Crop identification method and device based on leaf profile map
By performing three-dimensional grid subdivision and plane parameterization of crop leaves, extracting and screening phenotypic characteristic parameters, and constructing a two-dimensional leaf pattern map, the problem of difficulty in accurately identifying crop varieties in the existing technology is solved, and high-accuracy crop variety identification is achieved.
Patent Information
- Application Number
- CN202411632634.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-15
- Publication Date
- 2025-03-28
AI Technical Summary
The prior art is difficult to accurately identify crop varieties through leaf type differences.
By performing grid subdivision and plane parameterization of the leaf three-dimensional grid models of each variety of the target crop, a two-dimensional grid model of the leaf of each variety was obtained, multiple phenotypic characteristic parameters were extracted, key phenotypic characteristic parameters were screened out, and a two-dimensional leaf pattern map was constructed for variety identification of the crops to be identified.
Accurate identification of crop varieties based on leaf type is achieved, and the accuracy of crop varieties is improved.
Smart Images
Figure CN119722834A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of crop identification, and in particular to a crop identification method and device based on leaf profile atlas. Background Art
[0002] Crop plant type refers to the appearance of the plant as a whole or in its parts, which is composed of leaf type, stem type, ear type and root system configuration. Leaf type is an important component of crop plant type, including morphological and structural characteristics such as leaf spatial posture, local wrinkles, and leaf edge undulations. Leaf type varies among different species, different varieties of the same plant, and even among different individuals of the same variety. These differences are actually closely related to different plant genes. Crop leaf type can reflect the growth status of the plant to a certain extent. Differences in leaf type in three-dimensional space will directly affect the crop canopy structure, light energy utilization efficiency, etc., which is of great significance to agricultural production research.
[0003] Therefore, how to accurately identify crop varieties based on leaf shape has become a technical problem that needs to be solved urgently in the industry. Summary of the invention
[0004] The present application provides a crop identification method and device based on leaf shape atlas, which are used to solve the technical problem of how to accurately identify the variety of crops according to leaf shape.
[0005] The present application provides a crop identification method based on leaf type atlas, comprising: The three-dimensional mesh model of leaves of each variety of target crops is meshed and plane parameterized to obtain the two-dimensional mesh model of leaves of each variety; Determine a plurality of phenotypic characteristic parameters of the two-dimensional grid model of leaves of each variety and an average two-dimensional grid model of leaves of each variety; Screening the multiple phenotypic characteristic parameters to determine multiple key phenotypic characteristic parameters, and determining a two-dimensional leaf profile of the target crop based on the multiple key phenotypic characteristic parameters and an average two-dimensional grid model of leaves of each variety; Based on the two-dimensional leaf type map of the target crop, the variety of the crop to be identified is identified.
[0006] In some embodiments, the three-dimensional mesh model of leaves of each variety of target crops is meshed and plane parameterized to obtain the two-dimensional mesh model of leaves of each variety, including: Perform multiple mesh subdivisions on the three-dimensional mesh model of leaves of each variety of target crops to obtain the subdivided three-dimensional mesh model of leaves of each variety; Based on the mesh deformation algorithm that keeps the rigidity unchanged, the subdivided three-dimensional mesh model of each type of blade is plane parameterized to obtain the two-dimensional mesh model of each type of blade.
[0007] In some embodiments, the step of determining a plurality of phenotypic characteristic parameters of the two-dimensional grid model of leaves of each variety and an average two-dimensional grid model of leaves of each variety comprises: Extract the leaf contours of the two-dimensional mesh model of leaves of each variety to obtain edge contour line sets; A ray is emitted around the center point of the two-dimensional grid model of leaves of each variety to intersect with the edge contour line set to obtain a sampling point set of each variety; Based on the sampling point set of each variety, the average two-dimensional grid model of the leaves of each variety is determined.
[0008] In some embodiments, the screening of the multiple phenotypic characteristic parameters to determine multiple key phenotypic characteristic parameters includes: Correlation analysis and cluster analysis are performed on the multiple phenotypic characteristic parameters to determine multiple key phenotypic characteristic parameters.
[0009] In some embodiments, determining the two-dimensional leaf profile of the target crop based on the multiple key phenotypic characteristic parameters and the average two-dimensional grid model of leaves of each variety includes: Determine the impact weight corresponding to each key phenotypic characteristic parameter; Based on the key phenotypic characteristic parameters of each variety and the corresponding influence weights of each key phenotypic characteristic parameter, the leaf type characteristic index of each variety is determined; Based on the leaf shape characteristic index of each variety, the average two-dimensional grid model of the leaves of each variety is sorted to obtain a two-dimensional leaf shape map of the target crop.
[0010] In some embodiments, the performing variety identification on the crop to be identified based on the two-dimensional leaf type map of the target crop includes: Determining key phenotypic characteristic parameters of the crop to be identified based on a two-dimensional grid model of leaves of the crop to be identified; Comparing the key phenotypic characteristic parameters of the crop to be identified with the key phenotypic characteristic parameters of each variety in the two-dimensional leaf type map of the target crop for similarity; The variety with the highest similarity is determined as the variety of the crop to be identified.
[0011] In some embodiments, the performing variety identification on the crop to be identified based on the two-dimensional leaf type map of the target crop includes: Obtain key phenotypic characteristic parameters of the crops to be identified; Comparing the key phenotypic characteristic parameters of the crop to be identified with the key phenotypic characteristic parameters of each variety in the two-dimensional leaf type map of the target crop for similarity; The average two-dimensional grid model of leaves of the variety with the highest similarity is determined as the average two-dimensional grid model of leaves of the crop to be identified.
[0012] The present application provides a crop identification device based on leaf profile atlas, comprising: A model flattening unit is used to perform mesh subdivision and plane parameterization on the three-dimensional mesh model of leaves of various varieties of target crops to obtain a two-dimensional mesh model of leaves of various varieties; A parameter extraction unit, used to determine a plurality of phenotypic characteristic parameters of the two-dimensional grid model of leaves of each variety and an average two-dimensional grid model of leaves of each variety; A map construction unit is used to screen the multiple phenotypic characteristic parameters, determine multiple key phenotypic characteristic parameters, and determine the two-dimensional leaf map of the target crop based on the multiple key phenotypic characteristic parameters and the average two-dimensional grid model of leaves of each variety; The variety identification unit is used to perform variety identification on the crop to be identified based on the two-dimensional leaf type map of the target crop.
[0013] The present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the crop identification method based on leaf profiles when executing the computer program.
[0014] The present application provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the crop identification method based on leaf profile is implemented.
[0015] The crop identification method and device based on leaf type map provided in the present application perform grid subdivision and plane parameterization on the three-dimensional grid model of leaves of each variety of target crops to obtain the two-dimensional grid model of leaves of each variety; determine multiple phenotypic characteristic parameters of the two-dimensional grid model of leaves of each variety and the average two-dimensional grid model of leaves of each variety; screen multiple phenotypic characteristic parameters to determine multiple key phenotypic characteristic parameters, and determine the two-dimensional leaf type map of the target crop based on the multiple key phenotypic characteristic parameters and the average two-dimensional grid model of leaves of each variety; perform variety identification on the crop to be identified based on the two-dimensional leaf type map of the target crop; since the phenotypic characteristic parameters are used to quantitatively characterize the two-dimensional grid model of leaves, a two-dimensional leaf type map of crops is constructed through key phenotypic characteristic parameters and the average two-dimensional grid model of leaves, thereby realizing phenotypic big data analysis and knowledge mining for large-scale crop groups, realizing accurate identification of crop varieties based on leaf type, and improving the accuracy of crop variety identification. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.
[0017] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0018] Figure 1 This is one of the flow charts of the crop identification method based on leaf type atlas provided in this application.
[0019] Figure 2 It is a flattened schematic diagram of the three-dimensional mesh model of the blade provided in this application.
[0020] Figure 3 It is a schematic diagram of constructing the blade average two-dimensional grid model provided in this application.
[0021] Figure 4 It is a schematic diagram of the two-dimensional leaf shape map of corn leaves provided in this application.
[0022] Figure 5 This is the second flow chart of the crop identification method based on leaf type atlas provided in this application.
[0023] Figure 6 It is a structural schematic diagram of a crop identification device based on leaf profile atlas provided in this application.
[0024] Figure 7 It is a structural schematic diagram of the electronic device provided by this application. DETAILED DESCRIPTION
[0025] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.
[0026] It should be noted that the terms "first", "second" etc. in the present application are used to distinguish similar objects, and need not be used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable in appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, the process, method, system, product or equipment comprising a series of steps or units or modules need not be limited to those steps or units or modules clearly listed, but may include other steps or units or modules that are not clearly listed or inherent to these processes, methods, products or equipment.
[0027] The data collection of crop leaf shape is mostly through simple one-dimensional measurement, such as leaf length, leaf width, aspect ratio, etc. With the continuous deepening of leaf shape research, the methods of obtaining crop leaf shape are gradually enriched.
[0028] Taking corn leaves as an example, researchers placed corn leaves on a horizontal plane and flattened them as much as possible. They used devices such as ordinary cameras or hyperspectral cameras to obtain images of the leaves from a position perpendicular to the horizontal plane and extracted the two-dimensional phenotypic parameters related to the leaves. By extracting parameters from the image, two-dimensional parameters such as leaf area and leaf contour perimeter of corn leaves can be obtained. However, the flattening process will cause the loss of three-dimensional structural information of the leaves, such as the bending and wrinkling degree of the leaves, which is not conducive to a deeper study of the leaf shape from the perspective of three-dimensional structure. The emergence of three-dimensional data acquisition technologies such as laser radar (LiDAR), three-dimensional digitizers, and multi-view three-dimensional reconstruction has made it possible to fully obtain the three-dimensional structure of the leaf shape. In addition to obtaining the one-dimensional and two-dimensional phenotypic parameters of the previous methods, leaf shape research based on three-dimensional (3D) data can also obtain three-dimensional phenotypic parameters reflected by three-dimensional structures such as the bending degree and leaf inclination angle of the leaves. However, point cloud data obtained by laser radar (LiDAR) and other methods require complex processing to generate a 3D model containing semantic information and extract 3D leaf shape parameters. In contrast, 3D digitizers can directly obtain data containing semantic information through manual instrument collection. The method of using digitizers to obtain 3D structural data of plants has been developed for a long time and has been widely used in the study of plant roots, leaves and other organs or the whole plant.
[0029] The goal of variety identification based on leaf shape is to determine the variety to which the leaf belongs based on the differences in leaf shape between different varieties. The specific method can be achieved by screening representative key phenotypic parameters of leaf shape and calculating the differences in key leaf shape parameters to achieve variety identification. Leaf shape generation based on leaf shape characteristics also needs to rely on key phenotypic parameters to express leaf shape characteristics, thereby achieving the generation and construction of the target leaf shape.
[0030] The relevant technology lacks a quantitative characterization method for the two-dimensional (2D) leaf shape characteristics of crop leaves, the interaction between leaf shape characteristics is still unclear, and there is a lack of means for phenotypic big data analysis and knowledge mining for large-scale (such as inbred lines) populations. It is not yet possible to achieve variety identification based on leaf shape characteristics and leaf shape generation based on morphological and structural characteristics.
[0031] Figure 1 This is one of the flow charts of the crop identification method based on leaf type atlas provided in this application, such as Figure 1 As shown, the method includes step 110 , step 120 , step 130 and step 140 .
[0032] Step 110 , mesh subdivision and plane parameterization are performed on the three-dimensional mesh model of leaves of each variety of the target crop to obtain a two-dimensional mesh model of leaves of each variety.
[0033] Specifically, the crop identification method based on leaf type atlas provided in the embodiment of the present application is implemented by a crop identification device. The device can be implemented by software, such as a crop identification program running on a computer; or by hardware, such as a computer or server that executes the crop identification method based on leaf type atlas.
[0034] The target crop can be any plant cultivated in agriculture, including food crops and cash crops. For example, the target crop can be corn. Crop varieties refer to plant groups with specific genetic and agronomic traits selected through artificial breeding or natural variation. For example, corn varieties can include sweet corn, glutinous corn, and ordinary corn.
[0035] The blade 3D mesh model refers to a model that represents the geometric shape of the blade after the blade is reconstructed in 3D through digital technology. It is usually composed of a series of vertices, edges and faces, and can accurately describe the surface and structure of the blade in 3D space. This mesh model can help researchers better analyze the morphology and structure of the blade and its relationship with other environmental factors.
[0036] The blade 2D mesh model refers to a model that represents the geometric shape of the blade surface in a planar (two-dimensional) manner, and decomposes the two-dimensional shape of the blade into a series of regular or irregular mesh units (usually rectangles or triangles) by meshing. This model mainly focuses on the planar structure of the blade, that is, it only considers the geometric distribution of the blade on a plane (usually the main plane of the blade), and does not involve the thickness or three-dimensional spatial characteristics of the blade.
[0037] Mesh subdivision refers to the technique of increasing the fineness of a mesh by adding vertices and refining the mesh structure, especially when dealing with complex shapes or details.
[0038] Plane parameterization is the process of mapping a three-dimensional surface to a two-dimensional plane.
[0039] The three-dimensional mesh models of leaves of various varieties of target crops can be collected, and these three-dimensional mesh models of leaves are respectively meshed and plane parameterized to obtain the two-dimensional mesh models of leaves of various varieties.
[0040] Step 120: Determine a plurality of phenotypic characteristic parameters of the two-dimensional grid model of leaves of each variety and an average two-dimensional grid model of leaves of each variety.
[0041] Specifically, the phenotypic characteristic parameters refer to parameters describing the phenotypic characteristics of leaves, including leaf length, leaf width, leaf tip angle, leaf area, leaf edge tortuosity, leaf wrinkle degree, etc.
[0042] The average two-dimensional grid model of leaves refers to the two-dimensional grid model obtained by averaging the two-dimensional grid models of leaves of different individuals belonging to the same variety, which can be used to represent the common characteristics of leaves of a certain variety.
[0043] Parameters of the leaf two-dimensional grid model of each variety can be extracted to obtain multiple phenotypic characteristic parameters of the leaf two-dimensional grid model. The leaf two-dimensional grid models of multiple different individuals of each variety are averaged to obtain the average leaf two-dimensional grid model of each variety.
[0044] Step 130: Screen multiple phenotypic characteristic parameters to determine multiple key phenotypic characteristic parameters, and determine a two-dimensional leaf profile of the target crop based on the multiple key phenotypic characteristic parameters and the average two-dimensional grid model of leaves of each variety.
[0045] Specifically, multiple phenotypic characteristic parameters can be screened according to the degree of influence of each phenotypic characteristic parameter on the identification of the target crop variety, thereby determining multiple key phenotypic characteristic parameters. For example, phenotypic characteristic parameters with an influence greater than a preset threshold can be selected as key phenotypic characteristic parameters.
[0046] Based on multiple key phenotypic characteristic parameters, a comprehensive index that can represent leaf characteristics is calculated. The average two-dimensional grid models of leaves of various varieties are sorted according to the comprehensive index to obtain a two-dimensional leaf type map of the target crop.
[0047] The two-dimensional leaf profile map includes the average two-dimensional grid model and key phenotypic characteristic parameters of the leaves of each variety of the target crop, which can accurately describe the leaf characteristics of each variety of the target crop.
[0048] Step 140: Based on the two-dimensional leaf shape map of the target crop, perform variety identification on the crop to be identified.
[0049] Specifically, the crop to be identified refers to the crop that needs to be identified. On the one hand, the leaf two-dimensional grid model of the crop to be identified can be obtained, and the similarity of the leaf two-dimensional grid model is compared with the average two-dimensional grid models of each leaf in the two-dimensional leaf type map of the target crop, and the variety corresponding to the average two-dimensional grid model of the leaf with the highest similarity is determined as the variety of the crop to be identified; on the other hand, the key phenotypic characteristic parameters of the crop to be identified can be obtained, and the similarity of the key phenotypic characteristic parameters is compared with the key phenotypic characteristic parameters of each average two-dimensional grid model of the leaf in the two-dimensional leaf type map of the target crop, and the average two-dimensional grid model of the leaf with the highest similarity is determined as the two-dimensional grid model of the leaf of the crop to be identified.
[0050] The crop identification method based on leaf type map provided in the embodiment of the present application performs grid subdivision and plane parameterization on the three-dimensional grid model of the leaves of each variety of the target crop to obtain the two-dimensional grid model of the leaves of each variety; determines multiple phenotypic characteristic parameters of the two-dimensional grid model of the leaves of each variety and the average two-dimensional grid model of the leaves of each variety; screens multiple phenotypic characteristic parameters to determine multiple key phenotypic characteristic parameters, and determines the two-dimensional leaf type map of the target crop based on the multiple key phenotypic characteristic parameters and the average two-dimensional grid model of the leaves of each variety; performs variety identification on the crop to be identified based on the two-dimensional leaf type map of the target crop; since the phenotypic characteristic parameters are used to quantitatively characterize the two-dimensional grid model of the leaves, a two-dimensional leaf type map of the crop is constructed through the key phenotypic characteristic parameters and the average two-dimensional grid model of the leaves, thereby realizing phenotypic big data analysis and knowledge mining for large-scale crop groups, realizing accurate identification of crop varieties based on leaf type, and improving the accuracy of crop variety identification.
[0051] It should be noted that each implementation method of the present application can be freely combined, the order can be changed, or it can be executed separately, and does not need to rely on or depend on a fixed execution order.
[0052] In some embodiments, the three-dimensional mesh model of leaves of each variety of target crops is meshed and plane parameterized to obtain the two-dimensional mesh model of leaves of each variety, including: Perform multiple mesh subdivisions on the three-dimensional mesh model of leaves of each variety of target crops to obtain the subdivided three-dimensional mesh model of leaves of each variety; Based on the mesh deformation algorithm that keeps the rigidity unchanged, the subdivided three-dimensional mesh model of each type of blade is plane parameterized to obtain the two-dimensional mesh model of each type of blade.
[0053] Specifically, plane parameterization can also be called flattening. Figure 2 is a flattened schematic diagram of the blade three-dimensional mesh model provided in this application, such as Figure 2As shown in the figure, the flattening work is divided into two steps. The first step is to subdivide the original collected blade 3D mesh model (shown in part a of the figure) to obtain a blade subdivided 3D mesh model (shown in part b of the figure). The second step is to flatten the blade subdivided 3D mesh model to obtain a blade 2D mesh model (shown in part c of the figure).
[0054] Step 1. The original leaf 3D mesh model can be meshed using a uniform Sqrt3 subdivision algorithm. After the original leaf 3D mesh model is subdivided, the resulting leaf model has a shape that is closer to the real leaf than the original data, and is more conducive to reflecting the true phenotypic characteristics of crop leaves. The number of mesh subdivisions can be determined as needed. In this application, the original mesh is subjected to multiple Sqrt3 subdivision iterations, and after segmentation, the number of triangular meshes becomes nine times that of the original data, obtaining a leaf subdivision 3D mesh model, which can achieve a more uniform and smooth subdivision effect.
[0055] Step 2: Flatten the subdivided three-dimensional mesh model of the leaf in two dimensions. This process is called plane parameterization of the mesh. In this application, a mesh deformation algorithm that maintains rigidity is used, such as the (As-Rigid-As-Possible, ARAP) algorithm. After plane parameterization, the three-dimensional mesh model forms a corresponding plane two-dimensional mesh model. It can be observed that the two-dimensional leaf model after plane parameterization will produce certain deformations due to the wrinkles and distortions of the original three-dimensional model, so that the flattened two-dimensional leaf is not completely axially symmetrical along the midline of the leaf vein. On the one hand, the two-dimensional leaf model after plane parameterization retains the basic phenotypic characteristics of crop leaves, such as leaf length and leaf width, to the greatest extent; on the other hand, the deformations caused by these plane parameterizations can also reflect different three-dimensional crop leaf phenotypic characteristics to a certain extent, such as leaf distortion. The two-dimensional crop leaf mesh model obtained by plane parameterization of the ARAP algorithm can be well used to extract relevant phenotypic parameters for the analysis and study of leaf morphology.
[0056] The crop identification method based on leaf type maps provided in the embodiment of the present application performs grid subdivision and plane parameterization on the three-dimensional grid model of leaves of each variety of the target crop to obtain a two-dimensional grid model of the leaves of each variety, which is conducive to extracting relevant phenotypic parameters for leaf morphology analysis, improves the accuracy of the two-dimensional leaf type map, and improves the accuracy of crop variety identification.
[0057] In some embodiments, determining a plurality of phenotypic characteristic parameters of the two-dimensional grid model of leaves of each variety and an average two-dimensional grid model of leaves of each variety comprises: Extract the leaf contours of the two-dimensional mesh model of leaves of each variety to obtain edge contour line sets; Rays are emitted around the center point of the two-dimensional grid model of leaves of each variety and intersect with the edge contour line set to obtain the sampling point set of each variety; Based on the sampling point set of each variety, the average two-dimensional grid model of the leaves of each variety is determined.
[0058] Specifically, the average two-dimensional grid model of leaves of each variety can be determined before extracting the phenotypic characteristic parameters. Figure 3 is a schematic diagram of constructing the blade average two-dimensional grid model provided in this application, such as Figure 3 As shown, constructing the blade average two-dimensional grid model includes three steps.
[0059] Step 1: Extract the blade contour. The blade 2D mesh model (shown in part a of the figure) is a non-closed mesh, and the contour edge of the blade mesh is the boundary edge. By screening out all the boundary edges in the blade 2D mesh model, the edge contour line set of the blade mesh is obtained (shown in part b of the figure).
[0060] Step 2: Perform contour boundary sampling. Starting from the center point of the leaf, a ray is emitted around the center point and intersects with the edge contour line set to obtain a sampling point set (shown in part c in the figure). For example, for the corn leaf in the embodiment of the present application, the sampling accuracy is set to 600, so that a good sampling effect can still be obtained at the tip of the leaf and the base of the leaf far from the center point (shown in part d in the figure).
[0061] Step 3: Complete the construction of the average two-dimensional grid model of leaves. Calculate the average of the coordinates of the sampling point set belonging to the same variety, that is, the average point set. Connect the points in the average point set in sequence, and finally obtain the average two-dimensional grid model of leaves of each variety.
[0062] The crop identification method based on leaf type maps provided in the embodiments of the present application determines the average two-dimensional grid model of leaves of each variety, which is conducive to leaf morphology analysis and improves the accuracy of the two-dimensional leaf type maps.
[0063] In some embodiments, multiple phenotypic characteristic parameters are screened to determine multiple key phenotypic characteristic parameters, including: Correlation analysis and cluster analysis were performed on multiple phenotypic characteristic parameters to determine multiple key phenotypic characteristic parameters.
[0064] Specifically, the embodiment of the present application can extract 29 relevant phenotypic characteristic parameters for the two-dimensional grid model of the blade, which can be divided into two categories: the first category is conventional phenotypic characteristic parameters, and the second category is blade profile characteristic parameters.
[0065] Conventional phenotypic characteristic parameters include leaf length, leaf width, leaf tip angle, leaf area, etc. At the same time, parameters such as the length of the left and right edges of the leaf, the degree of tortuosity of the left and right edges, and the width of the leaf at different positions were also extracted based on the two-dimensional grid model of the leaf, as shown in Table 1.
[0066] Table 1 Conventional phenotypic characteristic parameters
[0067] The calculation method of the one-dimensional parameters in the above table is as follows (length, width and related scale parameters): The calculation is based on the distance and angle between points in the 2D grid model. These 1D features include median length (ML), maximum width (WM), mean width (WA), variance, and widest position (WP).
[0068] The calculation method of the two-dimensional parameters in the above table (leaf area, leaf tip angle, offset and twist) is as follows: The leaf area is calculated by summing the areas of all triangular mesh patches. Compared with the traditional method based on the formula (for example, for corn leaves: leaf length × maximum leaf width × 0.75), the method of calculating the area of mesh patches separately can obtain a more accurate leaf area.
[0069] The tip angle, the edge of the leaf, and the offset and curvature of the midrib are all expressed in the form of angles. In this application, the radian system is used to measure the angle-related parameters. The tip angle is represented by the angle value of the left and right edges at the tip of the leaf; the offset value of the left and right edges and the midline of the leaf is calculated by summing the angle values obtained by connecting the grid vertices at the corresponding positions point by point, and this value emphasizes the overall curvature of the leaf edge contour; the midline curvature is calculated by calculating the mean and variance of the angle between the leaf midline and the line connecting the leaf base-tip point, and this value emphasizes the left and right bending of the leaf after the two-dimensional plane parameterization.
[0070] The calculation method of the wrinkle degree in the above table is as follows: Wrinkles are an important morphological feature of leaves. This application proposes a quantitative calculation method for wrinkles based on three-dimensional digital data. For the parameters extracted from each sample, the average is calculated within the variety according to different varieties to obtain the average value of the phenotypic parameters representing the level of each group. The FA (leaf wrinkle degree) parameter mainly comprehensively reflects the degree of leaf wrinkles. The calculation of the FA parameter is as follows: In the formula, is the total number of all grids of crop leaves; For the grid The number of adjacent grids; For the The normal vector of the mesh, For the The adjacent The normal vector of the mesh; and is a vector; is the inverse cosine function.
[0071] The blade profile characteristic parameters are parameters that reflect the blade profile. The blade profile is divided into six parts: the blade tip, the upper middle part (left / right), the lower middle part (left / right) and the blade base. The distance between each part and the center point is calculated to obtain the quantitative value of the blade profile reflecting each part. The overall length-to-width ratio (LW) and the upper and lower width difference (UL) of the blade profile are obtained based on the quantitative values of different parts. The blade profile characteristic parameters are shown in Table 2.
[0072] Table 2 Blade profile characteristic parameters
[0073] Correlation analysis and cluster analysis can be performed on the above-mentioned multiple phenotypic characteristic parameters to determine multiple key phenotypic characteristic parameters.
[0074] For example, a hierarchical clustering method can be used. First, each feature is treated as a group separately, and then the closeness between the features is measured. Features with closer relationships are clustered into one category first, and this process is repeated until all samples are clustered into one category. In this application, the clustering threshold is set to 0.25, and the closeness between clusters is measured using correlation.
[0075] Correlation analysis reflects the degree of linear correlation between two related variables. If there is a large correlation between two phenotypic characteristic parameters, it means that the two parameters have a large degree of overlap and similarity in reflecting the leaf phenotype. Therefore, the more representative parameter of the two parameters is selected for feature description to achieve simplification and refinement of feature information.
[0076] For example, for corn leaves, the above method can be used to determine 11 key phenotypic characteristic parameters (WP, LWR, L, LTA, WV, W, FA, MTV, MTA, LOA, ROA) from 29 phenotypic characteristic parameters. That is, these 11 key phenotypic characteristic parameters can be used to describe most of the characteristics of the two-dimensional leaf shape. Since these 11 key phenotypic characteristic parameters have clear meanings, they can also be called semantic characteristic parameters.
[0077] The crop identification method based on leaf type maps provided in the embodiment of the present application performs correlation analysis and cluster analysis on multiple phenotypic characteristic parameters, determines multiple key phenotypic characteristic parameters, achieves a high degree of condensation and compression of characteristics, reduces the amount of data, and improves the efficiency of crop identification.
[0078] In some embodiments, determining a two-dimensional leaf profile of a target crop based on a plurality of key phenotypic characteristic parameters and an average two-dimensional grid model of leaves of each variety includes: Determine the impact weight corresponding to each key phenotypic characteristic parameter; Based on the key phenotypic characteristic parameters of each variety and the corresponding influence weights of each key phenotypic characteristic parameter, the leaf type characteristic index of each variety is determined; Based on the leaf shape characteristic indexes of each variety, the average two-dimensional grid models of leaves of each variety are sorted to obtain the two-dimensional leaf shape map of the target crop.
[0079] Specifically, the influence weight corresponding to each key phenotypic characteristic parameter can be determined according to the degree of influence of each key phenotypic characteristic parameter on the crop variety identification result. For example, in order to select the appropriate influence weight, different weights are sampled for each key phenotypic characteristic parameter, and the weights of other key phenotypic characteristic parameters are set to 1. The optimal weight value is obtained through data verification, and finally all the influence weights are normalized.
[0080] The key phenotypic characteristic parameters of each variety can be expressed using the characteristic vector Expressed as , For the Key phenotypic characteristic parameters; the influence weight can be expressed as a weight vector Expressed as , For the The influence weights corresponding to the key phenotypic characteristic parameters.
[0081] Taking corn leaves as an example, the eigenvector and influence vector can be expressed as: ; .
[0082] After weighted summation of the corresponding items in the eigenvector and the influence vector, the weighted vector can be obtained , the magnitude of this vector Determined as the leaf type characteristic index of each variety. This index can be used as a comprehensive index. It can reflect the degree of expression of leaf characteristics. When the phenotypic characteristics of the leaves are more obvious (such as longer length, wider width, more tortuous edges, etc.), the expression parameters When the phenotypic characteristics of the leaves are less obvious, the expression parameters The smaller the value of .
[0083] .
[0084] .
[0085] In the formula, is the weight vector Middle parameters, is the feature vector Middle parameters.
[0086] According to the leaf shape characteristic indexes of each variety, the average two-dimensional grid models of leaves of each variety are sorted to obtain the two-dimensional leaf shape map of the target crop.
[0087] Figure 4 is a schematic diagram of a two-dimensional leaf profile of a corn leaf provided in this application, such as Figure 4 As shown in FIG. 1 , the two-dimensional leaf shape atlas includes the average two-dimensional grid models of leaves of multiple varieties of corn. These two-dimensional grid models gradually increase from the inside to the outside according to the leaf shape characteristic index.
[0088] The crop identification method based on leaf type maps provided in the embodiments of the present application constructs a two-dimensional leaf type map of crops, realizes phenotypic big data analysis and knowledge mining for large-scale crop populations, realizes accurate identification of crop varieties based on leaf type, and improves the accuracy of crop variety identification.
[0089] In some embodiments, performing variety identification on the crop to be identified based on the two-dimensional leaf type map of the target crop includes: Based on the two-dimensional grid model of the leaves of the crop to be identified, key phenotypic characteristic parameters of the crop to be identified are determined; Comparing the key phenotypic characteristic parameters of the crop to be identified with the key phenotypic characteristic parameters of each variety in the two-dimensional leaf type map of the target crop for similarity; The variety with the highest similarity is determined as the variety of the crop to be identified.
[0090] Specifically, for crops to be identified, key phenotypic characteristic parameters can be extracted based on the two-dimensional grid model of leaves.
[0091] The key phenotypic characteristic parameters of the crop to be identified are compared with the key phenotypic characteristic parameters of each variety in the two-dimensional leaf type map of the target crop. The similarity comparison method can be cosine similarity comparison. Cosine similarity measures the difference between vectors by calculating the cosine value of the angle between two different vectors in multidimensional space.
[0092] Finally, the variety with the highest similarity is determined as the variety of the crop to be identified.
[0093] The crop identification method based on leaf type atlas provided in the embodiment of the present application determines the variety of the crop to be identified based on the comparison of key phenotypic characteristic parameters, thereby improving the accuracy of crop variety identification.
[0094] In some embodiments, performing variety identification on the crop to be identified based on the two-dimensional leaf type map of the target crop includes: Obtain key phenotypic characteristic parameters of the crops to be identified; Comparing the key phenotypic characteristic parameters of the crop to be identified with the key phenotypic characteristic parameters of each variety in the two-dimensional leaf type map of the target crop for similarity; The average two-dimensional grid model of the leaves of the variety with the highest similarity is determined as the average two-dimensional grid model of the leaves of the crop to be identified.
[0095] Specifically, when the key phenotypic characteristic parameters of the crop to be identified are known, the key phenotypic characteristic parameters of the crop to be identified can be compared with the key phenotypic characteristic parameters of each variety in the two-dimensional leaf type map of the target crop for similarity, and the similarity comparison method can adopt cosine similarity comparison, etc. Finally, the average two-dimensional grid model of the leaves of the variety with the highest similarity is determined as the average two-dimensional grid model of the leaves of the crop to be identified, so that it can be used to analyze and study the leaf type of the crop to be identified.
[0096] The crop identification method based on leaf type atlas provided in the embodiment of the present application determines the average two-dimensional grid model of the leaves of the crop to be identified based on the comparison of key phenotypic characteristic parameters, thereby realizing the unique drawing (visualization) of the leaves of the crop to be identified based on the phenotypic characteristic parameters.
[0097] Figure 5 This is the second flow chart of the crop identification method based on leaf type atlas provided in this application, such as Figure 5 As shown, the method is applied to corn leaves and comprises: Step 510: flattening the crop leaf 3D mesh model; Step 520, quantitative extraction of 2D leaf features of crop leaves; Step 530, 2D leaf shape semantic feature screening and leaf shape map construction; Step 540: Variety identification based on leaf shape data and leaf shape matching generation based on semantic features.
[0098] The crop identification method based on leaf type atlas provided in the embodiment of the present application has the following technical effects: (1) The ARAP plane parameterization method is used to convert the three-dimensional mesh model into a two-dimensional plane mesh. Compared with traditional methods such as obtaining the leaf plane mesh by projection or manual flattening, the ARAP plane parameterization technology can more accurately retain the structural information of the three-dimensional leaf and display it through the tortuosity of the mesh edge. Unlike the projection method, it will not be affected by factors such as the projection angle and cause information loss. For example, the leaf area of the 3D leaf mesh model can be highly accurately restored.
[0099] (2) A method for accurately quantifying and extracting leaf phenotypic characteristics was developed for the 2D leaf mesh model, and the phenotypic parameters of the leaves were fully extracted. In related studies, many indicators for expressing leaf phenotypic parameters have been proposed. However, due to different research directions and progress, there are duplications and redundancies between many phenotypic indicators. In the face of this situation, how to screen features - while ensuring the simplicity of feature description and covering as much as possible, plays a vital role in the smooth progress of subsequent work. In this application, based on the collection of phenotypic indicators, a hierarchical clustering method was used to screen and classify the features, and finally the parameters of 2D corn leaf shape features that can be relatively comprehensive and non-repetitive were determined, achieving a high degree of condensation and compression of 2D leaf shape features, laying the foundation for the construction of feature vectors and the identification of leaf shape varieties.
[0100] (3) Since there are many phenotypic indicators used to describe leaf shape, it is necessary to comprehensively rank all leaf types to fully reflect the differences and changes between leaf types. Currently, there is a lack of methods for quantitative ranking and integration of leaf types. This application obtains a comprehensive 2D leaf type index by weighted calculation of the modulus of the selected semantic feature parameters. The 2D leaf types are sorted and organized according to the comprehensive 2D leaf type index to obtain a 2D leaf type atlas, which realizes the ranking based on the degree of feature expression of the main leaf type features. The atlas shows the expression change relationship from weak to strong of the 2D leaf type phenotypic indicators of a large number of varieties.
[0101] (4) Compared with previous classification work that was mostly carried out between different species and plants with large differences, the work of identifying the variety to which a given leaf type belongs through a given leaf type enables phenotypic analysis to reach variety-level resolution, which has an important role in promoting the development of phenotypic analysis. Leaf type drawing based on semantic features is highly interpretable. This application is of great significance for promoting 2D generation and even 3D generation based on semantic information. This application has formed a research model for the analysis of crop phenotypic big data and portrait work based on phenotypic big data: including feature extraction, semantic feature extraction and screening, comprehensive indicator construction and map construction, and map-based variety identification. It provides new methods and approaches for knowledge mining of phenotypic big data, and proposes a method and full workflow with reference significance for plant variety identification.
[0102] The device provided in the embodiment of the present application is described below. The device described below and the method described above can be referenced to each other.
[0103] Figure 6 is a schematic diagram of the structure of a crop identification device based on leaf type atlas provided in this application, such as Figure 6 As shown, the device comprises: The model flattening unit 610 is used to perform mesh subdivision and plane parameterization on the three-dimensional mesh model of leaves of each variety of the target crop to obtain a two-dimensional mesh model of leaves of each variety; A parameter extraction unit 620, used to determine a plurality of phenotypic characteristic parameters of the two-dimensional grid model of leaves of each variety and an average two-dimensional grid model of leaves of each variety; A map construction unit 630 is used to screen multiple phenotypic characteristic parameters, determine multiple key phenotypic characteristic parameters, and determine a two-dimensional leaf map of the target crop based on the multiple key phenotypic characteristic parameters and the average two-dimensional grid model of leaves of each variety; The variety identification unit 640 is used to perform variety identification on the crop to be identified based on the two-dimensional leaf type map of the target crop.
[0104] The crop identification device based on leaf type atlas provided in the embodiment of the present application performs grid subdivision and plane parameterization on the three-dimensional grid model of leaves of each variety of target crops to obtain the two-dimensional grid model of leaves of each variety; determines multiple phenotypic characteristic parameters of the two-dimensional grid model of leaves of each variety and the average two-dimensional grid model of leaves of each variety; screens multiple phenotypic characteristic parameters to determine multiple key phenotypic characteristic parameters, and determines the two-dimensional leaf type atlas of the target crop based on the multiple key phenotypic characteristic parameters and the average two-dimensional grid model of leaves of each variety; performs variety identification on the crop to be identified based on the two-dimensional leaf type atlas of the target crop; since the phenotypic characteristic parameters are used to quantitatively characterize the two-dimensional grid model of leaves, a two-dimensional leaf type atlas of crops is constructed through key phenotypic characteristic parameters and the average two-dimensional grid model of leaves, thereby realizing phenotypic big data analysis and knowledge mining for large-scale crop groups, realizing accurate identification of crop varieties based on leaf type, and improving the accuracy of crop variety identification.
[0105] In some embodiments, the model flattening unit is used to: Perform multiple mesh subdivisions on the three-dimensional mesh model of leaves of each variety of target crops to obtain the subdivided three-dimensional mesh model of leaves of each variety; Based on the mesh deformation algorithm that keeps the rigidity unchanged, the subdivided three-dimensional mesh model of each type of blade is plane parameterized to obtain the two-dimensional mesh model of each type of blade.
[0106] In some embodiments, the parameter extraction unit is used to: Extract the leaf contours of the two-dimensional mesh model of leaves of each variety to obtain edge contour line sets; Rays are emitted around the center point of the two-dimensional grid model of leaves of each variety and intersect with the edge contour line set to obtain the sampling point set of each variety; Based on the sampling point set of each variety, the average two-dimensional grid model of the leaves of each variety is determined.
[0107] In some embodiments, the map construction unit is used to: Correlation analysis and cluster analysis were performed on multiple phenotypic characteristic parameters to determine multiple key phenotypic characteristic parameters.
[0108] In some embodiments, the map construction unit is used to: Determine the identification weight corresponding to each key phenotypic characteristic parameter; Based on the key phenotypic characteristic parameters of each variety and the identification weights corresponding to each key phenotypic characteristic parameter, the leaf type characteristic index of each variety is determined; Based on the leaf shape characteristic indexes of each variety, the average two-dimensional grid models of leaves of each variety are sorted to obtain the two-dimensional leaf shape map of the target crop.
[0109] In some embodiments, the variety identification unit is used to: Based on the two-dimensional grid model of the leaves of the crop to be identified, key phenotypic characteristic parameters of the crop to be identified are determined; Comparing the key phenotypic characteristic parameters of the crop to be identified with the key phenotypic characteristic parameters of each variety in the two-dimensional leaf type map of the target crop for similarity; The variety with the highest similarity is determined as the variety of the crop to be identified.
[0110] In some embodiments, the variety identification unit is used to: Obtain key phenotypic characteristic parameters of the crops to be identified; Comparing the key phenotypic characteristic parameters of the crop to be identified with the key phenotypic characteristic parameters of each variety in the two-dimensional leaf type map of the target crop for similarity; The average two-dimensional grid model of the leaves of the variety with the highest similarity is determined as the average two-dimensional grid model of the leaves of the crop to be identified.
[0111] Figure 7 is a schematic diagram of the structure of the electronic device provided by this application, such as Figure 7As shown, the electronic device may include: a processor (Processor) 710, a communication interface (Communications Interface) 720, a memory (Memory) 730 and a communication bus (Communications Bus) 740, wherein the processor 710, the communication interface 720, and the memory 730 communicate with each other through the communication bus 740. The processor 710 may call the logic command in the memory 730 to execute the method described in the above embodiment, for example: The three-dimensional leaf mesh models of each variety of the target crop are meshed and plane parameterized to obtain the two-dimensional leaf mesh models of each variety; multiple phenotypic characteristic parameters of the two-dimensional leaf mesh models of each variety and the average two-dimensional leaf mesh model of each variety are determined; multiple phenotypic characteristic parameters are screened to determine multiple key phenotypic characteristic parameters, and based on the multiple key phenotypic characteristic parameters and the average two-dimensional leaf mesh model of each variety, a two-dimensional leaf type map of the target crop is determined; based on the two-dimensional leaf type map of the target crop, the variety of the crop to be identified is identified.
[0112] In addition, the logical commands in the above-mentioned memory can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several commands to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program code.
[0113] The processor in the electronic device provided in the embodiment of the present application can call the logic instructions in the memory to implement the above method. Its specific implementation method is consistent with the implementation method of the aforementioned method and can achieve the same beneficial effects, which will not be repeated here.
[0114] An embodiment of the present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the methods provided in the above embodiments are implemented.
[0115] Its specific implementation is consistent with the aforementioned method implementation and can achieve the same beneficial effects, so it will not be repeated here.
[0116] An embodiment of the present application provides a computer program product, including a computer program, which implements the above method when executed by a processor.
[0117] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative effort.
[0118] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary 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 prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.
[0119] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.
Claims
1. A crop identification method based on leaf type atlas, characterized in that: include: The three-dimensional mesh model of leaves of each variety of target crops is meshed and plane parameterized to obtain the two-dimensional mesh model of leaves of each variety; Determine a plurality of phenotypic characteristic parameters of the two-dimensional grid model of leaves of each variety and an average two-dimensional grid model of leaves of each variety; Screening the multiple phenotypic characteristic parameters to determine multiple key phenotypic characteristic parameters, and determining a two-dimensional leaf profile of the target crop based on the multiple key phenotypic characteristic parameters and an average two-dimensional grid model of leaves of each variety; Based on the two-dimensional leaf type map of the target crop, the variety of the crop to be identified is identified.
2. The crop identification method based on leaf type atlas according to claim 1, characterized in that: The three-dimensional mesh model of leaves of each variety of target crops is meshed and plane parameterized to obtain the two-dimensional mesh model of leaves of each variety, including: Perform multiple mesh subdivisions on the three-dimensional mesh model of leaves of each variety of target crops to obtain the subdivided three-dimensional mesh model of leaves of each variety; Based on the mesh deformation algorithm that keeps the rigidity unchanged, the subdivided three-dimensional mesh model of each type of blade is plane parameterized to obtain the two-dimensional mesh model of each type of blade.
3. The crop identification method based on leaf type atlas according to claim 2, characterized in that: The method of determining a plurality of phenotypic characteristic parameters of the two-dimensional grid model of leaves of each variety and an average two-dimensional grid model of leaves of each variety comprises: Extract the leaf contours of the two-dimensional mesh model of leaves of each variety to obtain edge contour line sets; A ray is emitted around the center point of the two-dimensional grid model of leaves of each variety to intersect with the edge contour line set to obtain a sampling point set of each variety; Based on the sampling point set of each variety, the average two-dimensional grid model of the leaves of each variety is determined.
4. The crop identification method based on leaf type atlas according to claim 3, characterized in that: The screening of the multiple phenotypic characteristic parameters to determine multiple key phenotypic characteristic parameters includes: Correlation analysis and cluster analysis are performed on the multiple phenotypic characteristic parameters to determine multiple key phenotypic characteristic parameters.
5. The crop identification method based on leaf type atlas according to claim 4, characterized in that: The step of determining the two-dimensional leaf profile of the target crop based on the plurality of key phenotypic characteristic parameters and the average two-dimensional grid model of leaves of each variety comprises: Determine the impact weight corresponding to each key phenotypic characteristic parameter; Based on the key phenotypic characteristic parameters of each variety and the corresponding influence weights of each key phenotypic characteristic parameter, the leaf type characteristic index of each variety is determined; Based on the leaf shape characteristic index of each variety, the average two-dimensional grid model of the leaves of each variety is sorted to obtain a two-dimensional leaf shape map of the target crop.
6. The crop identification method based on leaf type atlas according to claim 5, characterized in that: The method of performing variety identification on the crop to be identified based on the two-dimensional leaf type map of the target crop includes: Determining key phenotypic characteristic parameters of the crop to be identified based on a two-dimensional grid model of leaves of the crop to be identified; Comparing the key phenotypic characteristic parameters of the crop to be identified with the key phenotypic characteristic parameters of each variety in the two-dimensional leaf type map of the target crop for similarity; The variety with the highest similarity is determined as the variety of the crop to be identified.
7. The crop identification method based on leaf type atlas according to claim 5, characterized in that: The method of performing variety identification on the crop to be identified based on the two-dimensional leaf type map of the target crop includes: Obtain key phenotypic characteristic parameters of the crops to be identified; Comparing the key phenotypic characteristic parameters of the crop to be identified with the key phenotypic characteristic parameters of each variety in the two-dimensional leaf type map of the target crop for similarity; The average two-dimensional grid model of leaves of the variety with the highest similarity is determined as the average two-dimensional grid model of leaves of the crop to be identified.
8. A crop identification device based on leaf profile, characterized in that: include: A model flattening unit is used to perform mesh subdivision and plane parameterization on the three-dimensional mesh model of leaves of various varieties of target crops to obtain a two-dimensional mesh model of leaves of various varieties; A parameter extraction unit, used to determine a plurality of phenotypic characteristic parameters of the two-dimensional grid model of leaves of each variety and an average two-dimensional grid model of leaves of each variety; A map construction unit is used to screen the multiple phenotypic characteristic parameters, determine multiple key phenotypic characteristic parameters, and determine the two-dimensional leaf map of the target crop based on the multiple key phenotypic characteristic parameters and the average two-dimensional grid model of leaves of each variety; The variety identification unit is used to perform variety identification on the crop to be identified based on the two-dimensional leaf type map of the target crop.
9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the crop identification method based on leaf type atlas described in any one of claims 1 to 7 is implemented.
10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the crop identification method based on leaf type atlas described in any one of claims 1 to 7 is implemented.