Root point cloud skeleton extraction method, system and device

Abstract

The present invention discloses a root point cloud skeleton extraction method, system, and device. The method comprises: obtaining a plant root point cloud, dividing the plant root point cloud into point cloud segments to form a root voxel grid set; constructing a root voxel map, wherein edges of the root voxel map are obtained by connecting the current root voxel grid and adjacent root voxel grids based on their spatial positions, and obtaining the weight of the root voxel map by constructing a point cloud search space and obtaining the point cloud density within the point cloud search space; constructing a root voxel spanning tree based on the root voxel map, determining noise in the root voxel map based on the root voxel spanning tree, and obtaining an initial root skeleton map by filtering; obtaining root connection data of the current root voxel grid, and filtering root voxel grids corresponding to burrs by determining to obtain a root skeleton map. The present invention reduces skeleton misjudgment by extracting skeletons from point cloud data and solves problems such as smearing, noise, and burrs on the root surface.

Description

Technical Field

[0001] The present invention relates to the technical field of point cloud extraction, and in particular to a method, system and device for extracting a root point cloud skeleton. Background Art

[0002] Point cloud skeleton extraction is a method for extracting skeleton structures from point cloud data. It can be used in areas such as object shape analysis and 3D model simplification. Existing point cloud skeleton extraction methods include Laplace Contraction-based skeleton extraction methods, Rotational Symmetry Axis-based skeleton extraction methods, and L1-Medial-based skeleton extraction methods. Among them, the Laplace Contraction-based skeleton extraction method uses the Laplace operator to process point cloud data to obtain an estimate. This method can be used to process complex geometric shapes, but is susceptible to point cloud noise. The Rotational Symmetry Axis-based skeleton extraction method extracts the rotational symmetry axis of the object in a certain direction and uses the symmetry axis as part of the skeleton. This method is effective for objects with obvious rotational symmetry, but the calculation process of the symmetry axis in this method is relatively complex. The L1-Medial-based skeleton extraction method uses the idea of ​​spatial median to extract the geometric features of point cloud data as a skeleton. This method can process point cloud data with complex geometric shapes, but requires adjusting a large number of parameters during the extraction process.

[0003] Due to the uneven distribution of root point clouds, the presence of ghosting on the root surface, the rough surface, and the presence of discrete noise points near the roots, the root skeletons obtained by existing root extraction methods have many misjudgments. In addition, the existing methods have many parameters that need to be fine-tuned, resulting in a low accuracy rate in root skeleton extraction. Summary of the Invention

[0004] In view of the shortcomings of the prior art, the present invention provides a root point cloud skeleton extraction method, system and device.

[0005] In order to solve the above technical problems, the present invention is solved by the following technical solutions:

[0006] A root system point cloud skeleton extraction method comprises the following steps:

[0007] Obtain plant root point cloud and divide the point cloud to form a root voxel grid set;

[0008] Constructing a root voxel map, wherein the edges of the root voxel map are obtained by connecting the spatial positions of the current root voxel grid and the spatial positions of the adjacent root voxel grids, and the weights of the root voxel map are obtained by constructing a point cloud search space and obtaining the point cloud density in the point cloud search space;

[0009] According to the weight of the root voxel map and the root voxel map, a root voxel spanning tree is constructed based on the minimum spanning tree algorithm to determine whether there is noise in the root voxel map, and the edges corresponding to the noise are filtered to obtain the initial root skeleton map;

[0010] The connection relationship between the current root voxel grid and the adjacent root voxel grid in the initial root skeleton map is obtained to obtain the root connection data. It is also determined whether there are burrs in the initial root skeleton map. The root voxel grids corresponding to the burrs are filtered to obtain the root skeleton map.

[0011] As an implementable method, the point cloud segmentation of the plant root point cloud includes the following steps:

[0012] Obtaining a plant root point cloud and determining a three-dimensional spatial boundary of the plant root point cloud, wherein the three-dimensional spatial boundary includes a horizontal boundary, a longitudinal boundary, and a vertical boundary;

[0013] The plant root point cloud within the three-dimensional space boundary is divided according to the preset grid threshold to obtain the initial root voxel grid set;

[0014] Obtain the number of point clouds contained in all initial root voxel grids in the initial root voxel grid set to obtain a point cloud quantity set, set a point cloud quantity threshold, obtain the number of point clouds in the point cloud quantity set that do not meet the point cloud quantity threshold, and filter the corresponding initial root voxel grids to obtain a filtered voxel grid set;

[0015] In the filtering voxel grid set, if the filtering voxel grid set still has connectivity after deleting the current filtering voxel grid, the current filtering voxel grid is deleted, and the root voxel grid set is obtained by iterative deletion.

[0016] As an implementable method, the step of obtaining the edge of the root voxel graph by connecting the spatial position of the current root voxel grid and the spatial positions of adjacent root voxel grids includes the following steps:

[0017] The spatial position of the root voxel grid is obtained through the center position of the root voxel grid;

[0018] Preset a search threshold, construct a search space of the current root voxel grid based on the search threshold, and determine whether there are adjacent root voxel grids in the search space;

[0019] If it exists, the spatial position of the current root voxel grid and the spatial position of the adjacent root voxel grid are connected to obtain the edge of the root voxel graph.

[0020] As an implementable method, the method of obtaining the weight of the root voxel map by constructing a point cloud search space and obtaining the point cloud density in the point cloud search space includes the following steps:

[0021] The spatial position of the root voxel grid is used as the sphere center and the preset grid threshold is used as the diameter to construct the point cloud search area.

[0022] Obtain the number of plant root point clouds in the point cloud search area and perform density calculation to obtain the root point cloud density;

[0023] Normalize the root point cloud density to obtain the normalized point cloud density;

[0024] The weight of the root voxel map is obtained by calculation based on the adjacent normalized point cloud density;

[0025] Among them, the normalized point cloud density and the weight of the root voxel map are expressed as follows:

[0026]

[0027]

[0028] in, , Indicates the maximum value of the root point cloud density, represents the density of root point cloud, Indicates the number of root voxel grids, Indicates the Normalized point cloud density, Indicates the A root voxel grid, Indicates the The root voxel grid and the The weight of the root voxel map between the root voxel grids, Indicates the Normalized point cloud density.

[0029] As an implementation method, the method constructs a root voxel spanning tree based on the minimum spanning tree algorithm according to the weight of the root voxel map and the root voxel map, and then determines whether there is noise in the root voxel map, and filters the edges corresponding to the noise to obtain an initial root skeleton map, including the following steps:

[0030] Sort the root voxel map according to its weight to obtain a sorted voxel map;

[0031] Initialize the root spanning set, determine whether the root voxel grid corresponding to the edge of the current sorted voxel graph belongs to the root spanning set, if not, add the edge of the current sorted voxel graph to the root spanning set;

[0032] If yes, filter the edges of the current sorted voxel graph and traverse the sorted voxel graph until the edge of the root spanning set can connect all the root voxel grids in the sorted voxel graph, and the sum of the weights of the root voxel spanning set is the smallest, and obtain the root voxel spanning tree;

[0033] The edges in the root voxel map that do not belong to the root voxel spanning tree are filtered out to obtain the initial root skeleton map.

[0034] As an implementable method, obtaining the connection relationship between the current root voxel grid and the adjacent root voxel grids to obtain the root connection data includes the following steps:

[0035] Initialize the root connection data of the current root voxel grid in the initial root skeleton map to 0;

[0036] The edges of the initial root skeleton graph are traversed to obtain the number of edges of the initial root skeleton graph connected to the current root voxel grid, and the root connection data corresponding to all root voxel grids are obtained.

[0037] As an implementation method, the determining whether there are burrs in the initial root skeleton map and filtering the root voxel grids corresponding to the burrs to obtain the root skeleton map includes the following steps:

[0038] Obtain root connection data of the current root voxel grid in the initial root skeleton image to obtain first voxel data, and obtain root connection data of the adjacent root voxel grid to obtain second voxel data;

[0039] Based on the first voxel data and the second voxel data, determining whether the root voxel grid meets the burr condition, and if so, determining it as a burr, wherein the burr condition is at least: the first voxel data is 1 and the second voxel data is greater than 2;

[0040] Traverse the initial root skeleton map, filter the root voxel grid corresponding to the burr, and obtain the root skeleton map. A root point cloud skeleton extraction system includes a point cloud segmentation module, a voxel map construction module, a noise filtering module, and a burr filtering module;

[0041] The point cloud segmentation module obtains the plant root point cloud and performs point cloud segmentation to form a root voxel grid set;

[0042] The voxel map construction module constructs a root voxel map, wherein the edges of the root voxel map are obtained by connecting the spatial positions of the current root voxel grid and the spatial positions of the adjacent root voxel grids, and the weights of the root voxel map are obtained by constructing a point cloud search space and obtaining the point cloud density in the point cloud search space;

[0043] The noise filtering module constructs a root voxel spanning tree based on the minimum spanning tree algorithm according to the weight of the root voxel map and the root voxel map, and then determines whether there is noise in the root voxel map, and filters the edges corresponding to the noise to obtain an initial root skeleton map;

[0044] The burr filtering module obtains the connection relationship between the current root voxel grid and the adjacent root voxel grid in the initial root skeleton map, obtains root connection data, and determines whether there are burrs in the initial root skeleton map, filters the root voxel grids corresponding to the burrs, and obtains the root skeleton map.

[0045] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following method:

[0046] Obtain plant root point cloud and divide the point cloud to form a root voxel grid set;

[0047] Constructing a root voxel map, wherein the edges of the root voxel map are obtained by connecting the spatial positions of the current root voxel grid and the spatial positions of the adjacent root voxel grids, and the weights of the root voxel map are obtained by constructing a point cloud search space and obtaining the point cloud density in the point cloud search space;

[0048] According to the weight of the root voxel map and the root voxel map, a root voxel spanning tree is constructed based on the minimum spanning tree algorithm to determine whether there is noise in the root voxel map, and the edges corresponding to the noise are filtered to obtain the initial root skeleton map;

[0049] The connection relationship between the current root voxel grid and the adjacent root voxel grid in the initial root skeleton map is obtained to obtain the root connection data. It is also determined whether there are burrs in the initial root skeleton map. The root voxel grids corresponding to the burrs are filtered to obtain the root skeleton map.

[0050] A root system point cloud skeleton extraction device includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the following method is implemented:

[0051] Obtain plant root point cloud and divide the point cloud to form a root voxel grid set;

[0052] Constructing a root voxel map, wherein the edges of the root voxel map are obtained by connecting the spatial positions of the current root voxel grid and the spatial positions of the adjacent root voxel grids, and the weights of the root voxel map are obtained by constructing a point cloud search space and obtaining the point cloud density in the point cloud search space;

[0053] According to the weight of the root voxel map and the root voxel map, a root voxel spanning tree is constructed based on the minimum spanning tree algorithm to determine whether there is noise in the root voxel map, and the edges corresponding to the noise are filtered to obtain the initial root skeleton map;

[0054] The connection relationship between the current root voxel grid and the adjacent root voxel grid in the initial root skeleton map is obtained to obtain the root connection data. It is also determined whether there are burrs in the initial root skeleton map. The root voxel grids corresponding to the burrs are filtered to obtain the root skeleton map.

[0055] The present invention has significant technical effects due to the adoption of the above technical solutions:

[0056] The present invention divides a plant root point cloud and constructs a root voxel map. Based on the edges and weights of the root voxel map, the noise present in the root voxels is filtered to extract a root skeleton map. This method solves the problem of frequent misjudgments in existing root point cloud skeleton extraction methods. Furthermore, by filtering out noise, the present invention addresses root misjudgments caused by uneven distribution of plant root point clouds, smearing on the root surface, and uneven smoothness. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0058] Figure 1 It is a schematic flow diagram of the method of the present invention;

[0059] Figure 2 It is an overall schematic diagram of the system of the present invention;

[0060] Figure 3 This is a schematic diagram of a plant root point cloud and a root voxel grid according to the present invention;

[0061] Figure 4 is a schematic diagram of the edge of the noise and the corresponding root voxel map in the present invention;

[0062] Figure 5 It is a schematic diagram of the burr of the initial root system skeleton diagram of the present invention. DETAILED DESCRIPTION

[0063] The present invention will be further described in detail below with reference to the examples. The following examples are intended to explain the present invention but the present invention is not limited to the following examples.

[0064] Example 1:

[0065] A root point cloud skeleton extraction method, such as Figure 1 As shown, the following steps are included:

[0066] S100, obtaining a plant root point cloud and dividing the point cloud to form a root voxel grid set;

[0067] S200, constructing a root voxel map, wherein edges of the root voxel map are obtained by connecting the spatial positions of the current root voxel grid and the spatial positions of adjacent root voxel grids, and weights of the root voxel map are obtained by constructing a point cloud search space and obtaining the point cloud density in the point cloud search space;

[0068] S300, constructing a root voxel spanning tree based on the minimum spanning tree algorithm according to the weight of the root voxel map and the root voxel map, and then determining whether there is noise in the root voxel map, and filtering the edges corresponding to the noise to obtain an initial root skeleton map;

[0069] S400, obtaining the connection relationship between the current root voxel grid and the adjacent root voxel grid in the initial root skeleton map, obtaining root connection data, and determining whether there are burrs in the initial root skeleton map, filtering the root voxel grids corresponding to the burrs, and obtaining a root skeleton map.

[0070] The present invention obtains a plant root point cloud and forms a root voxel grid set by dividing the point cloud. In the prior art, voxel is a three-dimensional concept, commonly used in fields such as computer graphics, medical imaging, and three-dimensional modeling. It is a basic unit in three-dimensional space, similar to a pixel in a two-dimensional image. A voxel can be defined as a cube or a more complex geometric shape, and is usually used to construct a three-dimensional model or represent three-dimensional data, similar to a pixel used to construct a two-dimensional image. A root voxel map is constructed based on the root voxel grid set, and a root voxel spanning tree is obtained by using a minimum spanning tree algorithm in combination with the weight of the root voxel map. The noise in the root voxel map is judged, and the edges of the root voxel map corresponding to the noise are filtered. Then, through the connection relationship between the root voxel points, the root voxel grids corresponding to the burrs in the initial root skeleton map are filtered to obtain a root skeleton map. Since the plant root point cloud obtained by the three-dimensional reconstruction algorithm has problems such as uneven distribution, ghosting, burrs and discrete noise points on the root surface, the method of the present invention is used to filter the noise and burrs in the plant root point cloud, thereby improving the accuracy of the root skeleton map. Without the need for fine-tuning of parameters, the applicability and generalization ability are stronger.

[0071] In this embodiment, the plant root point cloud is a dense point cloud. This high-density point cloud data, acquired through high-precision 3D scanning technology, can detail the root's morphology, structure, and spatial distribution. Composed of a large number of points with coordinate information in 3D space, this point cloud provides highly detailed 3D spatial information, clearly reflecting surface details and geometric features. It is typically acquired using technologies such as LiDAR, depth cameras, or photogrammetry.

[0072] Plant root point cloud in this embodiment The three-dimensional space boundary includes horizontal boundary, vertical boundary and vertical boundary, among which the horizontal boundary is , the vertical boundary is , the vertical boundary is , preset grid threshold , the plant root point cloud within the three-dimensional space boundary is divided into point clouds by a preset grid threshold to obtain an initial root voxel grid set, the number of plant root point clouds in all initial root voxel grids in the initial root voxel grid set is calculated to form a point cloud quantity set, a preset point cloud quantity threshold is set, the point cloud quantity set is compared, and the initial root voxels corresponding to the point cloud quantity that does not meet the point cloud quantity threshold in the point cloud quantity set are filtered to obtain a filtered voxel grid set. In this embodiment, the idea of ​​the VoxelThinning algorithm is used to further refine and delete the filtered root voxel grid set, and the filtered root voxels in the filtered root voxel grid set are deleted by a preset template. If the connectivity of the filtered root voxel grid set is not affected after the current filtered root voxel grid is deleted, the current filtered root voxel grid is deleted, otherwise it is retained, and the root voxel grid set is obtained by iteratively deleting the filtered root voxel grid set. ,in, Indicates the A root voxel grid, Indicates the number of root voxel grids. In the method of this embodiment, , each root voxel grid is a cube, the center of the cube is in the spatial position , the side length, i.e. the grid threshold, is , the schematic diagram of plant root point cloud and root voxel grid is as follows Figure 3 shown.

[0073] For the root voxel grid set All In this embodiment, The search space of the current root voxel grid is constructed within the range of the 26 adjacent root voxel grids, and a search is performed in the search space to determine whether there are other adjacent root voxel grids. , if present, the current root voxel grid spatial location and adjacent root voxel grids spatial location Connected to form the edges of the root voxel graph ; Using the current root voxel grid spatial location is the center of the sphere, the grid threshold Construct a spherical point cloud search area for the diameter, and search for plant root point clouds within the point cloud search area. The number of point clouds that fall within the point cloud search area is calculated, and the point cloud density within the point cloud search area is calculated to obtain the root point cloud density. , where the specific process of calculating the root point cloud density is as follows:

[0074] [IDX, ~]= rangesearch(PC.Location, voxelCenter, radius,

[0075] 'Distance','euclidean','NSMethod','kdtree');

[0076] PC.Location: Plant root point cloud The three-dimensional space coordinates of

[0077] voxelCenter: Each line is ;

[0078] Radius: The radius of the point cloud search space, which is set to GridSize / 2;

[0079] IDX: cells, each cell contains The spherical search range Point cloud index;

[0080] Obtain the root point cloud density of all root voxel grids in the root voxel grid set, calculate the maximum value of all root point cloud densities, and obtain the normalized point cloud density through normalization, which is expressed as follows:

[0081]

[0082]

[0083] The edge of the root voxel map is obtained by calculating the normalized point cloud density of the current root voxel grid and the normalized point cloud density of the adjacent root voxel grids. The weight of is expressed as follows:

[0084]

[0085] in, Indicates the maximum value of the root point cloud density, represents the density of root point cloud, Indicates the number of root voxel grids, Indicates the Normalized point cloud density, Indicates the A root voxel grid, Indicates the The root voxel grid and the The weight of the root voxel map between the root voxel grids, , Indicates the Normalized point cloud density.

[0086] After obtaining the edges and weights of the root voxel map, construct the root voxel map , which is expressed as follows:

[0087]

[0088]

[0089]

[0090] in, represents the root voxel map, represents the set of edges of the root voxel graph, A collection of weights representing the root voxel map.

[0091] In this embodiment, the noise includes smear and noise points. The noise and the corresponding edge of the root voxel map are as follows: Figure 4 As shown in the figure, since the plant root point cloud has a high density of point clouds on the root system and a low density of point clouds in the noise area, a root voxel map is constructed so that the point clouds on the dense root system are easier to connect to form a root voxel map. In the root voxel map, the noise needs to be filtered. The process of filtering the edges of the root voxel map generated by the noise in the root voxel map to obtain the initial root skeleton map includes the following steps:

[0092] Step 1: Sort the edges of the corresponding root voxel map according to the weight of the root voxel map to obtain the sorted voxel map;

[0093] Step 2: Initialize the root voxel generation set , determine whether the vertex corresponding to the edge of the sorted voxel graph belongs to the root system spanning set. If not, add the edge of the sorted voxel graph to the root system voxel spanning set. If it does, skip it;

[0094] Step 3: Iterate through the edges of the sorted voxel graph until the edges in the root spanning set can connect all the root voxel grids in the sorted voxel graph and the sum of the weights of the root voxel spanning set is minimized, thus obtaining the root voxel spanning tree.

[0095] Step 4: Use the root voxel spanning tree to determine the noise in the root voxel map, filter the edges in the root voxel map that do not belong to the root voxel spanning tree, and obtain the initial root skeleton map .

[0096] In this embodiment, filtering the edges of the root voxel map is implemented by a minimum spanning tree, specifically using the minimum spanning tree minspantree function in MATLAB. The pseudo code of this process is as follows:

[0097] Gske = minspantree(Gvox,'Method','sparse')

[0098] in, represents the root voxel map, Represents the initial root system skeleton diagram.

[0099] After filtering the edges of the root voxel map corresponding to the smear and noise points in the root voxel map, there are burrs in the initial root skeleton map, such as Figure 5 As shown, it is necessary to filter the root voxel grids corresponding to the burrs in the initial root skeleton map to obtain the root skeleton map, which includes the following steps:

[0100] Step 1: Initialize the root connection data of the current root voxel grid to 0, traverse the edges of the initial root skeleton graph, obtain the number of all edges connected to the current root voxel grid, and obtain the root connection data of the current root voxel grid;

[0101] Step 2: Traverse the root voxel grids in the initial root skeleton image to obtain the root connection data of all root voxel grids. In this embodiment, the root connection data is calculated using the degree function in MATLAB and is expressed as follows:

[0102] Deg = degree( )

[0103] in, represents the initial root system skeleton diagram, Represents the initial root system skeleton diagram Root connectivity data for the meso-root voxel grid;

[0104] Step 3: Determine whether the first voxel data and the second voxel data meet the burr condition. In this embodiment, the burr condition is set as: the first voxel data is 1 and the second voxel data is greater than 2;

[0105] Step 4: If the root connection data of the current root voxel grid is 1 and the root connection data of the adjacent root voxel grid is greater than 2, the current root voxel grid is judged to be a burr. Through iteration, the burrs in the initial root skeleton map are filtered out to obtain the root skeleton map.

[0106] The root point cloud skeleton extraction method of the present invention can eliminate the influence of smear, discrete noise points and burrs in the plant root point cloud on skeleton extraction, thereby improving the accuracy of the root skeleton map.

[0107] Example 2:

[0108] A root point cloud skeleton extraction system, such as Figure 2 As shown, it includes a point cloud segmentation module 100, a voxel map construction module 200, a noise filtering module 300 and a burr filtering module 400;

[0109] The point cloud segmentation module 100 obtains a plant root point cloud and performs point cloud segmentation to form a root voxel grid set;

[0110] The voxel map construction module 200 constructs a root voxel map, wherein the edges of the root voxel map are obtained by connecting the spatial positions of the current root voxel grid and the spatial positions of adjacent root voxel grids, and the weights of the root voxel map are obtained by constructing a point cloud search space and obtaining the point cloud density in the point cloud search space;

[0111] The noise filtering module 300 constructs a root voxel spanning tree based on the minimum spanning tree algorithm according to the weights of the root voxel map and the root voxel map, and then determines whether there is noise in the root voxel map, and filters the edges corresponding to the noise to obtain an initial root skeleton map;

[0112] The burr filtering module 400 obtains the connection relationship between the current root voxel grid and the adjacent root voxel grid in the initial root skeleton map, obtains root connection data, and determines whether there are burrs in the initial root skeleton map, filters the root voxel grids corresponding to the burrs, and obtains the root skeleton map.

[0113] Various changes and modifications can be made without departing from the spirit and scope of the present invention, and all equivalent technical solutions also fall within the scope of the present invention.

[0114] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments, and the same or similar parts between the various embodiments can be referenced to each other.

[0115] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, apparatus, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0116] The present invention is described with reference to the flowcharts and / or block diagrams of the method, terminal device (system), and computer program product according to the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the process in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0117] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0118] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0119] It should be noted that:

[0120] References in this specification to "one embodiment" or "an embodiment" mean that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the present invention. Therefore, appearances of the phrases "one embodiment" or "an embodiment" in various places throughout this specification do not necessarily refer to the same embodiment.

[0121] Furthermore, it should be noted that the specific embodiments described in this specification may vary in the shapes and names of their components. Any equivalent or simple variations based on the structure, features, and principles described in the patented concept of this invention are included within the scope of protection of this patent. Persons skilled in the art may make various modifications, additions, or substitutions to the described specific embodiments, and these modifications, as long as they do not deviate from the structure of the invention or exceed the scope defined by the claims, shall fall within the scope of protection of this invention.

Claims

1. A root point cloud skeleton extraction method, characterized in that: The following steps are involved: Obtain plant root point cloud and perform point cloud segmentation to form a root voxel grid set; Constructing a root voxel map, wherein the edges of the root voxel map are obtained by connecting the spatial positions of the current root voxel grid and the spatial positions of the adjacent root voxel grids, and the weights of the root voxel map are obtained by constructing a point cloud search space and obtaining the point cloud density in the point cloud search space; The spatial position of the root voxel grid is used as the sphere center and the preset grid threshold is used as the diameter to construct the point cloud search area. Obtain the number of plant root point clouds in the point cloud search area and perform density calculation to obtain the root point cloud density; Normalize the root point cloud density to obtain the normalized point cloud density; The weight of the root voxel map is obtained by calculation based on the adjacent normalized point cloud density; Among them, the normalized point cloud density and the weight of the root voxel map are expressed as follows: in, , Indicates the maximum value of the root point cloud density, represents the density of root point cloud, Indicates the number of root voxel grids, Indicates the Normalized point cloud density, Indicates the A root voxel grid, Indicates the The root voxel grid and the The weight of the root voxel map between the root voxel grids, Indicates the Normalized point cloud density; According to the weight of the root voxel map and the root voxel map, a root voxel spanning tree is constructed based on the minimum spanning tree algorithm to determine whether there is noise in the root voxel map, and the edges corresponding to the noise are filtered to obtain the initial root skeleton map; The connection relationship between the current root voxel grid and the adjacent root voxel grid in the initial root skeleton map is obtained to obtain the root connection data. It is also determined whether there are burrs in the initial root skeleton map. The root voxel grids corresponding to the burrs are filtered to obtain the root skeleton map.

2. The root point cloud skeleton extraction method according to claim 1, characterized in that: The step of obtaining a plant root point cloud and dividing the point cloud to form a root voxel grid set includes the following steps: Obtaining a plant root point cloud and determining a three-dimensional spatial boundary of the plant root point cloud, wherein the three-dimensional spatial boundary includes a horizontal boundary, a longitudinal boundary, and a vertical boundary; The plant root point cloud within the three-dimensional space boundary is divided according to the preset grid threshold to obtain the initial root voxel grid set; Obtain the number of point clouds contained in all initial root voxel grids in the initial root voxel grid set to obtain a point cloud quantity set, set a point cloud quantity threshold, obtain the number of point clouds in the point cloud quantity set that do not meet the point cloud quantity threshold, and filter the corresponding initial root voxel grids to obtain a filtered voxel grid set; In the filtering voxel grid set, if the filtering voxel grid set still has connectivity after deleting the current filtering voxel grid, the current filtering voxel grid is deleted, and the root voxel grid set is obtained by iterative deletion.

3. The root point cloud skeleton extraction method according to claim 1, characterized in that: The step of connecting the spatial position of the current root voxel grid and the spatial position of the adjacent root voxel grid to obtain the edge of the root voxel graph includes the following steps: The spatial position of the root voxel grid is obtained through the center position of the root voxel grid; Preset a search threshold, construct a search space of the current root voxel grid based on the search threshold, and determine whether there are adjacent root voxel grids in the search space; If it exists, the spatial position of the current root voxel grid and the spatial position of the adjacent root voxel grid are connected to obtain the edge of the root voxel graph.

4. The root point cloud skeleton extraction method according to claim 1, characterized in that: The method comprises the following steps: constructing a root voxel spanning tree based on the minimum spanning tree algorithm according to the weight of the root voxel map and the root voxel map, and then determining whether there is noise in the root voxel map, and filtering the edges corresponding to the noise to obtain an initial root skeleton map. Sort the root voxel map according to its weight to obtain a sorted voxel map; Initialize the root spanning set, determine whether the root voxel grid corresponding to the edge of the current sorted voxel graph belongs to the root spanning set, if not, add the edge of the current sorted voxel graph to the root spanning set; If yes, filter the edges of the current sorted voxel graph and traverse the sorted voxel graph until the edge of the root spanning set can connect all the root voxel grids in the sorted voxel graph, and the sum of the weights of the root voxel spanning set is the smallest, and obtain the root voxel spanning tree; The edges in the root voxel map that do not belong to the root voxel spanning tree are filtered out to obtain the initial root skeleton map.

5. The root point cloud skeleton extraction method according to claim 1, characterized in that: The method of obtaining the connection relationship between the current root voxel grid and the adjacent root voxel grids in the initial root system skeleton image to obtain the root system connection data includes the following steps: Initialize the root connection data of the current root voxel grid in the initial root skeleton map to 0; The edges of the initial root skeleton graph are traversed to obtain the number of edges of the initial root skeleton graph connected to the current root voxel grid, and the root connection data corresponding to all root voxel grids are obtained.

6. The root point cloud skeleton extraction method according to claim 1, characterized in that: The method of determining whether there are burrs in the initial root system skeleton map and filtering the root system voxel grids corresponding to the burrs to obtain the root system skeleton map includes the following steps: Obtain root connection data of the current root voxel grid in the initial root skeleton image to obtain first voxel data, and obtain root connection data of the adjacent root voxel grid to obtain second voxel data; Based on the first voxel data and the second voxel data, determining whether the root voxel grid meets the burr condition, and if so, determining it as a burr, wherein the burr condition is at least: the first voxel data is 1 and the second voxel data is greater than 2; The initial root skeleton graph is traversed and the root voxel grid corresponding to the burrs is filtered to obtain the root skeleton graph.

7. A root point cloud skeleton extraction system, characterized in that: Including point cloud segmentation module, voxel map construction module, noise filtering module and burr filtering module; The point cloud segmentation module obtains the plant root point cloud and performs point cloud segmentation to form a root voxel grid set; The voxel map construction module constructs a root voxel map, wherein the edges of the root voxel map are obtained by connecting the spatial positions of the current root voxel grid and the spatial positions of the adjacent root voxel grids, and the weights of the root voxel map are obtained by constructing a point cloud search space and obtaining the point cloud density in the point cloud search space; The spatial position of the root voxel grid is used as the sphere center and the preset grid threshold is used as the diameter to construct the point cloud search area. Obtain the number of plant root point clouds in the point cloud search area and perform density calculation to obtain the root point cloud density; Normalize the root point cloud density to obtain the normalized point cloud density; The weight of the root voxel map is obtained by calculation based on the adjacent normalized point cloud density; Among them, the normalized point cloud density and the weight of the root voxel map are expressed as follows: in, , Indicates the maximum value of the root point cloud density, represents the density of root point cloud, Indicates the number of root voxel grids, Indicates the Normalized point cloud density, Indicates the A root voxel grid, Indicates the The root voxel grid and the The weight of the root voxel map between the root voxel grids, Indicates the Normalized point cloud density; The noise filtering module constructs a root voxel spanning tree based on the minimum spanning tree algorithm according to the weight of the root voxel map and the root voxel map, and then determines whether there is noise in the root voxel map, and filters the edges corresponding to the noise to obtain an initial root skeleton map; The burr filtering module obtains the connection relationship between the current root voxel grid and the adjacent root voxel grid in the initial root skeleton map, obtains root connection data, and determines whether there are burrs in the initial root skeleton map, filters the root voxel grids corresponding to the burrs, and obtains the root skeleton map.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

9. A root system point cloud skeleton extraction device, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 6 is implemented.

Images