Deep learning-based potato fusarium wilt identification method
Through deep learning and Shuffle Attention module, the YOLOv8n model is improved, which solves the problem of time-consuming and labor-consuming traditional manual identification, and realizes fast and accurate potato blight detection, improving detection efficiency and accuracy.
Patent Information
- Application Number
- CN202510269026.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-07
- Publication Date
- 2025-07-25
AI Technical Summary
The traditional method of identifying potato blight relies on manual observation, which is time-consuming and labor-intensive and susceptible to subjective factors, making it difficult to achieve efficient and accurate disease detection.
Using a deep learning approach, the YOLOv8n model was improved using the Shuffle Attention module, and the model was trained to identify healthy, early and late blight by collecting and pre-processing the potato blight image dataset.
It realizes fast and accurate potato blight recognition, improves detection efficiency and model accuracy, and reduces calculation complexity and parameter quantity.
Smart Images
Figure CN120375032A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of computer vision technology, and in particular, to a method for identifying potato wilt disease based on deep learning. Background Art
[0002] With the continuous development of agricultural technology, the early detection and intelligent management of crop diseases have become a crucial part of agricultural production. As one of the important food crops globally, the health status of potatoes directly affects the yield and quality. Potato wilt disease is a serious disease, and its symptoms include yellowing and withering of leaves, as well as browning of the vascular bundles of underground stems and tubers. Traditional disease identification methods usually rely on manual observation, which is not only time-consuming and laborious but also easily affected by subjective factors. Summary of the Invention
[0003] Based on this, it is necessary to provide a method for identifying potato wilt disease based on deep learning to solve the technical problems mentioned in the above background art.
[0004] The present invention adopts the following technical solutions:
[0005] A method for identifying potato wilt disease based on deep learning, the steps are as follows:
[0006] S1: Collect potato wilt disease images at different times to form a dataset and perform preprocessing;
[0007] S2: Introduce Shuffle Attention into the backbone network of YOLOv8n and use the training set images to train the improved YOLOv8n model;
[0008] S3: Input the test set pictures into the improved YOLOv8n network;
[0009] S4: Output the recognition result map, detection accuracy, number of parameters, etc. of the improved model.
[0010] As a preferred embodiment of the method for identifying potato wilt disease based on deep learning provided by the present invention, in the S1 step, the images in the dataset are 3546 potato wilt disease images.
[0011] As a preferred embodiment of the method for identifying potato wilt disease based on deep learning provided by the present invention, in the S1 step, the collected dataset is sheared and the size of the dataset is adjusted to complete the preprocessing.
[0012] As a preferred embodiment of the potato blight recognition method based on deep learning provided by the present invention, in the step S1, the images in the preprocessed data set are classified into healthy, early blight, and late blight.
[0013] As a preferred embodiment of the potato blight recognition method based on deep learning provided by the present invention, in the step S2, Shuffle Attention is introduced into the backbone network of YOLOv8n, and the steps are as follows:
[0014] The attention mechanism simulates human visual attention, which enables the neural network to precisely focus on the target area in the image and allocate more computing resources for in-depth exploration of the rich details of the target, while suppressing redundant information. Currently, spatial attention and channel attention are widely used in computer vision research to obtain the spatial relationship and channel dependence relationship of target features. The present invention introduces Shuffle Attention into the backbone network, which effectively combines two types of attention mechanisms and minimizes computing resources.
[0015] The calculation formula for the feature output after being processed by the channel attention mechanism is as follows:
[0016] X i ′1 = σ(F c (s))·X i1 = σ(W1s + b1)·X i1
[0017] Among them, F c represents a linear function, W1 ∈ R C / 2G×1×1 and b1 ∈ R C / 2G×1×1 respectively represent the weight and bias for linearly processing s.
[0018] The calculation formula for the feature output after being processed by the spatial attention mechanism is as follows:
[0019] X i ′2 = σ(W2·GN(X i2 )) + b2)·X i2
[0020] Among them, W2 ∈ R C / 2G×1×1 and b2 ∈ R C / 2G×1×1 respectively represent the weight and bias for linearly processing X i2 ).
[0021] The features output from the two attention branches are combined into new sub-features. In addition, the new sub-features are aggregated into features of the complete channel, and finally, the complete features achieve cross-channel information exchange through channel shuffle operation to enhance the association between sub-features.
[0022] It can be clearly seen that through the above technical solutions of the present application, the technical problems to be solved by the present application can surely be solved.
[0023] Meanwhile, through the above technical solutions, the present invention has at least the following beneficial effects:
[0024] A method for identifying potato fusarium wilt based on deep learning provided by the present invention can quickly and accurately identify potato fusarium wilt, thus greatly improving the detection efficiency;
[0025] The present invention introduces the Shuffle Attention module into the backbone network of YOLOv8n, achieving the improvement of the target detection accuracy without changing the computational complexity and the number of parameters, enhancing the attention of the model to the useful information of the image, effectively suppressing the useless information of the image, and improving the training accuracy of the model. Description of the Drawings
[0026] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.
[0027] Figure 1 It is the program flow chart of the potato fusarium wilt identification and detection based on deep learning of the present invention. Detailed Embodiments
[0028] In order to make the purpose, technical solutions and advantages of the present invention clearer, the following further details the present invention with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.
[0029] In order to enable those skilled in the art of this technology to better understand the solution of the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings.
[0030] It should be noted that, without conflict, the embodiments in the present invention and the features and technical solutions in the embodiments can be combined with each other.
[0031] Embodiment 1
[0032] Refer to Figure 1 , a method for identifying potato fusarium wilt based on deep learning, the steps are as follows:
[0033] Collect potato blight images. A total of 3546 potato blight images were collected, and the pictures were marked and the dataset was classified into: healthy, early blight, and late blight.
[0034] Perform operations such as cropping and resizing the marked pictures on the dataset.
[0035] Introduce the Shuffle Attention module into the YOLOv8n model, import the processed training set pictures into the improved model for training, and then use the trained model to test the performance of the model with the test set images.
[0036] Example 2
[0037] Disclosed on the basis of the above Example 1.
[0038] Table 1: Results after improvement
[0039]
[0040] Compare the improvement results of introducing Shuffle Attention into the YOLOv8n backbone network with the results before introduction
[0041] Finally, the present invention uses three evaluation indicators: mean Average Precision (mAP), number of parameters, and computational complexity. The calculation formula for mAP is as follows
[0042]
[0043] Among them, TP, FP, and FN represent the number of positive samples, negative samples, and missed detection samples respectively; P and R represent the accuracy rate and recall rate respectively; AP represents the area enclosed by the P-R curve of a certain category and the coordinate axes; C represents the total number of categories.
[0044] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor limit the present invention to the specific embodiments described. Obviously, many modifications and changes can be made according to the content of this specification. These embodiments are selected and specifically described in this specification to better explain the principle and practical application of the present invention, so that those skilled in the relevant technical field can well understand and utilize the present invention. The present invention is only limited by the claims and their full scope and equivalents.
Claims
1. A method for identifying potato blight based on deep learning, characterized in that, The steps are as follows: S1: Collect potato blight images, form a dataset, and perform preprocessing; S2: Introduce Shuffle Attention into the backbone network of YOLOv8n, and use the training set images to train the improved YOLOv8n model; S3: Input the test set images into the improved YOLOv8n network; S4: Output the recognition result map, detection accuracy, number of parameters, etc. of the improved model.
2. The potato blight recognition method based on deep learning according to claim 1, wherein In the step S1, the images in the dataset are 3546 banana maturity images.
3. The potato blight recognition method based on deep learning according to claim 1, characterized in that, In the step S1, the collected dataset is sheared and the size of the dataset is adjusted to complete the preprocessing.
4. A method for identifying potato blight based on deep learning according to claim 1, characterized in that, In the step S1, the images in the preprocessed dataset are classified into healthy, early blight, and late blight.
5. A method for potato blight based on deep learning according to claim 1, characterized in that, In the step S2, the steps of introducing Shuffle Attention into the backbone network of YOLOv8n are as follows: The attention mechanism simulates human visual attention. It enables the neural network to precisely focus on the target area in the image and allocate more computing resources for in-depth exploration of the rich details of the target, while suppressing redundant information. Currently, spatial attention and channel attention are widely used in computer vision research to obtain the spatial relationship and channel dependence relationship of target features. In the present invention, Shuffle Attention is introduced into the backbone network, which effectively combines two types of attention mechanisms and minimizes computing resources.