Crop canopy phenotype parameter extraction method and device, medium and product

Abstract

The invention discloses a crop canopy phenotypic parameter extraction method and device, a medium and a product, and relates to the field of phenotypic parameter extraction, and the method comprises the steps: obtaining a three-dimensional point cloud set of a target crop; according to the three-dimensional point cloud set of the target crop, determining a point cloud skeleton of the target crop by using a Laplacian algorithm; then, optimization and instance segmentation are carried out on the target crop point cloud skeleton, and organ-level fine instance segmentation is completed; finally, crop phenotypic parameter extraction is achieved based on the three-dimensional point cloud and skeleton extraction, and phenotypic parameters comprise the plant height, the crown breadth, the leaf number, the leaf length, the leaf width and the leaf included angle. According to the method, the canopy phenotype parameters of the target crop can be accurately determined.

Description

Technical Field

[0001] The present application relates to the field of phenotypic parameter extraction, and in particular to a method, device, medium and product for extracting phenotypic parameters of crop canopy. Background Art

[0002] Phenomics is a discipline that observes and analyzes the morphology, structure, physiology, ecology, etc. of plants and animals. By accurately extracting the phenotypes of crops, we can understand their growth and development mechanisms, the interaction between genetics and the environment, and other information, which is of great significance for breeding. Traditional crop phenotype extraction mainly relies on manual observation and measurement of plant external morphological characteristics. The relevant methods involve the observation and measurement of a large number of samples and the analysis and comparison of massive data, which has the disadvantages of being time-consuming, inefficient, and having large subjective errors. The crop phenotype extraction method based on computer vision overcomes the drawbacks of manual measurement and can automatically obtain various characteristics of crops, such as leaf morphology, chlorophyll fluorescence, root structure, etc., and then conduct analysis of crop growth and development and adverse stress. It has relatively low cost, wide applicability, and high efficiency.

[0003] At present, most of the research on crop phenotype extraction focuses on the two-dimensional direction of computer vision. However, when displaying a three-dimensional object in reality in the form of a two-dimensional image, the information that can be obtained is very limited, and some spatial information may even be lost. Therefore, the current method of crop phenotype extraction based on two-dimensional images has the problem of inaccuracy. Summary of the invention

[0004] The purpose of this application is to provide a method, device, medium and product for extracting crop canopy phenotypic parameters, which can accurately extract crop canopy phenotypic parameters.

[0005] To achieve the above objectives, this application provides the following solutions: In a first aspect, the present application provides a method for extracting crop canopy phenotypic parameters, comprising: Obtain a three-dimensional point cloud set of the target crop; According to the target crop three-dimensional point cloud set, a target crop point cloud skeleton is determined by using a Laplace algorithm; the target crop point cloud skeleton is represented by an undirected graph; Performing skeleton pruning, skeleton calibration and instance segmentation on the target crop point cloud skeleton to obtain a processed target crop point cloud skeleton; The plant height and canopy phenotypic parameters of the target crop are determined according to the processed target crop point cloud skeleton; the canopy phenotypic parameters include canopy width, number of leaves, leaf length, leaf width and leaf angle.

[0006] Optionally, according to the target crop three-dimensional point cloud set, using the Laplace algorithm to determine the target crop point cloud skeleton specifically includes: Using a K-nearest neighbor algorithm and a Laplace algorithm, the three-dimensional point cloud of the target crop is subjected to point cloud contraction to obtain a point cloud skeleton point set after point cloud contraction; According to the point cloud skeleton point set after the point cloud is shrunk, the point cloud skeleton vertex set is obtained by using the farthest point sampling algorithm; According to the point cloud skeleton vertex set, a minimum spanning tree algorithm is used to generate a target crop point cloud skeleton.

[0007] Optionally, skeleton pruning, skeleton calibration and instance segmentation are performed on the target crop point cloud skeleton to obtain a processed target crop point cloud skeleton, which specifically includes: Performing skeleton pruning on the target crop point cloud skeleton to obtain a pruned point cloud skeleton; Performing skeleton calibration on the pruned point cloud skeleton to obtain a calibrated point cloud skeleton; According to the calibrated point cloud skeleton, instance segmentation is performed using a KD tree algorithm to obtain a processed target crop point cloud skeleton.

[0008] Optionally, performing skeleton pruning on the target crop point cloud skeleton to obtain a pruned point cloud skeleton specifically includes: The skeleton points in the undirected graph of the target crop point cloud skeleton are divided into stem nodes, leaf tip nodes and connection nodes; the stem node is a skeleton point connected by at least three edges in the undirected graph; the leaf tip node is a skeleton point connected by only one edge in the undirected graph; and the connection node is a skeleton point connected by two edges in the undirected graph; Determine the node set of leaf tips and the skeleton points of the stem root; Determine the shortest connection path between each leaf tip node in the leaf tip node set and the stem root skeleton point to obtain a shortest connection path set; Comparing the shortest connected paths in the shortest connected path set in pairs; If the overlap between one shortest connected path and another shortest connected path exceeds 75%, and the length difference is less than the set gap, the shortest path in the two shortest connected paths is deleted to obtain the pruned point cloud skeleton.

[0009] Optionally, skeleton calibration is performed on the pruned point cloud skeleton to obtain a calibrated point cloud skeleton, specifically including: Determine the point cloud center according to the calibrated point cloud skeleton; the point cloud center is the centroid of the point cloud skeleton; Decentralizing all point clouds in the calibrated point cloud skeleton according to the point cloud center to obtain a decentralized point cloud; Performing matrix singular value decomposition on the point cloud coordinate matrix to determine the direction vector of the fitting line; the point cloud coordinate matrix is ​​determined based on the decentralized point cloud; Determine a point cloud fitting line according to the point cloud center and the direction vector; The stem skeleton nodes are mapped to the point cloud fitting line to complete the skeleton calibration and obtain a calibrated point cloud skeleton.

[0010] Optionally, determining the plant height and canopy phenotypic parameters of the target crop according to the processed target crop point cloud skeleton specifically includes: Determining the plant height of the target crop according to the maximum and minimum values ​​of the z coordinates in the processed point cloud skeleton of the target crop; Projecting the processed target crop point cloud skeleton onto a two-dimensional plane, finding two-dimensional convex hull points in the projection point set using a Graham scanning method, and determining the crown width according to the Euclidean distance of the two-dimensional convex hull points; Extracting the edge points of the leaf point cloud in the processed target crop point cloud skeleton, obtaining the farthest point pair among the leaf edge points as the leaf base point and the leaf tip point, the shortest path between the leaf base point and the leaf tip point in the leaf point cloud as the leaf length path, and determining the leaf length according to the leaf length path; Determine the leaf width using the Euclidean distance according to the midpoint of the leaf in the processed target crop point cloud skeleton; Determining the number of blades according to the number of blade tip nodes; The leaf angle is determined according to the angle between the edge formed by the leaf midpoint and the leaf base point in the processed target crop point cloud skeleton and the stem.

[0011] In a second aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the crop canopy phenotypic parameter extraction methods described above.

[0012] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for extracting crop canopy phenotypic parameters.

[0013] In a fourth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-described methods for extracting crop canopy phenotypic parameters.

[0014] According to the specific embodiments provided in this application, this application has the following technical effects: The present application provides a crop canopy phenotypic parameter extraction method, equipment, medium and product. First, a crop point cloud skeleton extraction method based on the Laplace algorithm is used to obtain a complete target crop point cloud skeleton; then, the target crop point cloud skeleton is optimized and instance segmented to complete organ-level fine instance segmentation; finally, crop phenotypic parameter extraction is achieved based on three-dimensional point cloud and skeleton extraction. The phenotypic parameters include: plant height and canopy width, number of leaves, leaf length, leaf width, and leaf angle. The present application can accurately determine the canopy phenotypic parameters of the target crop. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 A schematic diagram of a process for extracting crop canopy phenotypic parameters provided in an embodiment of the present application; Figure 2 A schematic diagram of a process flow of a crop canopy phenotypic parameter extraction method based on three-dimensional point cloud and Laplace skeleton extraction provided in an embodiment of the present application; Figure 3 A schematic diagram of a target crop skeleton extracted in one embodiment of the present application; Figure 4 This is a schematic diagram of key points of the stem skeleton extracted in one embodiment of the present application; Figure 5 A schematic diagram of a redundant skeleton provided in an embodiment of the present application; Figure 6 A schematic diagram of removing redundant skeletons provided in an embodiment of the present application; Figure 7 A crop stem skeleton calibration diagram provided in an embodiment of the present application; Figure 8 A schematic diagram of crop organ segmentation example provided in an embodiment of the present application; Fig. 9 This is an example segmentation effect diagram of different crop organs provided in an embodiment of the present application; Fig.10 A schematic diagram of crop canopy phenotype extraction provided in an embodiment of the present application; Fig.11 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions 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 are within the scope of protection of this application.

[0018] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0019] In an exemplary embodiment, Figure 1 and Figure 2 As shown, a method for extracting crop canopy phenotypic parameters is provided, comprising the following steps: S1: Obtain the target crop 3D point cloud set.

[0020] In practical applications, obtain the target crop 3D point cloud set P , It can be expressed as ,in, Refers to the number of points in the point cloud. is a real number space, and each point in three-dimensional space is represented by It consists of three real coordinates.

[0021] S2: Determine a target crop point cloud skeleton based on the target crop three-dimensional point cloud set using a Laplace algorithm; the target crop point cloud skeleton is represented by an undirected graph.

[0022] As an optional implementation, S2 specifically includes: S21: Utilizing a K-nearest neighbor algorithm and a Laplace algorithm, the three-dimensional point cloud of the target crop is subjected to point cloud contraction to obtain a point cloud skeleton point set after point cloud contraction.

[0023] S22: According to the point cloud skeleton point set after the point cloud is shrunk, a point cloud skeleton vertex set is obtained by using a farthest point sampling algorithm.

[0024] S23: Generate a point cloud skeleton of a target crop using a minimum spanning tree algorithm according to the point cloud skeleton vertex set.

[0025] In practical applications, crop point cloud skeleton extraction is performed based on the Laplace algorithm to obtain a complete crop skeleton.

[0026] First, construct the point cloud adjacency graph and calculate the Laplacian matrix for point cloud contraction, including: (1) Use the K-nearest neighbor algorithm to construct a point cloud adjacency graph, capture the local and global structural features in the point cloud, calculate the Laplacian matrix to shrink the point cloud, and effectively extract a representative point cloud skeleton. The point cloud shrinkage equation is: .

[0027] in, is the point cloud matrix (the target crop 3D point cloud set), is the Laplace matrix, and the initial value is the weighted cotangent matrix constructed from the neighborhood points. and Represent the diagonal matrices of contraction force and attraction force respectively.

[0028] (2) According to the point cloud shrinkage equation, use the known point cloud matrix Find the point cloud matrix after iteration , then update the weight matrix and , the point cloud iterative shrinkage equation is: .

[0029] in, represents the weighted transformation of the Laplace matrix, and Indicates the contraction point The neighborhood length before and after contraction.

[0030] (3) Based on the updated point cloud matrix A new Laplacian matrix can be constructed , repeat this step until the inequality is satisfied: .

[0031] in, is a manually defined shrinkage threshold, and its value is 0.01. So far, this embodiment completes the target crop skeleton extraction based on the Laplace algorithm. Figure 3 As shown, (a) is the three-dimensional point cloud of the target crop, (b) is the crop point cloud skeleton extracted based on the Laplace algorithm, and (c) is the effect of the skeleton in the point cloud.

[0032] Then, the crop point cloud skeleton is simplified, including: Point cloud skeleton point set after point cloud shrinkage To refine, the Farthest Point Sampling (FPS) algorithm is used to extract the points from the point set. Get the point cloud skeleton vertex set , the points in this set are regarded as key points, such as Figure 4As shown, (a) is the crop point cloud skeleton extracted based on the Laplace algorithm, and (b) is the effect diagram of the key nodes of the skeleton.

[0033] Finally, since the directly connected skeleton topology may have loops or fail to form a graph, redundant skeletons are formed, such as Figure 5 As shown, the minimum spanning tree algorithm is used to simplify the connected skeleton structure again, aiming to find a tree so that all vertices in the graph can be connected through the edges in the tree, and the sum of the weights of these edges is minimal.

[0034] Point cloud skeleton vertex collection The points in are taken as vertices of the undirected graph, and the Euclidean distances between the points are taken as the weights of the edges of the undirected graph. The generated minimum spanning tree regards it as a plant skeleton undirected graph (target crop point cloud skeleton) to achieve skeleton refinement.

[0035] S3: performing skeleton pruning, skeleton calibration and instance segmentation on the target crop point cloud skeleton to obtain a processed target crop point cloud skeleton.

[0036] As an optional implementation, S3 specifically includes: S31: performing skeleton pruning on the target crop point cloud skeleton to obtain a pruned point cloud skeleton.

[0037] As an optional implementation, S31 specifically includes: The skeleton points in the undirected graph of the target crop point cloud skeleton are divided into stem nodes, leaf tip nodes and connection nodes; the stem nodes are skeleton points connected by at least three edges in the undirected graph; the leaf tip nodes are skeleton points connected by only one edge in the undirected graph; and the connection nodes are skeleton points connected by two edges in the undirected graph.

[0038] Determine the leaf tip node set and the stem root skeleton point.

[0039] The shortest connection path between each leaf tip node in the leaf tip node set and the stem root skeleton point is determined to obtain a shortest connection path set.

[0040] The shortest connected paths in the shortest connected path set are compared pairwise.

[0041] If the overlap between one shortest connected path and another shortest connected path exceeds 75%, and the length difference is less than the set gap, the shortest path in the two shortest connected paths is deleted to obtain the pruned point cloud skeleton.

[0042] In practical applications, the target crop point cloud skeleton pruning method includes: According to the number of edges connecting the skeleton points, the skeleton points are divided into three categories: stem nodes, leaf tip nodes, and connection nodes: (1) Stem node: In an undirected graph A skeleton node that has at least three edges connecting it.

[0043] (2) Leaf tip node: In an undirected graph There is only one edge connecting the skeleton nodes.

[0044] (3) Connecting nodes: In an undirected graph A skeleton node with two connected edges.

[0045] In addition, the skeleton node at the bottom of the stem has only one connecting edge and belongs to the stem node. Since plants grow vertically on the xOy plane, the point with the smallest z value among the skeleton nodes connected by only one edge is regarded as the stem skeleton node and belongs to the stem node rather than the leaf tip node.

[0046] like Figure 6 As shown in the figure, (a) is the target crop point cloud skeleton containing redundant skeletons, (b) is the identification map of the redundant skeleton in the target crop point cloud skeleton, and (c) is the pruned point cloud skeleton, that is, the redundant skeleton is pruned. The steps for pruning the skeleton are as follows: (1) Input the undirected graph of the target crop point cloud skeleton , find the leaf tip node set The skeleton point at the root of the stem .

[0047] (2) Find the i-th leaf tip node in sequence and The shortest path between , forming the shortest connected path set .

[0048] (3) Assume and is the shortest connected path set For any two paths in and If the overlap between two paths exceeds 75% and there is no large length difference, the shorter of the two paths is deleted.

[0049] (4) Reconstruct the skeleton undirected graph based on the newly obtained path after pruning ,in, are all the key nodes in the path, The connected nodes.

[0050] S32: performing skeleton calibration on the pruned point cloud skeleton to obtain a calibrated point cloud skeleton.

[0051] As an optional implementation, S32 specifically includes: The point cloud center is determined according to the calibrated point cloud skeleton; the point cloud center is the centroid of the point cloud skeleton.

[0052] All point clouds in the calibrated point cloud skeleton are decentralized according to the point cloud center to obtain a decentralized point cloud.

[0053] Perform matrix singular value decomposition on the point cloud coordinate matrix to determine the direction vector of the fitting line; the point cloud coordinate matrix is ​​determined based on the decentralized point cloud.

[0054] A point cloud fitting line is determined according to the point cloud center and the direction vector.

[0055] The stem skeleton nodes are mapped to the point cloud fitting line to complete the skeleton calibration and obtain a calibrated point cloud skeleton.

[0056] In practical applications, the stem skeleton points in the pruned point cloud skeleton are subjected to three-dimensional linear fitting and mapping transformation to calibrate the stem skeleton, such as Figure 7 As shown, (a) is the point cloud skeleton after pruning, and (b) is the point cloud skeleton after calibration. The steps are: (1) Obtain the centroid of the point cloud and regard it as the center of the point cloud .

[0057] (2) Subtract the coordinates of the center of the point cloud from the coordinates of all points in the point cloud to decentralize the point cloud.

[0058] (3) Extract the decentralized point cloud coordinate matrix and perform matrix singular value decomposition to obtain the eigenvector corresponding to the maximum eigenvalue, which is the direction vector of the fitting line. .

[0059] (4) Based on the direction vector of the point cloud center and the fitted line The point cloud fitting straight line expression can be obtained, and the calculation formula is as follows: .

[0060] (5) The stem skeleton nodes Mapped onto the fitting line, it becomes , complete the crop point cloud skeleton calibration, the calculation formula is as follows: .

[0061] S33: Based on the calibrated point cloud skeleton, the KD tree algorithm is used to perform instance segmentation to obtain a processed target crop point cloud skeleton, such as Figure 8 As shown, (a) is the crop point cloud skeleton before instance segmentation, and (b) is the point cloud skeleton after instance segmentation.

[0062] In practical applications, after completing the crop point cloud optimization, the shortest path connecting the leaf tip node and the connecting node is taken as the leaf skeleton path, and the skeleton points contained in the leaf skeleton path belong to the leaf skeleton points. The shortest connected path of the stem skeleton points is taken as the stem skeleton path, and the points contained in the stem skeleton path belong to the stem skeleton points.

[0063] A KD tree is constructed to query all points whose distance from the anchor point is less than a threshold distance, which is regarded as the current organ point cloud. The anchor point is a skeleton point, and the threshold distance is set to the average Euclidean distance between the anchor point and two adjacent skeleton points on the current organ skeleton path.

[0064] If some point clouds are not correctly classified in the rough segmentation, further fine segmentation is required to calculate the number of points in the unclassified successful point set. The organ point cloud to which the skeleton point with the smallest Euclidean distance belongs is The organ to which it belongs is the organ, and the crop point cloud organ-level fine instance segmentation is completed, such as Fig. 9 As shown, (a) is the three-dimensional point cloud of the target crop, (b) is the point cloud skeleton of the target crop, (c) is the effect of the point cloud after coarse segmentation, and (d) is the effect of the point cloud after fine segmentation.

[0065] S4: Determine the plant height and canopy phenotypic parameters of the target crop according to the processed target crop point cloud skeleton, such as Fig.10 As shown, (a) is a schematic diagram for calculating leaf length, (b) is a schematic diagram for calculating leaf width, (c) is a schematic diagram for calculating leaf angle, (d) is a schematic diagram for calculating plant height, and (e) is a schematic diagram for calculating crown width; the canopy phenotypic parameters include crown width, number of leaves, leaf length, leaf width and leaf angle.

[0066] As an optional implementation, S4 specifically includes: S41: determining the plant height of the target crop according to the maximum value and the minimum value of the z coordinate in the processed target crop point cloud skeleton.

[0067] S42: Project the processed target crop point cloud skeleton onto a two-dimensional plane, use Graham scanning to find the two-dimensional convex hull points in the projection point set, and determine the crown width according to the Euclidean distance of the two-dimensional convex hull points. In practical applications, find a pair of points with the farthest Euclidean distance in the two-dimensional convex hull points as the crown width of the target crop.

[0068] S43: extracting the edge points of the leaf point cloud in the processed target crop point cloud skeleton, obtaining the farthest point pair in the leaf edge points as the leaf base point and the leaf tip point, the shortest path between the leaf base point and the leaf tip point in the leaf point cloud as the leaf length path, and determining the leaf length according to the leaf length path. In practical applications, the leaf length path distance value is used as the leaf length.

[0069] S44: Determine the leaf width using the Euclidean distance according to the midpoint of the leaf in the processed target crop point cloud skeleton.

[0070] In practical applications, the leaf length path point at the midpoint of the leaf is expressed as ,from Start and construct directions respectively and Vector and , using vector and as the normal vector, and For the points on the cutting plane, derive the cutting plane and The equation of the cutting plane in the leaf point cloud and The points between are considered as the leaf width position range, within which the leaf width position range is along the vector and Linear scan is performed in the direction of to obtain a set of point pairs with the largest Euclidean distance on the scan line segment. The Euclidean distance between the point pairs is the leaf width of the point, and the maximum leaf width of all points in the position range is taken as the leaf width of the position range.

[0071] S45: determining the leaf angle according to the angle between the edge formed by the leaf midpoint and the leaf base point in the processed target crop point cloud skeleton and the stem.

[0072] The leaf angle is the angle between the edge formed by the midpoint of the leaf and the base of the leaf and the stem. Let the base of the leaf be point O, point M be a point on the z-axis, and point N be a point on the leaf length path that is one-quarter the length of the intersection point O. The calculation formula for the leaf angle α is as follows: .

[0073] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the above-mentioned crop canopy phenotypic parameter extraction method when executing the computer program.

[0074] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, which implements the above-mentioned crop canopy phenotypic parameter extraction method when executed by a processor.

[0075] In an exemplary embodiment, a computer program product is provided, including a computer program, which implements the above-mentioned crop canopy phenotypic parameter extraction method when executed by a processor.

[0076] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Fig.11 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for extracting crop canopy phenotypic parameters is implemented.

[0077] Those skilled in the art will understand that Fig.11 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0078] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0079] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0080] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0081] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0082] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for extracting crop canopy phenotypic parameters, characterized in that: include: Obtain a three-dimensional point cloud set of the target crop; According to the target crop three-dimensional point cloud set, using the Laplace algorithm, determining the target crop point cloud skeleton; The target crop point cloud skeleton is represented by an undirected graph; Performing skeleton pruning, skeleton calibration and instance segmentation on the target crop point cloud skeleton to obtain a processed target crop point cloud skeleton; The plant height and canopy phenotypic parameters of the target crop are determined according to the processed target crop point cloud skeleton; the canopy phenotypic parameters include canopy width, number of leaves, leaf length, leaf width and leaf angle.

2. The method for extracting crop canopy phenotypic parameters according to claim 1, characterized in that: According to the target crop three-dimensional point cloud set, the target crop point cloud skeleton is determined by using the Laplace algorithm, specifically including: Using a K-nearest neighbor algorithm and a Laplace algorithm, the three-dimensional point cloud of the target crop is subjected to point cloud contraction to obtain a point cloud skeleton point set after point cloud contraction; According to the point cloud skeleton point set after the point cloud is shrunk, the point cloud skeleton vertex set is obtained by using the farthest point sampling algorithm; According to the point cloud skeleton vertex set, a minimum spanning tree algorithm is used to generate a target crop point cloud skeleton.

3. The method for extracting crop canopy phenotypic parameters according to claim 1, characterized in that: Performing skeleton pruning, skeleton calibration and instance segmentation on the target crop point cloud skeleton to obtain a processed target crop point cloud skeleton, specifically including: Performing skeleton pruning on the target crop point cloud skeleton to obtain a pruned point cloud skeleton; Performing skeleton calibration on the pruned point cloud skeleton to obtain a calibrated point cloud skeleton; According to the calibrated point cloud skeleton, instance segmentation is performed using a KD tree algorithm to obtain a processed target crop point cloud skeleton.

4. The method for extracting crop canopy phenotypic parameters according to claim 3, characterized in that: Performing skeleton pruning on the target crop point cloud skeleton to obtain a pruned point cloud skeleton specifically includes: The skeleton points in the undirected graph of the target crop point cloud skeleton are divided into stem nodes, leaf tip nodes and connection nodes; the stem node is a skeleton point connected by at least three edges in the undirected graph; the leaf tip node is a skeleton point connected by only one edge in the undirected graph; and the connection node is a skeleton point connected by two edges in the undirected graph; Determine the node set of leaf tips and the skeleton points of the stem root; Determine the shortest connection path between each leaf tip node in the leaf tip node set and the stem root skeleton point to obtain a shortest connection path set; Comparing the shortest connected paths in the shortest connected path set in pairs; If the overlap between one shortest connected path and another shortest connected path exceeds 75%, and the length difference is less than the set gap, the shortest path in the two shortest connected paths is deleted to obtain the pruned point cloud skeleton.

5. The method for extracting crop canopy phenotypic parameters according to claim 3, characterized in that: Performing skeleton calibration on the pruned point cloud skeleton to obtain a calibrated point cloud skeleton specifically includes: Determine the point cloud center according to the calibrated point cloud skeleton; the point cloud center is the centroid of the point cloud skeleton; Decentralizing all point clouds in the calibrated point cloud skeleton according to the point cloud center to obtain a decentralized point cloud; Performing matrix singular value decomposition on the point cloud coordinate matrix to determine the direction vector of the fitting line; the point cloud coordinate matrix is ​​determined based on the decentralized point cloud; Determine a point cloud fitting line according to the point cloud center and the direction vector; The stem skeleton nodes are mapped to the point cloud fitting line to complete the skeleton calibration and obtain a calibrated point cloud skeleton.

6. The method for extracting crop canopy phenotypic parameters according to claim 4, characterized in that: According to the processed target crop point cloud skeleton, the plant height and canopy phenotypic parameters of the target crop are determined, specifically including: Determining the plant height of the target crop according to the maximum and minimum values ​​of the z coordinates in the processed point cloud skeleton of the target crop; Projecting the processed target crop point cloud skeleton onto a two-dimensional plane, finding two-dimensional convex hull points in the projection point set using a Graham scanning method, and determining the crown width according to the Euclidean distance of the two-dimensional convex hull points; Extracting the edge points of the leaf point cloud in the processed target crop point cloud skeleton, obtaining the farthest point pair among the leaf edge points as the leaf base point and the leaf tip point, the shortest path between the leaf base point and the leaf tip point in the leaf point cloud as the leaf length path, and determining the leaf length according to the leaf length path; Determine the leaf width using the Euclidean distance according to the midpoint of the leaf in the processed target crop point cloud skeleton; Determining the number of blades according to the number of blade tip nodes; The leaf angle is determined according to the angle between the edge formed by the leaf midpoint and the leaf base point in the processed target crop point cloud skeleton and the stem.

7. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for extracting crop canopy phenotypic parameters according to any one of claims 1 to 6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the crop canopy phenotypic parameter extraction method described in any one of claims 1 to 6 is implemented.

9. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the crop canopy phenotypic parameter extraction method described in any one of claims 1 to 6 is implemented.

Images