Plant leaf inclination angle measurement method and device

Abstract

The present invention provides a method and device for measuring the inclination angle of plant leaves. The method for measuring the inclination angle of plant leaves includes: obtaining a plant point cloud image of a target plant; clustering and segmenting the plant point cloud image to obtain a plurality of leaf point cloud images; performing surface reconstruction on any leaf point cloud image to determine the actual area of ​​any leaf point cloud image, and performing plane projection on any leaf point cloud image to determine the projection area of ​​any leaf point cloud image; determining the inclination angle of any leaf point cloud image based on the actual area and the projection area; and determining the plant leaf inclination angle of the target plant based on the inclination angles of the plurality of leaf point cloud images. The method for measuring the inclination angle of plant leaves provided by the present invention obtains a plurality of leaf point cloud images, and obtains the inclination angle of a single leaf based on the actual area and the projection area of ​​the leaf point cloud images, thereby obtaining the plant leaf inclination angle, which can avoid causing damage to the plant, and can improve the accuracy of measuring the plant leaf inclination angle, and improve the efficiency of measuring the plant leaf inclination angle.

Description

Technical Field

[0001] The present invention relates to the field of computer vision technology, and in particular to a method and device for measuring the inclination angle of plant leaves. Background Art

[0002] Leaf inclination is an important parameter for describing the structure of the vegetation canopy. It determines the amount of radiation intercepted by the vegetation canopy and affects the photosynthesis rate of the entire plant. In addition, leaf inclination can be used as an indicator of plant growth status, such as water conditions. Therefore, measuring the phenotypic parameter of leaf inclination is crucial to improving plant yield.

[0003] At present, in the technology related to measuring the inclination angle of plant leaves, the inclination angle of plant leaves is mainly measured on the spot through inclinometers, protractors and compasses. This kind of field measurement may cause damage to the plants and relies on human subjectivity, and the accuracy cannot be guaranteed. This is time-consuming and labor-intensive, and the measurement efficiency is low. Summary of the invention

[0004] The present invention provides a plant leaf inclination angle measurement method and device, which are used to solve the defects in the prior art that field measurement may cause damage to plants, rely on human subjectivity, cannot guarantee accuracy, is time-consuming and labor-intensive, and has low measurement efficiency. The method avoids damage to plants and can improve the accuracy of measuring the plant leaf inclination angle and improve the efficiency of measuring the plant leaf inclination angle.

[0005] The present invention provides a method for measuring the inclination angle of plant leaves, which comprises: obtaining a plant point cloud image of a target plant; clustering and segmenting the plant point cloud image to obtain a plurality of leaf point cloud images; performing surface reconstruction on any of the leaf point cloud images to determine the actual area of ​​any of the leaf point cloud images, and performing plane projection on any of the leaf point cloud images to determine the projection area of ​​any of the leaf point cloud images; determining the inclination angle of any of the leaf point cloud images based on the actual area and the projection area; and determining the plant leaf inclination angle of the target plant based on the inclination angles of the plurality of leaf point cloud images.

[0006] According to a plant leaf inclination angle measurement method provided by the present invention, obtaining a plant point cloud image of a target plant includes: obtaining an original three-dimensional point cloud image of the target plant; and preprocessing the original three-dimensional point cloud image to obtain the plant point cloud image.

[0007] According to a plant leaf inclination angle measurement method provided by the present invention, the original three-dimensional point cloud image is preprocessed to obtain the plant point cloud image, including: based on the original three-dimensional point cloud image, a rectangular parallelepiped model is generated, the rectangular parallelepiped model covers the three-dimensional boundary points of the original three-dimensional point cloud image, and any face of the rectangular parallelepiped model is parallel to at least one coordinate axis of the coordinate system of the original three-dimensional point cloud image; based on a preset scale feature, the rectangular parallelepiped model is divided into multiple grids; based on a preset range threshold, point cloud data whose distance from the center of the grid exceeds the range threshold is deleted to obtain the plant point cloud image.

[0008] According to a plant leaf inclination angle measurement method provided by the present invention, the original three-dimensional point cloud image is preprocessed to obtain the plant point cloud image, including: based on the depth value of the original three-dimensional point cloud image and a preset depth threshold, deleting noise point cloud data to obtain the plant point cloud image.

[0009] According to a plant leaf inclination angle measurement method provided by the present invention, the original three-dimensional point cloud image is preprocessed to obtain the plant point cloud image, including: selecting marking points based on statistical analysis of the original three-dimensional point cloud image; deleting noise point cloud data based on the distance between the point cloud data of the original three-dimensional point cloud image and the marking points and a preset distance threshold to obtain the plant point cloud image.

[0010] According to a plant leaf inclination angle measurement method provided by the present invention, the surface reconstruction of any leaf point cloud image is performed to determine the actual area of ​​any leaf point cloud image, including: based on the Delaunay triangulation method, the surface reconstruction of any leaf point cloud image is performed to obtain a plurality of triangular units; based on the formula

[0011]

[0012]

[0013]

[0014] Get the actual area of ​​any of the leaf point cloud images, where S (TLA) is the actual area of ​​any of the blade point cloud images, k is the serial number of the triangle unit, n is the number of triangle units, A is the area of ​​the kth triangle unit, and a, b, and c are the three side lengths of the triangle unit; performing plane projection on any of the blade point cloud images to determine the projection area of ​​any of the blade point cloud images includes: projecting the coordinates of any of the blade point cloud images onto a grid plane, performing plane triangulation on any of the blade point cloud images, and obtaining a plurality of triangle units;

[0015] Based on the formula

[0016]

[0017]

[0018]

[0019] Among them, S (PLA) is the projection area of ​​any of the leaf point cloud images, i is the serial number of the triangular unit, m is the number of triangular units, B is the area of ​​the i-th triangular unit, and x, y, and z are the three side lengths of the triangular unit.

[0020] According to a method for measuring the inclination angle of plant leaves provided by the present invention, the inclination angle of any leaf point cloud image is determined based on the actual area and the projection area, including: taking the ratio of the actual area to the projection area as the inclination angle of any leaf point cloud image.

[0021] The present invention also provides a plant leaf inclination angle measuring device, which includes: an acquisition unit, used to acquire a plant point cloud image of a target plant; a segmentation unit, used to cluster and segment the plant point cloud image to obtain a plurality of leaf point cloud images; a first determination unit, used to perform surface reconstruction on any of the leaf point cloud images, determine the actual area of ​​any of the leaf point cloud images, and perform plane projection on any of the leaf point cloud images to determine the projection area of ​​any of the leaf point cloud images; a second determination unit, used to determine the inclination angle of any of the leaf point cloud images based on the actual area and the projection area; and a third determination unit, used to determine the plant leaf inclination angle of the target plant based on the inclination angles of a plurality of the leaf point cloud images.

[0022] The present invention also 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 steps of any of the above-mentioned plant leaf inclination angle measurement methods when executing the computer program.

[0023] The present invention also 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 steps of any of the above-mentioned methods for measuring the inclination angle of plant leaves are implemented.

[0024] The plant leaf inclination angle measurement method provided by the present invention obtains a three-dimensional plant point cloud image, obtains multiple leaf point cloud images using a clustering segmentation method, and obtains the inclination angle of a single leaf of the plant based on the actual area and projected area of ​​the leaf point cloud image, thereby obtaining the plant leaf inclination angle, which can avoid causing damage to the plant, improve the accuracy of measuring the plant leaf inclination angle, and improve the efficiency of measuring the plant leaf inclination angle. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions in the present invention or 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 some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0026] Figure 1 This is one of the flow charts of the method for measuring the inclination angle of plant leaves provided by the present invention;

[0027] Figure 2 It is a schematic diagram of the collection position of the plant leaf inclination angle measurement method provided by the present invention;

[0028] Figure 3 It is a cluster segmentation schematic diagram of the plant leaf inclination angle measurement method provided by the present invention;

[0029] Figure 4 It is a schematic diagram of obtaining the actual area and the projected area of ​​the plant leaf inclination angle measurement method provided by the present invention;

[0030] Figure 5 It is a schematic diagram of determining the inclination angle of a leaf point cloud image in the method for measuring the inclination angle of a plant leaf provided by the present invention;

[0031] Figure 6 This is the second flow chart of the method for measuring the inclination angle of plant leaves provided by the present invention;

[0032] Figure 7 It is one of the schematic diagrams of the preprocessing process of the plant leaf inclination angle measurement method provided by the present invention;

[0033] Figure 8 This is the second schematic diagram of the preprocessing process of the plant leaf inclination angle measurement method provided by the present invention;

[0034] Fig. 9 This is the third schematic diagram of the preprocessing process of the plant leaf inclination angle measurement method provided by the present invention;

[0035] Fig.10 This is a fourth schematic diagram of the preprocessing process of the plant leaf inclination angle measurement method provided by the present invention;

[0036] Fig.11 This is a fifth schematic diagram of the preprocessing process of the plant leaf inclination angle measurement method provided by the present invention;

[0037] Fig.12 This is one of the statistical graphs of the effect verification test of the plant leaf inclination angle measurement method provided by the present invention;

[0038] Fig.13 This is the second statistical graph of the effect verification test of the plant leaf inclination angle measurement method provided by the present invention;

[0039] Fig.14 This is the third statistical graph of the effect verification test of the plant leaf inclination angle measurement method provided by the present invention;

[0040] Fig.15 It is a structural schematic diagram of the plant leaf inclination angle measuring device provided by the present invention;

[0041] Fig.16 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] Combine the following Figure 1-Figure 16 The invention describes a method and device for measuring the inclination angle of plant leaves.

[0044] The plant leaf inclination angle measurement method provided in the embodiment of the present invention can be applied to the leaf inclination angle detection of various plants, for example, the leaf inclination angle of soybean plants can be measured and calculated.

[0045] For example, the yield of full-canopy pod-bearing crops such as soybeans is closely related to the canopy structure, and leaf inclination is an important parameter that describes the canopy structure of vegetation. It determines the amount of radiation intercepted by the canopy and affects the photosynthesis rate of the entire plant. In addition, leaf inclination can be used as an indicator of plant growth status, such as water conditions. Therefore, studying the phenotypic parameter of plant leaf inclination is crucial to improving plant yield.

[0046] like Figure 1 As shown, an embodiment of the present invention provides a method for measuring the inclination angle of plant leaves, and the method for measuring the inclination angle of plant leaves includes the following steps 110 to 150.

[0047] Wherein, in step 110, a plant point cloud image of the target plant is obtained.

[0048] It should be noted that if Figure 2As shown in the figure, a Kinect 2.0 camera can be used to obtain the plant point cloud image of the target plant. A data acquisition platform is built with the Kinect 2.0 camera as the center. The platform consists of four parts: a Kinect 2.0 camera, a laptop, an adjustable iron frame, and a coil. The Kinect 2.0 camera is placed at a fixed position on the iron frame, and the height of the iron frame is controlled to change the position of the Kinect 2.0 camera to ensure the effectiveness of the information obtained.

[0049] In order to reduce the impact of light on the acquisition of plant point cloud data, the "top view method" can be used to obtain the original three-dimensional point cloud data of soybean plants in the four stages of branching, flowering, pod formation and grain filling. In other words, the Kinect 2.0 camera can be placed on the top of the iron frame, and the Kinect 2.0 camera lens can be used to shoot the plants downward.

[0050] The size of the three-dimensional image of the plant captured by the Kinect 2.0 camera can be 512 pixels × 424 pixels. The depth value of each pixel contains the distance information between the target plant and the Kinect 2.0 camera. The point cloud information represented by each pixel can be obtained by transforming the depth value. The original three-dimensional point cloud data can be pre-processed by simplifying operations and denoising operations to obtain the plant point cloud image of the target plant.

[0051] Step 120: cluster and segment the plant point cloud image to obtain a plurality of leaf point cloud images.

[0052] like Figure 2 As shown, it is understandable that plants, such as soybeans, often have many leaves. The three-dimensional image of the plant taken from above by the Kinect2.0 camera has multiple leaves. Therefore, the plant point cloud image needs to be segmented to obtain multiple separate leaf point cloud images.

[0053] The K-means clustering algorithm can be used to cluster and segment the plant point cloud images.

[0054] K-means is a dynamic clustering algorithm based on partitioning, which uses the sum of squared errors (SSE) as the clustering criterion function.

[0055] The process of clustering segmentation can be:

[0056] First, randomly select f cluster center points from the plant point cloud image.

[0057] Second, the distance between each sample point in the plant point cloud image and f cluster center points is calculated respectively, and these points are classified into the cluster with the nearest cluster center point.

[0058] Third, recalculate the center points of the f clusters based on the existing objects in the cluster. The formula is as follows:

[0059]

[0060] Among them, L j Represents sample point data; represents the average value of sample points; Q j Represents the set of clusters.

[0061] Fourth, repeat the second and third steps above until the cluster center points do not change twice or are less than the set accuracy.

[0062] Fifth, output the results of clustering segmentation.

[0063] When dividing different plants, the t value can be set to the corresponding number of leaves.

[0064] Step 130 , performing surface reconstruction on any blade point cloud image to determine the actual area of ​​any blade point cloud image, and performing plane projection on any blade point cloud image to determine the projection area of ​​any blade point cloud image.

[0065] It can be understood that the actual area of ​​the blade point cloud image is the real area of ​​a surface of the blade in three-dimensional space, and the projection image of the blade point cloud image is the projection area of ​​the blade on a certain plane, such as the ground, a horizontal plane, or a preset XY plane.

[0066] like Figure 4 As shown in the figure, in order to accurately calculate the leaf area due to the irregular shape of the leaf surface, triangulation processing is performed on the basis of the segmentation of a single leaf point cloud image. Taking a single leaf 3D point cloud as the target, the surface reconstruction of the leaf point cloud image is realized using a method based on Delaunay triangulation. The basic principle is to first construct a seed triangle, expand the edges of the triangle to generate new triangles, and repeat the above process until there are no triangle edges that can be expanded.

[0067] The blade point cloud image after surface reconstruction by point cloud triangulation is composed of multiple triangular meshes. This study uses the Helen Keller formula to calculate the area of ​​a single mesh triangle to derive the actual area of ​​the blade point cloud image.

[0068] The process of performing surface reconstruction on any blade point cloud image and determining the actual area of ​​any blade point cloud image may include:

[0069] Based on the Delaunay triangulation method, the surface of any leaf point cloud image is reconstructed to obtain multiple triangle units.

[0070] Based on the formula

[0071]

[0072]

[0073]

[0074] Get the actual area of ​​any leaf point cloud image, where S (TLA) is the actual area of ​​any leaf point cloud image, k is the serial number of the triangle unit, n is the number of triangle units, A is the area of ​​the kth triangle unit, and a, b, and c are the three side lengths of the triangle unit.

[0075] Performing a planar projection on any leaf point cloud image to determine the projection area of ​​any leaf point cloud image may include:

[0076] Project the coordinates of any leaf point cloud image onto a grid plane, perform plane triangulation on any leaf point cloud image, and obtain a plurality of triangle units;

[0077] Based on the formula

[0078]

[0079]

[0080]

[0081] Among them, S (PLA) is the projection area of ​​any leaf point cloud image, i is the serial number of the triangular unit, m is the number of triangular units, B is the area of ​​the i-th triangular unit, and x, y, and z are the three side lengths of the triangular unit.

[0082] Step 140: Determine the inclination angle of any blade point cloud image based on the actual area and the projected area.

[0083] The actual area and the projected area can be divided or subtracted, or trigonometric functions can be used to calculate the inclination angle of the leaf point cloud image.

[0084] like Figure 5 As shown, the ratio of the actual area to the projected area can be used as the inclination angle of any blade point cloud image.

[0085] It should be noted that the leaf inclination angle is the angle θ between the normal line L of the leaf ventral surface and the zenith axis (z axis) L The leaf inclination angle takes the z-axis as 0 degrees. In other words, the leaf inclination angle is also the angle between the leaf surface and the ground plane, which can be calculated based on the formula

[0086]

[0087] Calculate the inclination angle θ of the blade point cloud imageL .

[0088] Step 150: Determine the inclination angle of the leaves of the target plant based on the inclination angles of the plurality of leaf point cloud images.

[0089] It should be noted that the inclination angles of the multiple leaf point cloud images may be averaged, and the average of the inclination angles of the multiple leaf point cloud images may be used as the plant leaf inclination angle of the target plant.

[0090] Alternatively, a weighted average of the inclination angles of multiple leaf point cloud images can be taken, the actual area of ​​the leaf point cloud image can be used as a weight to weight the leaf point cloud image, and the weighted average result of the inclination angles of multiple leaf point cloud images can be used as the plant leaf inclination angle of the target plant.

[0091] The plant leaf inclination angle measurement method provided by the embodiment of the present invention obtains a three-dimensional plant point cloud image, obtains multiple leaf point cloud images using a clustering segmentation method, and calculates the inclination angle of a single leaf of the plant based on the actual area and projection area of ​​the leaf point cloud image, thereby obtaining the plant leaf inclination angle. This can avoid causing damage to the plant, and can improve the accuracy of measuring the plant leaf inclination angle, and improve the efficiency of measuring the plant leaf inclination angle.

[0092] like Figure 6 As shown, in some embodiments, the above step 110, obtaining the plant point cloud image of the target plant, includes the following steps 111-112.

[0093] Wherein, step 111 is to obtain an original three-dimensional point cloud image of the target plant.

[0094] It can be understood that a Kinect 2.0 camera can be used to obtain a plant point cloud image of the target plant. The size of the three-dimensional image of the plant captured by the Kinect 2.0 camera can be 512 pixels × 424 pixels. The depth value of each pixel contains the distance information between the target plant and the Kinect 2.0 camera. The point cloud information represented by each pixel can be obtained by transforming the depth value to obtain the original three-dimensional point cloud image of the target plant.

[0095] Step 112: pre-process the original three-dimensional point cloud image to obtain a plant point cloud image.

[0096] It is understandable that the original three-dimensional point cloud data can be pre-processed by using methods such as simplification operations and denoising operations to obtain a plant point cloud image of the target plant.

[0097] By preprocessing the original three-dimensional point cloud data through simplification and denoising operations, a point cloud image of the plant with prominent leaf features can be obtained, other irrelevant features are excluded, and the efficiency of leaf inclination angle measurement can be improved.

[0098] like Figure 7 and Figure 8 As shown, in some embodiments, the above step 112, preprocessing the original three-dimensional point cloud image to obtain the plant point cloud image, includes:

[0099] Based on the original 3D point cloud image, a cuboid model is generated, the cuboid model covers the 3D boundary points of the original 3D point cloud image, and any face of the cuboid model is parallel to at least one coordinate axis of the coordinate system of the original 3D point cloud image.

[0100] It is understandable that the original 3D point cloud image can be simplified and read to build a cuboid model that surrounds the entire original 3D point cloud image and has three sides parallel to the three coordinate axes of the coordinate system of the original 3D point cloud image.

[0101] The three sides of the cuboid model are x 、b y and c z Based on the formula

[0102]

[0103] Indicates that, a max and a min represents the coordinates of the two boundary points of the original 3D point cloud image along the x-axis, b max and b min Represents the coordinates of the two boundary points of the original 3D point cloud image along the y-axis, c max and c min Represents the coordinates of the two boundary points of the original 3D point cloud image along the z-axis.

[0104] Based on the preset scale characteristics, the rectangular model is divided into multiple grids.

[0105] It is understandable that the rectangular model is divided into multiple grids of the same shape and size. The side length of the grid can be adjusted. The larger the side length, the smaller the density of the simplified point cloud. This can adjust the effect of the simplification operation.

[0106] The side length d of the grid is based on the formula

[0107]

[0108] Determine, where β represents a preset scale feature, and t represents the total number of point clouds in the original three-dimensional point cloud image.

[0109] Based on the preset range threshold, the point cloud data whose distance from the center of the grid exceeds the range threshold is deleted to obtain the plant point cloud image.

[0110] It can be understood that the preset range threshold, with the center of the grid as the base point, deletes the point cloud data exceeding the range threshold, and retains the point cloud data within the range threshold, thereby simplifying the original three-dimensional point cloud image and achieving a streamlining effect.

[0111] It should be noted that the original 3D point cloud image may have noise, which is generally divided into two categories: subjective noise and objective noise. Subjective noise is caused by expanding the scanning range of the object being measured. Objective noise is caused by the equipment itself and the external environment. Different noise types require different denoising methods.

[0112] like Fig. 9 As shown, in some embodiments, the above step 112, preprocessing the original three-dimensional point cloud image to obtain a plant point cloud image, includes: based on the depth value of the original three-dimensional point cloud image and a preset depth threshold, deleting noise point cloud data to obtain a plant point cloud image.

[0113] It is understandable that this is a depth threshold filtering method used to remove subjective noise. The subjective noise caused by expanding the scanning range of the object under test is usually called background point cloud. In order to reduce the size of the point cloud file to achieve fast and efficient processing of effective point cloud data, it is necessary to segment the background point cloud of the object, that is, select the rectangular area containing the reconstructed object for segmentation. The depth threshold filtering method is used to select the depth effective area, that is, the point cloud that does not exceed the preset depth threshold is considered to be valid data, and the point cloud that exceeds the depth threshold, that is, the subjective noise point cloud, is directly deleted.

[0114] like Fig.10 As shown, in some embodiments, the above step 112, pre-processing the original three-dimensional point cloud image to obtain a plant point cloud image, includes: selecting marker points based on statistical analysis of the original three-dimensional point cloud image; deleting noise point cloud data based on the distance between the point cloud data of the original three-dimensional point cloud image and the marker points and a preset distance threshold to obtain a plant point cloud image.

[0115] It is understandable that the objective noise problem caused by the device itself and the external environment can be processed by statistical filtering. The principle is to first count all the points in the original 3D point cloud data, select one of them as a marker point, calculate the average distance between the marker point and the surrounding adjacent points, and delete the points outside the preset distance threshold.

[0116] Assume that the original 3D point cloud image has g points, one of which is marked as P g (X g , Y g , Z g), g = 1, 2, 3 ... h, where h is the total number of point clouds, and the average distance du between the marked point and the surrounding adjacent points can be based on the formula

[0117]

[0118]

[0119] Indicates, where r represents the number of adjacent points around the marked point Pg, It is expressed as the average distance between the marked point Pg and its neighboring points.

[0120] The distribution of multiple critical points can be represented by Gaussian distribution, with a mean of μ and a standard deviation of σ. The standard deviation multiple threshold std can be introduced as the distance threshold. If the average distance corresponding to the marked point Pg is When , it is a valid point, otherwise it is an outlier point, and the outlier point is removed from the original 3D point cloud image.

[0121] like Fig.11 As shown, in some embodiments, the original three-dimensional point cloud image of a single-pot soybean plant in each period is obtained vertically using a constructed data acquisition platform under natural lighting conditions, and is preprocessed. In order to illustrate the effect of preprocessing, a single-pot soybean plant is taken as an example, and the changes in the point cloud images of the soybean plant in four periods of branching, flowering, podding and grain filling over time are shown. After preprocessing, the canopy characteristics can be clearly observed, the soybean leaf structure can be clearly observed, and high-precision three-dimensional reconstruction can be achieved. It can be seen that the preprocessing of the original three-dimensional point cloud image of soybeans can provide a solid preliminary foundation for the accuracy of soybean leaf inclination calculation.

[0122] For the soybean sample plants in a single pot at the branching stage, the number of point clouds in the original three-dimensional point cloud image is 180340, the number of point clouds after simplification is 126238, and the number of point clouds after denoising is 14663; for the soybean sample plants in a single pot at the flowering stage, the number of point clouds in the original three-dimensional point cloud image is 170714, the number of point clouds after simplification is 116085, and the number of point clouds after denoising is 13248; for the soybean sample plants in a single pot at the clamping stage, the number of point clouds in the original three-dimensional point cloud image is 182217, the number of point clouds after simplification is 127551, and the number of point clouds after denoising is 15564; for the soybean sample plants in a single pot at the grain filling stage, the number of point clouds in the original three-dimensional point cloud image is 185246, the number of point clouds after simplification is 127819, and the number of point clouds after denoising is 16598. This method is also applicable to the calculation of leaf inclination angles of soybean groups.

[0123] like Fig.12 , Fig.13 and Fig.14As shown in the figure, in order to evaluate the experimental results, 10 representative potted soybean plants of three varieties, Fudou No. 6, Kangxian No. 9 and Kangxian No. 13, were selected respectively in the vertical collection mode, and the leaf inclination parameters of the four reproductive growth stages of branching, flowering, pod setting and grain filling were calculated by the method proposed in the embodiment of the present invention for random combination. A total of 140 data were measured, and the calculated values ​​were compared with the measured values. The average relative error of the plant leaf inclination angle measurement method provided by the embodiment of the present invention was 3.58%, and the determination coefficient R 2 They are 0.8317, 0.9075, and 0.9186 respectively. The experimental results prove the effectiveness of the plant leaf inclination angle measurement method provided by the embodiment of the present invention.

[0124] The plant leaf inclination angle measuring device provided by the present invention is described below. The plant leaf inclination angle measuring device described below and the plant leaf inclination angle measuring method described above can be referenced to each other.

[0125] like Fig.15 The embodiment of the present invention provides a plant leaf inclination angle measuring device, including: an acquisition unit 151 , a segmentation unit 152 , a first determination unit 153 , a second determination unit 154 and a third determination unit 155 .

[0126] The acquisition unit 151 is used to acquire a plant point cloud image of a target plant;

[0127] The segmentation unit 152 is used to perform cluster segmentation on the plant point cloud image to obtain a plurality of leaf point cloud images;

[0128] The first determination unit 153 is used to perform surface reconstruction on any blade point cloud image to determine the actual area of ​​any blade point cloud image, and perform plane projection on any blade point cloud image to determine the projection area of ​​any blade point cloud image;

[0129] A second determining unit 154 is used to determine the inclination angle of any blade point cloud image based on the actual area and the projected area;

[0130] The third determining unit 155 is used to determine the inclination angle of the plant leaves of the target plant based on the inclination angles of the multiple leaf point cloud images.

[0131] The plant leaf inclination angle measuring device provided in the embodiment of the present invention is used to execute the plant leaf inclination angle measuring method described in the above embodiment, which will not be described in detail in this embodiment.

[0132] Fig.16 An example of a physical structure diagram of an electronic device is shown in FIG. Fig.16As shown, the electronic device may include: a processor 161, a communication interface 162, a memory 163 and a communication bus 164, wherein the processor 161, the communication interface 162 and the memory 163 communicate with each other through the communication bus 164. The processor 161 may call the logic instructions in the memory 163 to execute a plant leaf inclination angle measurement method, which includes: obtaining a plant point cloud image of a target plant; clustering and segmenting the plant point cloud image to obtain a plurality of leaf point cloud images; performing surface reconstruction on any leaf point cloud image to determine the actual area of ​​any leaf point cloud image, and performing plane projection on any leaf point cloud image to determine the projection area of ​​any leaf point cloud image; determining the inclination angle of any leaf point cloud image based on the actual area and the projection area; determining the plant leaf inclination angle of the target plant based on the inclination angles of the plurality of leaf point cloud images.

[0133] In addition, the logic instructions in the above-mentioned memory 163 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. 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 and other media that can store program codes.

[0134] On the other hand, the present invention also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the plant leaf inclination angle measurement method provided by the above-mentioned methods, and the plant leaf inclination angle measurement method includes: obtaining a plant point cloud image of a target plant; clustering and segmenting the plant point cloud image to obtain multiple leaf point cloud images; performing surface reconstruction on any leaf point cloud image to determine the actual area of ​​any leaf point cloud image, and performing plane projection on any leaf point cloud image to determine the projection area of ​​any leaf point cloud image; determining the inclination angle of any leaf point cloud image based on the actual area and the projection area; and determining the plant leaf inclination angle of the target plant based on the inclination angles of multiple leaf point cloud images.

[0135] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the above-mentioned plant leaf inclination angle measurement methods, the plant leaf inclination angle measurement methods comprising: obtaining a plant point cloud image of a target plant; clustering and segmenting the plant point cloud image to obtain a plurality of leaf point cloud images; performing surface reconstruction on any leaf point cloud image to determine the actual area of ​​any leaf point cloud image, and performing plane projection on any leaf point cloud image to determine the projection area of ​​any leaf point cloud image; determining the inclination angle of any leaf point cloud image based on the actual area and the projection area; and determining the plant leaf inclination angle of the target plant based on the inclination angles of multiple leaf point cloud images.

[0136] 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. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0137] 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, it 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.

[0138] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention 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 invention.

Claims

1. A method for measuring the inclination angle of plant leaves, characterized in that: include: Acquire a plant point cloud image of the target plant; Performing clustering segmentation on the plant point cloud image to obtain a plurality of leaf point cloud images; Performing surface reconstruction on any of the blade point cloud images to determine the actual area of ​​any of the blade point cloud images, and performing plane projection on any of the blade point cloud images to determine the projection area of ​​any of the blade point cloud images; Determine the inclination angle of any of the blade point cloud images based on the actual area and the projected area; Determining the inclination angle of a plant leaf of a target plant based on the inclination angles of the plurality of leaf point cloud images; Wherein, determining the inclination angle of the plant leaf of the target plant based on the inclination angles of the plurality of leaf point cloud images comprises: performing weighted average processing on the inclination angles of the plurality of leaf point cloud images to obtain the inclination angle of the plant leaf of the target plant; the weight of each leaf point cloud image is the actual area of ​​each leaf point cloud image; The step of obtaining a plant point cloud image of a target plant comprises: Obtaining the original three-dimensional point cloud image of the target plant; Preprocessing the original three-dimensional point cloud image to obtain the plant point cloud image; The preprocessing of the original three-dimensional point cloud image to obtain the plant point cloud image includes: Selecting marker points based on statistical analysis of the original three-dimensional point cloud image; Based on the distance between the point cloud data of the original three-dimensional point cloud image and the marking point and a preset distance threshold, the noise point cloud data is deleted to obtain the plant point cloud image; The method of deleting noise point cloud data based on the distance between the point cloud data of the original three-dimensional point cloud image and the marking point and a preset distance threshold to obtain the plant point cloud image includes: Assume that the original 3D point cloud image has g points, one of which is marked as Pg(Xg, Yg, Zg), g = 1, 2, 3...h, where h is the total number of point clouds. The average distance between the marked point and the surrounding adjacent points is calculated based on the following formula: Among them, (Xn, Yn, Zn) is the nth neighboring point of the marker point Pg, d u represents the distance between the marked point Pg and the nth adjacent point, r represents the number of adjacent points around the marked point Pg, It is expressed as the average distance between the marked point Pg and the adjacent points; If the average distance corresponding to the marked point Pg Then the marked point Pg is a valid point, otherwise the marked point Pg is an outlier, std is the standard deviation multiple threshold, μ is the mean of the Gaussian distribution of the surrounding adjacent points, and σ is the standard deviation; The outlier points in the original three-dimensional point cloud image are deleted to obtain the plant point cloud image.

2. The method for measuring the inclination angle of plant leaves according to claim 1, characterized in that: The preprocessing of the original three-dimensional point cloud image to obtain the plant point cloud image includes: Based on the original three-dimensional point cloud image, generate a cuboid model, wherein the cuboid model covers the three-dimensional boundary points of the original three-dimensional point cloud image, and any face of the cuboid model is parallel to at least one coordinate axis of the coordinate system of the original three-dimensional point cloud image; Based on a preset scale feature, the cuboid model is divided into a plurality of grids; Based on a preset range threshold, point cloud data whose distance from the center of the grid exceeds the range threshold is deleted to obtain the plant point cloud image.

3. The method for measuring the inclination angle of plant leaves according to claim 1, characterized in that: The preprocessing of the original three-dimensional point cloud image to obtain the plant point cloud image includes: Based on the depth value of the original three-dimensional point cloud image and a preset depth threshold, noise point cloud data is deleted to obtain the plant point cloud image.

4. The method for measuring the inclination angle of plant leaves according to any one of claims 1 to 3, characterized in that: The performing surface reconstruction on any of the blade point cloud images to determine the actual area of ​​any of the blade point cloud images comprises: Based on the Delaunay triangulation method, a surface reconstruction is performed on any of the blade point cloud images to obtain a plurality of triangular units; Based on the formula Get the actual area of ​​any of the leaf point cloud images, where S (TLA) is the actual area of ​​any of the leaf point cloud images, k is the serial number of the triangle unit, n is the number of triangle units, A is the area of ​​the kth triangle unit, and a, b, c are the three side lengths of the triangle unit; The performing planar projection on any of the blade point cloud images to determine the projection area of ​​any of the blade point cloud images comprises: Projecting the coordinates of any of the blade point cloud images onto a grid plane, performing plane triangulation on any of the blade point cloud images to obtain a plurality of triangle units; Based on the formula Among them, S (PLA) is the projection area of ​​any of the leaf point cloud images, i is the serial number of the triangular unit, m is the number of triangular units, B is the area of ​​the i-th triangular unit, and x, y, and z are the three side lengths of the triangular unit.

5. The method for measuring the inclination angle of plant leaves according to any one of claims 1 to 3, characterized in that: The determining the inclination angle of any one of the blade point cloud images based on the actual area and the projection area includes: taking the ratio of the actual area to the projection area as the inclination angle of any one of the blade point cloud images.

6. A plant leaf inclination angle measuring device, characterized in that: The method for measuring the inclination angle of plant leaves as claimed in claim 1 comprises: An acquisition unit, used for acquiring a plant point cloud image of a target plant; A segmentation unit, used for clustering and segmenting the plant point cloud image to obtain a plurality of leaf point cloud images; A first determination unit is used to perform surface reconstruction on any of the blade point cloud images to determine the actual area of ​​any of the blade point cloud images, and to perform plane projection on any of the blade point cloud images to determine the projection area of ​​any of the blade point cloud images; A second determining unit, configured to determine an inclination angle of any of the blade point cloud images based on the actual area and the projection area; The third determination unit is used to determine the inclination angle of the plant leaves of the target plant based on the inclination angles of the plurality of leaf point cloud images.

7. 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 program, the steps of the plant leaf inclination angle measurement method according to any one of claims 1 to 5 are implemented.

8. 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 steps of the plant leaf inclination angle measurement method according to any one of claims 1 to 5 are implemented.

Images