Crop disease and pest recognition algorithm and device based on deep learning, and storage medium

Through deep learning algorithms, crop leaves and environmental features are extracted, clustered feature vectors are calculated, and the time-consuming and laborious identification of crop diseases and pests in the existing technology is solved, efficient and accurate identification is achieved, and user experience is improved.

CN119048901APending Publication Date: 2024-11-29ZHOUKOU NORMAL UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411065884.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-05
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

The prior art cannot provide an effective deep learning-based crop pest recognition algorithm, resulting in time-consuming, laborious and inefficient identification and poor user experience.

Method used

Using a deep learning-based crop pest and disease recognition algorithm, we use crop leaf images to extract leaf feature vectors and environmental information feature vectors, calculate clustered feature vectors and membership, and use convolutional neural networks to identify to improve recognition accuracy.

Benefits of technology

It greatly improves the identification accuracy of disease categories, improves user experience, and achieves efficient and accurate identification of crop pests and diseases.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119048901A_ABST
    Figure CN119048901A_ABST
Patent Text Reader

Abstract

The invention is applicable to the technical field of plant disease and insect pest recognition algorithms, and provides a crop disease and insect pest recognition algorithm and device based on deep learning, and a storage medium. Shooting a plurality of disease leaf images containing a plurality of disease categories at a plurality of different times of the plant; extracting leaves in the plurality of disease leaf images one by one, scaling down the leaves to a preset size according to a proportion, and outputting a plurality of processed disease and pest leaf images; calculating a plurality of clustering feature vectors in one-to-one correspondence with the plurality of disease categories according to the plurality of processed pest leaf images; calculating membership degrees of the plurality of to-be-identified scab images and the plurality of clustering feature vectors; the disease category corresponding to the clustering feature vector corresponding to the maximum membership degree in the plurality of membership degrees is the disease category corresponding to the scab image to be identified; according to the method, the clustering feature vector corresponding to the disease category is obtained through training, and then the disease category corresponding to the disease spot image to be recognized is recognized, so that the recognition precision of the disease category is greatly improved, and the user experience is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of plant pest and disease identification algorithms, and in particular, relates to crop pest and disease identification algorithms, devices, and storage media based on deep learning. Background Art

[0002] Crops may be attacked by various diseases and pests during their growth period, resulting in a serious reduction in crop yield and quality.

[0003] Therefore, how to improve the effectiveness of crop disease and pest prevention is particularly important. Traditional methods of crop disease and pest identification rely more on personal experience and require field surveys to judge the characteristics of crops one by one. This is time-consuming and labor-intensive, and difficult to promote and apply on a large scale. How to identify crop diseases and pests efficiently and accurately has become an issue that must be considered in the current development of smart agriculture. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the purpose of the present invention is to provide a crop disease and pest identification algorithm, device, and storage medium based on deep learning, aiming to solve the problem that the prior art cannot provide an effective crop disease and pest identification algorithm based on deep learning, resulting in the identification of crop diseases and pests being time-consuming, labor-intensive, inefficient, and having a poor user experience.

[0005] In one aspect, the present invention provides a crop pest and disease identification algorithm based on deep learning, the algorithm comprising the following steps:

[0006] Shooting multiple diseased leaf images of plants containing multiple disease categories at multiple different times; wherein the multiple diseased leaf images have the same resolution and size, and the disease category includes the multiple diseased leaf images;

[0007] Extracting leaves from a plurality of the diseased leaf images one by one and reducing them proportionally to a preset size, and then outputting a plurality of the processed diseased and pest leaf images;

[0008] Calculate a plurality of clustering feature vectors corresponding to a plurality of the disease categories one by one according to a plurality of the processed diseased and pest leaf images;

[0009] Calculating the membership degree between a plurality of lesion images to be identified and a plurality of the clustering feature vectors;

[0010] The disease category corresponding to the cluster feature vector corresponding to the largest degree of membership among the multiple degrees of membership is the disease category corresponding to the disease spot image to be identified.

[0011] Furthermore, the step of calculating a plurality of clustering feature vectors corresponding to a plurality of disease categories based on a plurality of processed diseased and pest leaf images comprises:

[0012] Acquire environmental information when taking a plurality of images of the diseased leaves, and acquire a first environmental information feature vector based on the environmental information;

[0013] Acquire a first diseased leaf feature vector based on the processed diseased and insect-infested leaf image;

[0014] Constructing a disease joint classification feature based on the first environmental information feature vector and the first diseased leaf feature vector;

[0015] A plurality of features with greater contributions are extracted from the disease joint classification features and normalized to form a training feature vector.

[0016] Furthermore, the environmental information includes: season, temperature, humidity, rainfall and light;

[0017] The diseased leaf feature vector includes: color feature, shape feature and texture feature;

[0018] The color features include: the mean, variance, peak, skewness, energy and entropy of the color components R, G, B and H in the RGB, HIS and YCbCr color spaces of the processed diseased and pest leaf images, and the mean of Cb and Cr;

[0019] The shape features include: circularity, eccentricity, shape complexity and shape parameters of the processed pest leaf image;

[0020] The texture features include: contrast, correlation, energy, moment of inertia and entropy of the diseased spot area of ​​the processed diseased and pest leaf image;

[0021] Extracting a plurality of features with greater contributions from the disease joint classification features comprises at least: 10;

[0022] The training feature vector is expressed as: X=(χ1, χ2, ..., χ10).

[0023] Furthermore, the step of calculating a plurality of clustering feature vectors corresponding to a plurality of disease categories based on a plurality of processed diseased and pest leaf images further comprises:

[0024] Calculate the clustering feature vector corresponding to the disease category according to formula 1;

[0025] Formula 1:

[0026] in, m=1,2,...,10.

[0027] Furthermore, before calculating the cluster feature vector of the first disease category according to Formula 1, the method further includes:

[0028] The feature vector of the processed leaf image of the diseased and pest-infested leaf of the diseased category is recorded as:

[0029] Where i = 1, 2, ..., n j , j = 1, 2, ..., C, C is the number of the plurality of said disease categories, n is the number of the plurality of said processed diseased and pest leaf images of a certain said disease category, n j is the number of the processed diseased and pest leaf images of the first disease category.

[0030] Furthermore, before calculating the membership degree between the plurality of lesion images to be identified and the plurality of clustering feature vectors, the method further comprises:

[0031] Shooting a plurality of disease spot images to be identified of a plurality of disease categories at a plurality of different times of the plant; wherein the resolution and size of the plurality of disease spot images to be identified are the same as the resolution and size of the diseased leaf image, and the disease category includes the plurality of disease spot images to be identified;

[0032] Extracting leaves from a plurality of the diseased spot images to be identified one by one and reducing them proportionally to a preset size, and then outputting a plurality of processed diseased spot leaf images;

[0033] Acquire environmental information when taking a plurality of images of the lesions to be identified, and acquire a second environmental information feature vector based on the environmental information;

[0034] Acquire a second diseased leaf feature vector based on the processed diseased leaf image;

[0035] Based on the second environmental information feature vector and the second diseased leaf feature vector, a second disease joint classification feature is formed;

[0036] Extracting a plurality of features with greater contributions from the second disease joint classification features and performing normalization processing to form a test feature vector;

[0037] Extracting a plurality of features with greater contributions from the second disease joint classification features includes at least: 10;

[0038] The test feature vector is expressed as: Y = (y1, y2, ..., y 10 ).

[0039] Furthermore, the calculating of the membership degree between the plurality of lesion images to be identified and the plurality of clustering feature vectors comprises:

[0040] Calculate Y=(y1,y2,...,y 10 )and The degree of membership between represents a plurality of said cluster feature vectors.

[0041] Further, the degree of membership is calculated using one of Euclidean distance, cosine similarity, and Pearson correlation coefficient;

[0042] A convolutional neural network algorithm is used to extract leaves from the plurality of diseased leaf images or the plurality of diseased spot images to be identified one by one.

[0043] On the other hand, the present invention also provides a crop pest and disease identification device based on deep learning, the device comprising at least one processor; and

[0044] a memory communicatively connected to the at least one processor; wherein,

[0045] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the above-mentioned deep learning-based crop disease and pest identification algorithm.

[0046] On the other hand, the present invention also provides a non-volatile computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are executed by one or more processors, the one or more processors can execute the above-mentioned deep learning-based crop disease and pest identification algorithm.

[0047] The beneficial effects of the present invention are as follows: photographing multiple diseased leaf images of plants containing multiple disease categories at multiple different times; extracting leaves from the multiple diseased leaf images one by one and scaling them down to a preset size, and then outputting multiple processed diseased and insect leaf images; calculating multiple clustering feature vectors corresponding to multiple disease categories based on the multiple processed diseased and insect leaf images; calculating the membership of multiple disease spot images to be identified and the multiple clustering feature vectors; the disease category corresponding to the clustering feature vector corresponding to the largest membership among the multiple memberships is the disease category corresponding to the disease spot image to be identified; first training to obtain the clustering feature vector corresponding to the disease category, and then identifying the disease category corresponding to the disease spot image to be identified, thereby greatly improving the recognition accuracy of the disease category, and further improving the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 is a flowchart of an implementation of a crop disease and insect pest identification algorithm based on deep learning provided in Embodiment 1 of the present invention;

[0049] Figure 2 It is a structural schematic diagram of a crop disease and pest identification device based on deep learning provided in Embodiment 2 of the present invention. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0051] The specific implementation of the present invention is described in detail below in conjunction with specific embodiments:

[0052] Embodiment 1:

[0053] Figure 1 The implementation process of the crop pest and disease identification algorithm based on deep learning provided by the first embodiment of the present invention is shown. For the convenience of explanation, only the part related to the embodiment of the present invention is shown, which is described in detail as follows:

[0054] In step S101, multiple images of diseased leaves of a plant with multiple disease categories are captured at multiple different times;

[0055] In an embodiment of the present invention, the resolution and size of multiple diseased leaf images are the same, and the disease category includes multiple diseased leaf images; multiple disease categories, for example: downy mildew, brown spot and anthracnose of cucumber; the multiple diseased leaf images include at least 50 to improve the training accuracy.

[0056] In step S102, extract the leaves from the plurality of diseased leaf images one by one and scale them down to a preset size, and then output a plurality of processed diseased and pest leaf images;

[0057] In an embodiment of the present invention, a convolutional neural network algorithm is used to extract leaves from a plurality of diseased leaf images one by one.

[0058] In step S103, a plurality of clustering feature vectors corresponding to a plurality of disease categories are calculated based on a plurality of processed diseased and pest leaf images;

[0059] In an embodiment of the present invention, calculating a plurality of clustering feature vectors corresponding to a plurality of disease categories one by one according to a plurality of processed diseased and pest leaf images comprises:

[0060] Acquire environmental information when taking multiple diseased leaf images, and acquire a first environmental information feature vector based on the environmental information;

[0061] Acquire a first diseased leaf feature vector based on the processed diseased and insect-infested leaf image;

[0062] Constructing a disease joint classification feature based on the first environmental information feature vector and the first diseased leaf feature vector;

[0063] A number of features with greater contributions are extracted from the disease joint classification features and normalized to form a training feature vector.

[0064] Among them, environmental information includes: season, temperature, humidity and rainfall, and light;

[0065] The feature vector of diseased leaves includes: color features, shape features and texture features;

[0066] Color features include: mean, variance, peak, skewness, energy and entropy of color components R, G, B and H in RGB, HIS and YCbCr color spaces of processed diseased and pest leaf images, and mean of Cb and Cr;

[0067] Shape features include: circularity, eccentricity, shape complexity, and shape parameters of processed diseased and pest leaf images;

[0068] The texture features include: contrast, correlation, energy, moment of inertia and entropy of the diseased area of ​​the processed pest leaf images;

[0069] Extracting multiple features with greater contributions from the disease joint classification features includes at least 10;

[0070] The training feature vector is expressed as: X = (χ1, χ2, ..., χ10).

[0071] Furthermore, calculating a plurality of clustering feature vectors corresponding to a plurality of disease categories one by one according to a plurality of processed diseased and pest leaf images also includes:

[0072] According to formula 1, the cluster feature vector corresponding to a certain disease category is calculated;

[0073] Formula 1:

[0074] in, m=1,2,...,10.

[0075] Furthermore, before calculating the cluster feature vector of a certain disease category according to formula 1, the following steps are also included:

[0076] The feature vector of the processed leaf image of the diseased insect pest of the disease category is recorded as:

[0077] Where i = 1, 2, ..., n j , j = 1, 2, ..., C, C is the number of the plurality of said disease categories, n is the number of the plurality of said processed diseased and pest leaf images of a certain said disease category, n j is the number of the processed diseased and pest leaf images of the first disease category.

[0078] In step S104, the membership degree between a plurality of lesion images to be identified and a plurality of clustering feature vectors is calculated;

[0079] In an embodiment of the present invention, before calculating the membership degree of a plurality of lesion images to be identified and a plurality of clustering feature vectors, the method includes:

[0080] Shooting multiple disease spot images to be identified of multiple disease categories at multiple different times of the plant; wherein the resolution and size of the multiple disease spot images to be identified are the same as the resolution and size of the diseased leaf image, and the disease category includes the multiple disease spot images to be identified;

[0081] Extracting leaves from a plurality of diseased spot images to be identified one by one and reducing them proportionally to a preset size, and then outputting a plurality of processed diseased spot leaf images;

[0082] Furthermore, a convolutional neural network algorithm is used to extract leaves from a plurality of diseased spot images to be identified one by one;

[0083] Acquire environmental information when taking a plurality of images of disease spots to be identified, and acquire a second environmental information feature vector based on the environmental information;

[0084] Acquire a second diseased leaf feature vector based on the processed diseased leaf image;

[0085] Constructing a second disease joint classification feature based on the second environmental information feature vector and the second diseased leaf feature vector;

[0086] Extracting multiple features with greater contributions from the second disease joint classification features and performing normalization processing to form a test feature vector;

[0087] Extracting multiple features with greater contributions from the second disease joint classification features includes at least 10;

[0088] The test feature vector is expressed as: Y = (y1, y2, ..., y 10 ).

[0089] Furthermore, calculating the membership degree of the plurality of lesion images to be identified and the plurality of clustering feature vectors includes:

[0090] Calculate Y=(y1,y2,...,y 10 )and The degree of membership between represents a plurality of said cluster feature vectors.

[0091] Furthermore, the degree of membership is calculated using one of the Euclidean distance, cosine similarity, and Pearson correlation coefficient.

[0092] In step S105, the disease category corresponding to the cluster feature vector corresponding to the largest degree of membership among the multiple degrees of membership is the disease category corresponding to the disease spot image to be identified.

[0093] In an embodiment of the present invention, multiple diseased leaf images of a plant containing multiple disease categories are photographed at multiple different times; leaves in the multiple diseased leaf images are extracted one by one and scaled down to a preset size, and then multiple processed diseased and insect leaf images are output; multiple clustering feature vectors corresponding to multiple disease categories are calculated based on the multiple processed diseased and insect leaf images; the membership degrees of multiple disease spot images to be identified and the multiple clustering feature vectors are calculated; the disease category corresponding to the clustering feature vector corresponding to the largest membership degree among the multiple membership degrees is the disease category corresponding to the disease spot image to be identified; the clustering feature vector corresponding to the disease category is first trained to obtain the clustering feature vector corresponding to the disease category, and then the disease category corresponding to the disease spot image to be identified is identified, thereby greatly improving the recognition accuracy of the disease category and further improving the user experience.

[0094] Embodiment 2:

[0095] Figure 2 The invention shows a crop pest identification device based on deep learning provided in the second embodiment of the present invention. Figure 2 As shown, the device 10 comprises:

[0096] One or more processors 110 and memory 120, Figure 2 A processor 110 is used as an example for description. The processor 110 and the memory 120 may be connected via a bus or other means. Figure 2 The example of connecting through bus is taken in the following.

[0097] The processor 110 is used to complete various control logics of the device 10, and it can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), a single-chip microcomputer, an ARM (Acorn RISCMachine) or other programmable logic device, a discrete gate or transistor logic, a discrete hardware component, or any combination of these components. In addition, the processor 110 can also be any traditional processor, microprocessor or state machine. The processor 110 can also be implemented as a combination of computing devices, for example, a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors combined with a DSP core, or any other such configuration.

[0098] The memory 120 is a non-volatile computer-readable storage medium that can be used to store non-volatile software programs, non-volatile computer executable programs and modules, such as program instructions corresponding to the deep learning-based crop pest identification algorithm in the embodiment of the present invention. The processor 110 executes various functional applications and data processing of the device 10 by running the non-volatile software programs, instructions and units stored in the memory 120, that is, implementing the deep learning-based crop pest identification algorithm in the above method embodiment.

[0099] The memory 120 may include a program storage area and a data storage area, wherein the program storage area may store applications required for operating the device and at least one function; the data storage area may store data created according to the use of the device 10, etc. In addition, the memory 120 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 120 may optionally include a memory remotely arranged relative to the processor 110, and these remote memories may be connected to the device 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0100] One or more units are stored in the memory 120, and when executed by one or more processors 110, execute the crop pest identification algorithm based on deep learning in any of the above method embodiments, for example, execute the above described Figure 1 The method comprises steps S101 to S105.

[0101] Embodiment three:

[0102] Embodiment 3 of the present invention provides a non-volatile computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are executed by one or more processors, for example, to execute the above-described Figure 1 The method comprises steps S101 to S105.

[0103] As an example, the non-volatile storage medium can include a read-only memory (ROM), a programmable ROM (PROM), an electrically programmable ROM (EPROM), an electrically erasable ROM (EEPROM), or a flash memory. Volatile memory can include a random access memory (RAM) as an external cache memory. By way of illustration and not limitation, RAM can be obtained in many forms such as synchronous RAM (SRAM), dynamic RAM, (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), Synchlink DRAM (SLDRAM), and direct Rambus RAM (DRRAM). The disclosed memory components or memories of the operating environment described herein are intended to include one or more of these and / or any other suitable types of memory.

[0104] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a general hardware platform, and of course can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the relevant technology can be embodied in the form of a software product, and the computer software product can exist in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer electronic device (which can be a personal computer, a server, or a network electronic device, etc.) to execute the methods of each embodiment or some parts of the embodiment.

[0105] What has been described herein in this specification and the accompanying drawings includes examples that can provide crop pest and disease identification algorithms, devices, and storage media based on deep learning. Of course, it is not possible to describe every conceivable combination of elements and / or methods for the purpose of describing the various features of the present disclosure, but it can be recognized that many other combinations and permutations of the disclosed features are possible. Therefore, it is obvious that various modifications can be made to the present disclosure without departing from the scope or spirit of the present disclosure. In addition, or in an alternative, other embodiments of the present disclosure may be apparent from consideration of this specification and the accompanying drawings and from the practice of the present disclosure as presented herein. It is intended that the examples set forth in this specification and the accompanying drawings are considered to be illustrative and not restrictive in all respects. Although specific terms are employed herein, they are used in a general and descriptive sense and are not used for limiting purposes.

Claims

1. A crop pest and disease identification algorithm based on deep learning, characterized by: The algorithm comprises the following steps: Shooting multiple diseased leaf images of plants containing multiple disease categories at multiple different times; wherein the multiple diseased leaf images have the same resolution and size, and the disease category includes the multiple diseased leaf images; Extracting leaves from a plurality of the diseased leaf images one by one and reducing them proportionally to a preset size, and then outputting a plurality of the processed diseased and pest leaf images; Calculate a plurality of clustering feature vectors corresponding to a plurality of the disease categories one by one according to a plurality of the processed diseased and pest leaf images; Calculating the membership degree between a plurality of lesion images to be identified and a plurality of the clustering feature vectors; The disease category corresponding to the cluster feature vector corresponding to the largest degree of membership among the multiple degrees of membership is the disease category corresponding to the disease spot image to be identified.

2. The algorithm according to claim 1, characterized in that The step of calculating a plurality of clustering feature vectors corresponding to a plurality of disease categories according to a plurality of processed diseased and pest leaf images comprises: Acquire environmental information when taking a plurality of images of the diseased leaves, and acquire a first environmental information feature vector based on the environmental information; Acquire a first diseased leaf feature vector based on the processed diseased and insect-infested leaf image; Constructing a disease joint classification feature based on the first environmental information feature vector and the first diseased leaf feature vector; A plurality of features with greater contributions are extracted from the disease joint classification features and normalized to form a training feature vector.

3. The algorithm according to claim 2, characterized in that The environmental information includes: season, temperature, humidity, rainfall and light; The diseased leaf feature vector includes: color feature, shape feature and texture feature; The color features include: the mean, variance, peak, skewness, energy and entropy of the color components R, G, B and H in the RGB, HIS and YCbCr color spaces of the processed diseased and pest leaf images, and the mean of Cb and Cr; The shape features include: circularity, eccentricity, shape complexity and shape parameters of the processed pest leaf image; The texture features include: contrast, correlation, energy, moment of inertia and entropy of the diseased spot area of ​​the processed diseased and pest leaf image; Extracting a plurality of features with greater contributions from the disease joint classification features comprises at least: 10; The training feature vector is expressed as: X=(χ1, χ2, ..., χ10).

4. The algorithm as claimed in claim 3, characterized in that The step of calculating a plurality of clustering feature vectors corresponding to a plurality of disease categories according to a plurality of processed diseased and pest leaf images further comprises: Calculate the clustering feature vector corresponding to the disease category according to formula 1; Formula 1: in, m=1,2,...,10.

5. The algorithm according to claim 4, characterized in that Before calculating the cluster feature vector of the first disease category according to formula 1, the method further includes: The feature vector of the processed leaf image of the diseased and pest-infested leaf of the diseased category is recorded as: Where i = 1, 2, ..., n j , j = 1, 2, ..., C, C is the number of the plurality of said disease categories, n is the number of the plurality of said processed diseased and pest leaf images of a certain said disease category, n j is the number of the processed diseased and pest leaf images of the first disease category.

6. The algorithm according to claim 5, characterized in that The step of calculating the membership degree of the plurality of lesion images to be identified and the plurality of clustering feature vectors includes: Shooting a plurality of disease spot images to be identified of a plurality of disease categories at a plurality of different times of the plant; wherein the resolution and size of the plurality of disease spot images to be identified are the same as the resolution and size of the diseased leaf image, and the disease category includes the plurality of disease spot images to be identified; Extracting leaves from a plurality of the diseased spot images to be identified one by one and reducing them proportionally to a preset size, and then outputting a plurality of processed diseased spot leaf images; Acquire environmental information when taking a plurality of images of the lesions to be identified, and acquire a second environmental information feature vector based on the environmental information; Acquire a second diseased leaf feature vector based on the processed diseased leaf image; Based on the second environmental information feature vector and the second diseased leaf feature vector, a second disease joint classification feature is formed; Extracting a plurality of features with greater contributions from the second disease joint classification features and performing normalization processing to form a test feature vector; Extracting a plurality of features with greater contributions from the second disease joint classification features includes at least: 10; The test feature vector is expressed as: Y = (y1, y2, ..., y 10 ).

7. The algorithm of claim 6, wherein: The calculating of the membership degree between the plurality of lesion images to be identified and the plurality of clustering feature vectors comprises: Calculate Y=(y1,y2,...,y 10 )and The degree of membership between represents a plurality of said cluster feature vectors.

8. The algorithm of claim 7, wherein: Calculate the membership using one of Euclidean distance, cosine similarity, and Pearson correlation coefficient; A convolutional neural network algorithm is used to extract leaves from the plurality of diseased leaf images or the plurality of diseased spot images to be identified one by one.

9. A crop pest and disease identification device based on deep learning, characterized in that: The apparatus comprises at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the deep learning-based crop pest identification algorithm described in any one of claims 1-8.

10. A non-volatile computer-readable storage medium, characterized in that: The non-volatile computer-readable storage medium stores computer-executable instructions, which, when executed by one or more processors, enable the one or more processors to execute the deep learning-based crop disease and pest identification algorithm described in any one of claims 1-8.