving the model's ability to identify crop diseases.
[0012] Furthermore, the preprocessing and augmentation of the dataset includes:
[0013] Randomly crop, randomly flip horizontally, randomly rotate, or randomly erase the images in the dataset;
[0014] The images in the dataset are normalized, and the size of the input images is uniformly adjusted to 224×224 during training.
[0015] Furthermore, the pyramid convolution consists of S groups of convolution kernels of different sizes, and grouped convolution is used to reduce the amount of computation. The specific calculation process is expressed as follows:
[0016] X i =Split S (X)
[0017] k i =2(i+1)+1
[0018]
[0019]
[0020] F = Concat(F0, F1, ..., F) S-1 )
[0021] Where i = 0, 1, 2, ..., S-1, X ∈ R C×H×W Split represents the input feature map. S This means that the input feature map is uniformly divided into S groups according to the channel dimension, Xi ∈R C′×H×W This represents the partitioned input feature map, where k i G represents the size of the i-th convolutional kernel. i For a convolution kernel size of k i The number of convolutional groups, Indicates a size of k i The number of groups is G i The convolution kernel, F i ∈R C′×H×W This indicates that the kernel size is k. i The resulting feature map, where Concat represents the feature concatenation operation, F∈R C×H×W This represents the output feature map after pyramid convolution.
[0022] Furthermore, without introducing additional parameters, channel attention units are used to establish dependencies between feature map channels, spatial attention units are used to establish correlations between pixels, and the input feature map is grouped along the channel dimension into multiple sub-feature maps. The specific calculation process is expressed as follows:
[0023] X k =Split G (X)
[0024] (X k1 X k2 ) = Split2(X k )
[0025] X′ k1 =U CA (X k1 )
[0026] X′ k2 =U SA (X k2 )
[0027] X′ k =Concat(X′) k1 , X′ k2 )
[0028] X′=ChannelShuffle(Concat(X′1,X′2,…,X′ G ))
[0029] Where k = 1, 2, ..., G, X ∈ R C×H×W Split represents the input feature map. G This means that the input feature map is uniformly divided into G groups according to the channel dimension. This means that each sub-feature map is uniformly divided into two parts according to the channels, UCA Represented as a channel attention unit, U SA Represented as a spatial attention unit, Let X′ represent the outputs of the channel attention unit and the spatial attention unit, respectively. Let Concat represent the feature concatenation operation and ChannelShuffle represent the channel shuffling operation. Let X′∈R C×H×W This represents the output feature map after concatenating all sub-feature maps and then performing a channel shuffling operation.
[0030] Furthermore, the attention mechanism module includes a channel attention unit and a spatial attention unit, and the specific calculation process expression is as follows:
[0031] U CA =σ(W1(P) AVG (X)+P MAX (X))+b1)·X
[0032] U SA =σ(W2·GN(X)+b2)·X
[0033] Where X represents the input feature map, σ represents the non-linear activation function Sigmoid, and P AVG Represented as global average pooling, P MAX This is represented as global max pooling, where W1, W2, b1, and b2 represent the parameters of the linear transformation, and GN represents group normalization.
[0034] Furthermore, training the disease identification model using the obtained dataset includes:
[0035] The network parameters were updated using the RMSProp optimizer, with the initial learning rate set to 0.001, momentum set to 0.9, and weight decay set to 0.0001. The cross-entropy loss function was used, and the cosine annealing learning rate update strategy was employed. The training was conducted for 200 epochs, and after each training iteration, the model was tested on the validation set. The model with the highest accuracy on the validation set was saved.
[0036] Another technical solution adopted in this invention is:
[0037] A crop disease identification system, comprising:
[0038] The data acquisition module is used to acquire crop disease datasets, preprocess and augment the datasets to obtain preprocessed datasets.
[0039] The model training module is used to build a disease identification model. The obtained dataset is used to train the disease identification model, and the trained model is used for crop disease identification.
[0040] The disease identification model uses CSP ResNet50 as its backbone network. The residual blocks in the backbone network are replaced with pyramidal convolutions instead of 3×3 convolutions. This increases the receptive field of the convolutional kernels without increasing computational cost or model size, improving the network model's ability to extract multi-scale information. It also integrates feature information from receptive fields of different sizes to enhance the overall performance of the model. Furthermore, an attention mechanism module is added to the residual blocks of the backbone network to establish dependencies between feature map channels and correlations between pixels. This allows the model to focus more on important disease features and weaken irrelevant interference factors, thereby improving the model's ability to identify crop diseases.
[0041] Another technical solution adopted in this invention is:
[0042] A crop disease identification device, comprising:
[0043] At least one processor;
[0044] At least one memory for storing at least one program;
[0045] When the at least one program is executed by the at least one processor, the at least one processor implements the method described above.
[0046] Another technical solution adopted in this invention is:
[0047] A computer-readable storage medium storing a processor-executable program, which, when executed by a processor, performs the method described above.
[0048] The beneficial effects of this invention are as follows: By employing a pyramid convolution module, this invention increases the receptive field of the convolution kernel without increasing computational load or model size, thereby improving the network model's ability to extract multi-scale information and enhancing the overall performance of the model by fusing feature information from receptive fields of different sizes. Furthermore, by employing an attention mechanism module, it models the dependencies between feature map channels and the correlations between pixels, extracting and enhancing spatial and semantic information in the feature maps to weaken irrelevant interference factors. This allows the model to focus more on important disease features, improving its ability to identify crop diseases. Attached Figure Description
[0049] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following description is provided with accompanying drawings of the relevant technical solutions in the embodiments of the present invention or the prior art. It should be understood that the accompanying drawings described below are only for the purpose of clearly illustrating some embodiments of the technical solutions of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0050] Figure 1 This is a schematic diagram of a crop disease identification network based on pyramid convolution and attention mechanisms in an embodiment of the present invention;
[0051] Figure 2 This is a residual module diagram of pyramid convolution and attention mechanism in an embodiment of the present invention;
[0052] Figure 3 This is a diagram of the pyramid convolution module in an embodiment of the present invention;
[0053] Figure 4 This is a block diagram of the attention mechanism in an embodiment of the present invention;
[0054] Figure 5 This is a diagram of the channel attention unit and the spatial attention unit in an embodiment of the present invention. Detailed Implementation
[0055] The embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. The step numbers in the following embodiments are set only for ease of explanation, and there is no limitation on the order between the steps. The execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.
[0056] In the description of this invention, it should be understood that the orientation descriptions, such as up, down, front, back, left, right, etc., are based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention.
[0057] In the description of this invention, "several" means one or more, "more than" means two or more, "greater than," "less than," and "exceeding" are understood to exclude the stated number, while "above," "below," and "within" are understood to include the stated number. The use of "first" and "second" in the description is merely for distinguishing technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the order of the indicated technical features.
[0058] In the description of this invention, unless otherwise explicitly defined, terms such as "setting," "installing," and "connecting" should be interpreted broadly, and those skilled in the art can reasonably determine the specific meaning of the above terms in this invention in conjunction with the specific content of the technical solution.
[0059] like Figure 1 As shown, this embodiment provides a crop disease identification method based on pyramid convolution and attention mechanisms, including the following steps:
[0060] Step 1: Obtain the crop disease dataset, and perform preprocessing and data augmentation on the dataset to obtain the preprocessed dataset. Specifically, the dataset is preprocessed and augmented, including random cropping, random horizontal flipping, random rotation, random erasure, and normalization. Finally, during training, the input images are uniformly adjusted to 224×224.
[0061] The disease dataset in this embodiment comes from the publicly available AI Challenger 2018 dataset, which contains 61 different diseases across 10 crops. The crops include apples, cherries, corn, grapes, citrus, peaches, peppers, potatoes, strawberries, and tomatoes. The diseases include black spot, gray spot, powdery mildew, and mosaic virus, with severity levels ranging from mild to severe. After cleaning and removing two categories with insufficient data and duplicate images, a total of 36,020 images remain. The dataset is randomly divided into a training set (25,258 images, 70% of the total), a validation set (7,182 images, 20% of the total), and a test set (3,580 images, 10% of the total).
[0062] Step 2: Use CSP ResNet50 as the backbone network, see [link / reference] Figure 2 This approach replaces the 3×3 convolutions in the residual blocks of the backbone network with pyramid convolutions. Without increasing computation or model size, it increases the receptive field of the convolution kernels, improving the network model's ability to extract multi-scale information. It also integrates feature information from receptive fields of different sizes to enhance the overall performance of the model. The pyramid convolution module is as follows: Figure 3 As shown, it consists of S groups of convolutional kernels of different sizes, and uses grouped convolution to reduce computation. The specific calculation process is expressed as follows:
[0063] X i =Split S (X)
[0064] k i =2(i+1)+1
[0065]
[0066]
[0067] F = Concat(F0, F1, ..., F) S-1 )
[0068] Where i = 0, 1, 2, ..., S-1, X ∈ R C×H×WSplit represents the input feature map. S This means that the input feature map is uniformly divided into S groups according to the channel dimension, X i ∈R C′×H×W This represents the partitioned input feature map, where k i G represents the size of the i-th convolutional kernel. i For a convolution kernel size of k i The number of convolutional groups, Indicates a size of k i The number of groups is G i The convolution kernel, F i ∈R C′×H×W This indicates that the kernel size is k. i The resulting feature map, where Concat represents the feature concatenation operation, F∈R C×H×W This represents the output feature map after pyramid convolution.
[0069] Step 3: Add an attention mechanism module to the residual blocks of the backbone network to establish dependencies between feature map channels and correlations between pixels. This allows the model to focus more on important disease features and weaken irrelevant interference factors, thereby improving the model's ability to identify crop diseases. The attention mechanism module is as follows: Figure 4 As shown, the input feature map is grouped along the channel dimension into multiple sub-feature maps. The specific calculation process is expressed as follows:
[0070] X k =Split G (X)
[0071] (X k1 X k2 ) = Split2(X k )
[0072] X′ k1 =U CA (X k1 )
[0073] X′ k2 =U SA (X k2 )
[0074] X′ k =Concat(X′) k1 , X′ k2 )
[0075] X′=ChannelShuffle(Concat(X′1,X′2,…,X′ G ))
[0076] Where k = 1, 2, ..., G, X ∈ R C×H×W Split represents the input feature map. G This means that the input feature map is uniformly divided into G groups according to the channel dimension. This means that each sub-feature map is uniformly divided into two parts according to the channels, U CA Represented as a channel attention unit, U SA Represented as a spatial attention unit, Let X′ represent the outputs of the channel attention unit and the spatial attention unit, respectively. Let Concat represent the feature concatenation operation and ChannelShuffle represent the channel shuffling operation. Let X′∈R C×H×W This represents the output feature map after concatenating all sub-feature maps and then performing a channel shuffling operation.
[0077] The attention mechanism module includes channel attention units and spatial attention units, such as... Figure 5 As shown, the specific calculation process is expressed as follows:
[0078] U CA =σ(W1(P) AVG (X)+P MAX (X))+b1)·X
[0079] U SA =σ(W2·GN(X)+b2)·X
[0080] Where X represents the input feature map, σ represents the non-linear activation function Sigmoid, and P AVG Represented as global average pooling, P MAX This is represented as global max pooling, where W1, W2, b1, and b2 represent the parameters of the linear transformation, and GN represents group normalization.
[0081] Step 4: Train the model processed in Steps 2 and 3 using the dataset obtained in Step 1, and save the trained model. The specific training method is as follows: Update the network parameters using the RMSProp optimizer, initialize the learning rate to 0.001, momentum to 0.9, and weight decay to 0.0001, use the cross-entropy loss function, and employ a cosine annealing learning rate update strategy. Train for 200 epochs, and after each training iteration, test on the validation set, saving the model with the highest accuracy on the validation set.
[0082] To objectively evaluate the performance of the network model proposed in this embodiment, its experimental data was compared with several popular network models. The experimental results of different models on the disease dataset are shown in Table 1.
[0083] Table 1. Experimental comparison of different network models on crop disease datasets.
[0084]
[0085]
[0086] As shown in Table 1, the crop disease identification model proposed in this example improves the network model's ability to extract multi-scale information by introducing pyramid convolution, and further enhances the model's accuracy by incorporating an attention mechanism module, enabling the network to accurately locate diseased areas. Experimental results show that the crop disease identification model proposed in this example achieves an accuracy of 85.25%. Furthermore, the model achieves better identification results with significantly lower parameter and computational costs compared to other excellent methods, further validating the effectiveness of the proposed model.
[0087] This embodiment also provides a crop disease identification system, including:
[0088] The data acquisition module is used to acquire crop disease datasets, preprocess and augment the datasets to obtain preprocessed datasets.
[0089] The model training module is used to build a disease identification model. The obtained dataset is used to train the disease identification model, and the trained model is used for crop disease identification.
[0090] The disease identification model uses CSP ResNet50 as its backbone network. The residual blocks in the backbone network are replaced with pyramidal convolutions instead of 3×3 convolutions. This increases the receptive field of the convolutional kernels without increasing computational cost or model size, improving the network model's ability to extract multi-scale information. It also integrates feature information from receptive fields of different sizes to enhance the overall performance of the model. Furthermore, an attention mechanism module is added to the residual blocks of the backbone network to establish dependencies between feature map channels and correlations between pixels. This allows the model to focus more on important disease features and weaken irrelevant interference factors, thereby improving the model's ability to identify crop diseases.
[0091] This embodiment of the crop disease identification system can execute the crop disease identification method provided in the method embodiment of the present invention, and can execute any combination of the implementation steps of the method embodiment, and has the corresponding functions and beneficial effects of the method.
[0092] This embodiment also provides a crop disease identification device, including:
[0093] At least one processor;
[0094] At least one memory for storing at least one program;
[0095] When the at least one program is executed by the at least one processor, the at least one processor implements the method described above.
[0096] This embodiment of a crop disease identification device can execute a crop disease identification method provided in the method embodiment of the present invention, and can execute any combination of the implementation steps of the method embodiment, and has the corresponding functions and beneficial effects of the method.
[0097] This embodiment also provides a storage medium storing instructions or programs that can execute the crop disease identification method provided in the method embodiment of the present invention. When the instructions or programs are run, any combination of implementation steps of the method embodiment can be executed, and the method has the corresponding functions and beneficial effects.
[0098] In some alternative embodiments, the functions / operations mentioned in the block diagrams may not occur in the order shown in the operation diagrams. For example, depending on the functions / operations involved, two consecutively shown blocks may actually be executed substantially simultaneously, or the blocks may sometimes be executed in reverse order. Furthermore, the embodiments presented and described in the flowcharts of this invention are provided by way of example to provide a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented herein. Alternative embodiments are contemplated in which the order of various operations is altered and sub-operations described as part of a larger operation are executed independently.
[0099] Furthermore, although the invention has been described in the context of functional modules, it should be understood that, unless otherwise stated, one or more of the described functions and / or features may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the invention. Rather, given the properties, functions, and internal relationships of the various functional modules in the apparatus disclosed herein, the actual implementation of the module will be understood within the scope of conventional skill of an engineer. Therefore, those skilled in the art can implement the invention as set forth in the claims using ordinary techniques without excessive experimentation. It is also understood that the specific concepts disclosed are merely illustrative and not intended to limit the scope of the invention, which is determined by the full scope of the appended claims and their equivalents.
[0100] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0101] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.
[0102] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.
[0103] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.
[0104] In the foregoing description of this specification, references to terms such as "one embodiment," "another embodiment," or "some embodiments" indicate that a specific feature, structure, material, or characteristic described in connection with an embodiment or example is included in at least one embodiment or example of the present invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.
[0105] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.
[0106] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the above embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.
Claims
1. A method for identifying crop diseases, characterized in that, Includes the following steps: Obtain a crop disease dataset, preprocess and augment the dataset to obtain a preprocessed dataset; A disease identification model is constructed, and the obtained dataset is used to train the disease identification model. The trained model is then used for crop disease identification. The disease identification model uses CSP ResNet50 as the backbone network, and replaces the 3x3 convolutions in the residual blocks of the backbone network with pyramid convolutions.
3. Convolution; Add an attention mechanism module to the residual blocks of the backbone network; The pyramid convolution consists of S groups of convolution kernels of different sizes, and grouped convolution is used to reduce computation. The specific calculation process is expressed as follows: in, , Indicates the input feature map, This means that the input feature map is uniformly divided according to the channel dimension. Group, This represents the input feature map after partitioning, where , Indicates the first Group convolution kernel size, The kernel size is The number of convolutional groups, Indicates size is The number of groups is convolution kernel, This indicates that the kernel size is... The feature map after processing, This indicates a feature concatenation operation. This represents the output feature map after pyramid convolution; The attention mechanism module, without introducing additional parameters, employs channel attention units to establish dependencies between feature map channels, spatial attention units to establish correlations between pixels, and groups the input feature map along the channel dimension into multiple sub-feature maps. The specific calculation process is expressed as follows: in, , Indicates the input feature map, This means that the input feature map is uniformly divided according to the channel dimension. Group, , This means that each sub-feature map is evenly divided into two parts according to each channel. Represented as a channel attention unit. Represented as a spatial attention unit, These represent the outputs of the channel attention unit and the spatial attention unit, respectively. This indicates a feature concatenation operation. This indicates a channel rinsing operation. This represents the output feature map after concatenating all sub-feature maps and then performing a channel shuffling operation.
2. The method for identifying crop diseases according to claim 1, characterized in that, The preprocessing and augmentation of the dataset includes: Randomly crop, randomly flip horizontally, randomly rotate, or randomly erase the images in the dataset; The images in the dataset are normalized, and the size of the input images is uniformly adjusted to 224 during training.
224.
3. The method for identifying crop diseases according to claim 1, characterized in that, The attention mechanism module includes a channel attention unit and a spatial attention unit, and the specific calculation process is expressed as follows: in, This is represented as the input feature map. Represented as the non-linear activation function Sigmoid, This is represented as global average pooling. This is represented as global max pooling. The parameters are represented as linear transformation parameters. This is represented as group normalization.
4. The method for identifying crop diseases according to claim 1, characterized in that, The step of training the disease identification model using the obtained dataset includes: The network parameters were updated using the RMSProp optimizer, with the initial learning rate set to 0.001, momentum set to 0.9, and weight decay set to 0.0001. The cross-entropy loss function was used, and the cosine annealing learning rate update strategy was employed. The training was conducted for 200 epochs, and after each training iteration, the model was tested on the validation set. The model with the highest accuracy on the validation set was saved.
5. A crop disease identification system for implementing the method as described in any one of claims 1-4, characterized in that, include: The data acquisition module is used to acquire crop disease datasets, preprocess and augment the datasets to obtain preprocessed datasets. The model training module is used to build a disease identification model. The obtained dataset is used to train the disease identification model, and the trained model is used for crop disease identification. The disease identification model uses CSP ResNet50 as the backbone network, and replaces the 3x3 convolutions in the residual blocks of the backbone network with pyramid convolutions.
3. Convolution; add an attention mechanism module to the residual blocks of the backbone network.
6. A crop disease identification device, characterized in that, include: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method of any one of claims 1-4.
7. A computer-readable storage medium storing a processor-executable program, characterized in that, The processor-executable program, when executed by the processor, is used to perform the method as described in any one of claims 1-4.