Corn variety testing methods, apparatuses, systems, devices, and media

Abstract

The application provides a corn variety testing method, device, system, equipment and medium, and relates to the field of crop variety testing.The method comprises the following steps: determining all tassel point clouds of a target corn according to multi-layer corn tassel multi-view images, the multi-layer corn tassel multi-view images being identified target corn electronic tags, and the multi-layer corn tassel multi-view images corresponding to the target corn being obtained; inputting the tassel point clouds into a point cloud segmentation model to obtain tassel branch tip point clouds output by the point cloud segmentation model; determining tassel main axis point clouds and tassel branch point clouds according to the tassel branch tip point clouds and the tassel point clouds; determining corn tassel phenotype index data according to the tassel main axis point clouds and the tassel branch point clouds, so as to test corn varieties according to phenotype index data of all tassels of the target corn.The application can realize high-throughput collection of phenotype index data, and improves the fineness and accuracy of testing corn varieties.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of crop variety testing, and in particular to a corn variety testing method, device, system, equipment and medium. BACKGROUND

[0002] At present, based on computer vision technology can only extract three-dimensional distribution indicators of corn tassel, and cannot meet the demand of corn tassel specificity, uniformity and stability (Distinctness, Uniformity, Stability, DUS) indicators. SUMMARY

[0003] The present application provides a corn variety testing method, device, system, equipment and medium, to solve the technical defects of not obtaining each fine index of tassel branch in the prior art, and provides a technical scheme for obtaining tassel length, tassel curvature, tassel volume, tassel height, tassel dispersion, corn tassel short half-axis length, corn tassel long half-axis length, tassel branch number, tassel branch length, tassel branch angle and tassel branch curvature, to improve the precision of corn variety testing and improve the testing efficiency.

[0004] In a first aspect, the present application provides a corn variety testing method, comprising:

[0005] According to the multi-layer corn tassel multi-view image, all tassel point clouds of the target corn are determined, and the multi-layer corn tassel multi-view image is identified by the target corn electronic tag to obtain the corresponding multi-layer corn tassel multi-view image determined by the target corn;

[0006] For any tassel point cloud of the target corn, the tassel point cloud is input into a point cloud segmentation model, and a tassel branch tip point cloud output by the point cloud segmentation model is obtained;

[0007] According to the tassel branch tip point cloud and the tassel point cloud, a tassel main axis point cloud and a tassel branch point cloud are determined;

[0008] According to the tassel main axis point cloud and the tassel branch point cloud, tassel phenotype index data is determined, and the corn variety is tested according to all tassel phenotype index data of the target corn;

[0009] The tassel phenotype index data at least includes tassel length, tassel curvature, tassel volume, tassel height, tassel dispersion, corn tassel short half-axis length, corn tassel long half-axis length, tassel branch number, tassel branch length, tassel branch angle and tassel branch curvature;

[0010] The point cloud segmentation model is determined by training each tassel branch tip point cloud and non-tip point cloud in a plurality of groups of different types of corn variety tassel point cloud samples.

[0011] The tassel branch tip point cloud is all point clouds in a region formed after extending a preset length from a tassel branch tip end to a tassel branch bottom end.

[0012] According to the corn variety test method provided by the application, the all-tassel point cloud of the target corn is determined according to the multi-view image of the multi-layer corn tassel, and the method comprises the following steps:

[0013] The multi-view image of the multi-layer corn tassel is reconstructed by a three-dimensional point cloud, and a three-dimensional point cloud of the multi-layer corn tassel is obtained;

[0014] The fixed column point cloud is removed from the three-dimensional point cloud of the multi-layer corn tassel, and a tassel point cloud group is obtained;

[0015] The tassel point cloud group is processed according to a clustering algorithm, and all the tassel point clouds of the target corn are obtained.

[0016] According to the corn variety test method provided by the application, before the tassel point cloud is input into the point cloud segmentation model, the method further comprises the following steps:

[0017] For any tassel point cloud sample, all point clouds in a region formed after extending a preset length from a tassel branch tip end to a tassel branch bottom end are marked as a first label, and all point clouds in other regions except the region are marked as a second label;

[0018] The tassel branch tip point cloud corresponding to the first label and the non-tip point cloud corresponding to the second label are used to construct a sample set, all sample sets corresponding to all tassel point cloud samples are trained, and a point cloud segmentation model is determined.

[0019] According to the corn variety test method provided by the application, the tassel main axis point cloud and the tassel branch point cloud are determined according to the tassel branch tip point cloud and the tassel point cloud, and the method comprises the following steps:

[0020] The center of gravity of the tassel point cloud is calculated, a surrounding cylinder with a preset length as a radius and the center of gravity as a center is established, and the lowest point of the surrounding cylinder in the vertical direction is determined as the main axis base point of the tassel point cloud;

[0021] All point cloud paths from the main axis base point of the tassel point cloud to each tassel branch tip point cloud are determined.

[0022] All point clouds with the shortest point cloud path are determined as the tassel main axis point cloud, and the point clouds in other point cloud paths except the tassel main axis point cloud are determined as the tassel branch point cloud.

[0023] According to the corn variety test method provided by the application, the tassel phenotype index data is determined according to the tassel main axis point cloud and the tassel branch point cloud, and the method comprises the following steps:

[0024] correcting a direction of the tassel point cloud to a vertical direction according to a correction matrix determined by linear fitting a sequence of skeleton key nodes determined after skeleton contraction of the tassel main axis point cloud;

[0025] determining a tassel length according to a shortest distance from an apex of the tassel branch tip point cloud to a tassel main axis base point;

[0026] determining a tassel curvature according to a tassel height and a tassel length;

[0027] determining a tassel volume according to a convex hull volume formed by the tassel main axis point cloud and the tassel branch point cloud;

[0028] determining a tassel height according to a height difference between an apex of the tassel branch tip point cloud and the tassel main axis point cloud in a vertical direction;

[0029] determining a tassel short semi-axis length according to a length of the tassel main axis point cloud;

[0030] determining a tassel long semi-axis length according to a difference between the tassel length and the tassel short semi-axis length;

[0031] determining a tassel branch length according to a shortest distance between a tassel branch head node and a tassel branch tail node;

[0032] determining a tassel branch angle according to an included angle between the tassel main axis point cloud and the tassel branch point cloud;

[0033] determining a tassel branch curvature according to a straight-line distance between the tassel branch head node and the tassel branch tail node and the tassel branch length;

[0034] determining a tassel dispersion according to a mean value of all tassel branch curvatures.

[0035] In a second aspect, a corn variety testing system is provided, which includes a corn variety testing device of the corn variety testing method as described above;

[0036] Further comprising:

[0037] a field corn tassel sampling device for sampling corn tassels in a corn planting field;

[0038] a multi-layer corn tassel fixing device for fixing the sampled corn tassels and identifying a target corn electronic tag;

[0039] a multi-view image acquisition device for acquiring multi-view images of the multi-layer corn tassels corresponding to the target corn;

[0040] The field corn tassel sampling device of the corn variety testing system comprises a sampling vehicle, and a plurality of sampling plates are fixed on the sampling vehicle;

[0041] For each sampling plate, a plurality of sampling grooves arranged in a matrix are arranged, each sampling groove is provided with a sampling hole, the sampling hole comprises an outer hole, an inner hole and a hole bottom, the outer hole is made of a first material, the inner hole is made of a second material, and an electronic chip for reading and writing an electronic tag is arranged on the hole bottom, the electronic chip is in contact with a radio frequency chip hung on a field corn plant in a working state, and the elasticity of the second material is greater than that of the first material;

[0042] Further comprising a display panel for displaying the sampling results of the corn tassels in the sampling holes and the electronic tags.

[0043] The multi-layer corn tassel sampling device of the corn variety testing system comprises a fixed column and a plurality of sampling support units arranged on the fixed column;

[0044] Each sampling support unit is provided with a sampling branch unit at the end away from the fixed column, the sampling branch unit comprises a sliding groove and an identifier, the sliding groove is provided with a sampling groove, and the identifier is in contact with the sampling hole in a working state to read and write an electronic tag.

[0045] The multi-view image acquisition device of the corn variety testing system comprises a rotating shaft, a rotating arm and an acquisition arm, the rotating shaft is connected with the rotating arm, and the rotating arm is fixedly connected with the acquisition arm;

[0046] The acquisition arm comprises a first acquisition arm and a second acquisition arm, the first acquisition arm is fixedly provided with a first image collector, and the second acquisition arm is fixedly provided with a second image collector and a third image collector.

[0047] In a third aspect, a corn variety testing device is provided, comprising:

[0048] A first determination unit: determining all tassel point clouds of a target corn according to multi-layer corn tassel multi-view images, the multi-layer corn tassel multi-view images being identified target corn electronic tags, and the multi-layer corn tassel multi-view images corresponding to the target corn being acquired;

[0049] An acquisition unit: inputting any tassel point cloud of the target corn into a point cloud segmentation model to acquire a tassel branch tip point cloud output by the point cloud segmentation model;

[0050] A second determination unit: determining a tassel main shaft point cloud and a tassel branch point cloud according to the tassel branch tip point cloud and the tassel point cloud.

[0051] The third determining unit is configured to determine tassel phenotype index data according to the tassel main axis point cloud and the tassel branch point cloud, so as to test the corn variety according to all tassel phenotype index data of the target corn;

[0052] The tassel phenotype index data at least include tassel length, tassel curvature, tassel volume, tassel height, tassel dispersion, corn tassel short semi-axis length, corn tassel long semi-axis length, tassel branch number, tassel branch length, tassel branch angle and tassel branch curvature.

[0053] The point cloud segmentation model is determined according to each tassel branch tip point cloud and non-tip point cloud in a plurality of groups of different kinds of corn variety tassel point cloud samples.

[0054] The tassel branch tip point cloud is all point clouds in a region formed after extending a preset length from a tassel branch tip to a tassel branch bottom end.

[0055] In a fourth aspect, the present application further provides an electronic device, which comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the corn variety testing method when executing the program.

[0056] In a fifth aspect, the present application further provides a non-transitory computer readable storage medium, which stores a computer program, wherein the computer program is executable on a processor to implement the corn variety testing method of any one of the above aspects.

[0057] The present application provides a corn variety testing method, device, system, equipment and medium, by inputting the tassel point cloud to the point cloud segmentation model, obtaining the tassel branch tip point cloud, determining the tassel main axis point cloud and the tassel branch point cloud according to the tassel branch tip point cloud and the tassel point cloud, determining the branch length, the branch angle, the branch curvature and the semi-axis length of the refined phenotype index data according to the tassel main axis point cloud and the tassel branch point cloud, thereby realizing high-throughput collection of phenotype index data, improving the refinement and accuracy of testing corn varieties, and more accurately distinguishing different corn varieties. BRIEF DESCRIPTION OF DRAWINGS

[0058] In order to more clearly illustrate the technical solutions of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0059] Figure 1is one of the flowcharts of the corn variety testing method provided by the present application;

[0060] Figure 2 is the flowchart of determining the all-tassel point cloud of the target corn provided by the present application;

[0061] Figure 3 is the second flowchart of the corn variety testing method provided by the present application;

[0062] Figure 4 is the flowchart of determining the tassel main axis point cloud and the tassel branch point cloud provided by the present application;

[0063] Figure 5 is the structural schematic diagram of the corn variety testing system provided by the present application;

[0064] Figure 6 is the structural schematic diagram of the field corn tassel sampling device provided by the present application;

[0065] Figure 7 is the structural schematic diagram of the multi-layer corn tassel sampling device provided by the present application;

[0066] Figure 8 is the structural schematic diagram of the sampling branch unit provided by the present application;

[0067] Figure 9 is the structural schematic diagram of the multi-view image acquisition device provided by the present application;

[0068] Figure 10 is the structural schematic diagram of the multi-layer corn tassel three-dimensional point cloud provided by the present application;

[0069] Figure 11 is the structural schematic diagram of the tassel point cloud group provided by the present application;

[0070] Figure 12 is the structural schematic diagram of the shortest path of the two sampling branch unit point clouds provided by the present application;

[0071] Figure 13 is the structural schematic diagram of the corn tassel short semi-axis length provided by the present application;

[0072] Figure 14 is the structural schematic diagram of the tassel length provided by the present application;

[0073] Figure 15 is the structural schematic diagram of the tassel height provided by the present application;

[0074] Figure 16 is the structural schematic diagram of the tassel volume provided by the present application;

[0075] Figure 17is a structural schematic diagram of a tassel branching angle provided by the present application;

[0076] Figure 18 is a third flowchart of a corn variety testing method provided by the present application;

[0077] Figure 19 is a structural schematic diagram of a corn variety testing device provided by the present application;

[0078] Figure 20 is a structural schematic diagram of an electronic device provided by the present application. DETAILED DESCRIPTION

[0079] In order to make the objectives, technical solutions, and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below with reference to the drawings in the present application. Obviously, the described embodiments are some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0080] Specificity, consistency, and stability testing is a necessary condition for crop application of new variety protection, variety approval, and registration, and measurement of corn tassel DUS indicators is an important link in corn variety approval. However, due to the factors such as a large number of approved varieties, large sample repetition, strong timeliness of measurement, and high precision requirements of measured indicators, the development of corn tassel DUS indicator measurement is a time-consuming and labor-intensive work, and the precision of DUS indicators is greatly affected by human factors during the measurement process, making it difficult to guarantee the precision of indicators and the traceability of indicator data, and there is an urgent need for digital, automated, and high-throughput technical solutions.

[0081] For the types of corn tassel indicators, the current related technical solutions include manual observation and measurement, auxiliary measurement equipment, and computer vision technology. However, the method used by manual observation and measurement is to count by manual observation, calculate the number of corn tassel branches, calculate the angle of the branches by an angle measuring scale, and measure the length of the branches, the length of the tassel, and the height of the tassel by a scale, which is difficult to meet the requirements of measurement precision and efficiency; the auxiliary measurement equipment has low efficiency and can measure few indicators, and it is difficult to measure three-dimensional distribution and other DUS indicators by auxiliary equipment; the computer vision technology method does not provide a technical solution for branch segmentation, and the automatic segmentation of tassel branches is a key technical link, data collection can only realize data collection of one tassel sample, and the throughput and efficiency of data collection are not high, which limits the technical popularization, and it is difficult to meet the popularization and use of corn tassel DUS indicator measurement.

[0082] The present application aims at the problems of lack of corn tassel DUS phenotype trait automatic software and hardware system, low measurement accuracy and low efficiency of current technical solutions, and provides a corn variety testing method, device, system, equipment and medium, which can realize automatic corn tassel DUS phenotype trait analysis, realize automatic and accurate collection of corn tassel DUS phenotype traits, and provide a convenient tool and system for corn tassel DUS trait investigation, Figure 1 is one of the flowcharts of the corn variety testing method provided by the present application, and provides a corn variety testing method, which comprises the following steps:

[0083] According to the multi-layer corn tassel multi-view image, all tassel point clouds of the target corn are determined, the multi-layer corn tassel multi-view image is identified by an electronic tag of the target corn, and the multi-layer corn tassel multi-view image corresponding to the target corn is obtained.

[0084] For any tassel point cloud of the target corn, the tassel point cloud is input into a point cloud segmentation model, and a tassel branch tip point cloud output by the point cloud segmentation model is obtained.

[0085] According to the tassel branch tip point cloud and the tassel point cloud, a tassel main axis point cloud and a tassel branch point cloud are determined.

[0086] According to the tassel main axis point cloud and the tassel branch point cloud, tassel phenotype index data are determined, and the corn variety is tested according to all tassel phenotype index data of the target corn.

[0087] The tassel phenotype index data at least include tassel length, tassel curvature, tassel volume, tassel height, tassel dispersion, corn tassel short semi-axis length, corn tassel long semi-axis length, tassel branch number, tassel branch length, tassel branch angle and tassel branch curvature.

[0088] The point cloud segmentation model is determined by training each tassel branch tip point cloud and non-tip point cloud in a plurality of groups of different types of tassel point cloud samples of corn varieties.

[0089] The tassel branch tip point cloud is all point clouds in a region formed by extending a preset length from a tassel branch tip to a tassel branch bottom end.

[0090] In step 101, the number identity of the target corn is determined by identifying the electronic tag of the target corn, and then the multi-view image of the target corn is collected to obtain the multi-layer corn tassel multi-view image. Further, all tassel point clouds of the target corn are determined according to the multi-layer corn tassel multi-view image by means of three-dimensional point cloud reconstruction, solid column denoising and clustering algorithm.

[0091] In step 102, for the multi-layered tassel of the target corn collected, which includes multiple tassels, the present application inputs each tassel point cloud of the target corn into a point cloud segmentation model, obtains the tassel branch tip point cloud output by the point cloud segmentation model, and further obtains the tassel branch tip point cloud corresponding to each tassel point cloud. Those skilled in the art understand that the point cloud segmentation model is determined by training each tassel branch tip point cloud and non-tip point cloud in a plurality of different types of tassel point cloud samples of corn varieties. The tassel branch tip point cloud is all point clouds in the region formed after extending a preset length from the tassel branch tip to the tassel branch bottom. The preset length can be 3 cm, that is, 3 cm from the tassel branch tip to the tassel branch bottom as the cutoff point, and all point clouds in the region from the tip to the cutoff point.

[0092] In step 103, after obtaining all tassel branch point clouds in each tassel, the present application needs to determine the tassel main axis point cloud. The present application determines the center of gravity of the tassel point cloud, determines the lowest point of the surrounding cylinder along the vertical direction as the main axis base point of the tassel point cloud, determines all point cloud paths from the main axis base point of the tassel point cloud to each tassel branch tip point cloud, determines all point clouds with the shortest point cloud path as the tassel main axis point cloud, and determines the point clouds in other point cloud paths except the tassel main axis point cloud as the tassel branch point cloud.

[0093] In step 104, after determining the tassel main axis point cloud and the tassel branch point cloud, the tassel phenotype index data can be determined according to the tassel main axis point cloud and the tassel branch point cloud, including 11 index data of tassel length, tassel curvature, tassel volume, tassel height, tassel dispersion, corn tassel short semi-axis length, corn tassel long semi-axis length, tassel branch number, tassel branch length, tassel branch angle, and tassel branch curvature, thereby realizing automatic analysis and high-throughput acquisition of corn phenotype data. Compared with the low automation, low throughput and low measurement efficiency of the prior art, the present application can analyze all branch tips in each tassel of the target corn, thereby comprehensively and finely analyzing the phenotype of the tassel, testing the corn variety according to the phenotype index data of all tassels, and improving the testing ability of the corn variety.

[0094] Optionally, the determination of the tassel phenotype index data according to the tassel main axis point cloud and the tassel branch point cloud includes:

[0095] According to the correction matrix, the direction of the tassel point cloud is corrected to the vertical direction. The correction matrix is determined by linear fitting the skeleton key node sequence determined after skeleton contraction of the tassel main axis point cloud;

[0096] tassel length is determined according to the shortest distance from the highest point of the tassel branch tip point cloud to the tassel main axis base point;

[0097] tassel curvature is determined according to the tassel height and the tassel length;

[0098] tassel volume is determined according to the convex hull volume formed by the tassel main axis point cloud and the tassel branch point cloud;

[0099] tassel height is determined according to the highest point of the tassel branch tip point cloud in the vertical direction and the height difference of the tassel main axis point cloud;

[0100] tassel short semi-axis length is determined according to the length of the tassel main axis point cloud;

[0101] tassel long semi-axis length is determined according to the difference between the tassel length and the tassel short semi-axis length;

[0102] tassel branch length is determined according to the shortest distance between the tassel branch head node and the tassel branch tail node;

[0103] tassel branch angle is determined according to the included angle between the tassel main axis point cloud and the tassel branch point cloud;

[0104] tassel branch curvature is determined according to the straight-line distance between the tassel branch head node and the tassel branch tail node and the tassel branch length;

[0105] tassel dispersion is determined according to the average value of all tassel branch curvatures.

[0106] Optionally, a skeleton contraction algorithm is adopted to perform skeleton contraction on the tassel main axis point cloud MainAxis pc to obtain a contracted skeleton key node sequence {Lplsp i}, then linear fitting is performed on the skeleton key node sequence to calculate a fitting straight line vector LmainAxis, the tassel main axis point cloud MainAxis pc is corrected to the Z-axis direction through the fitting straight line vector LmainAxis, to obtain a correction matrix M ori , the tassel point cloud, i.e., the corn tassel point cloud Tpc ij is corrected to the Z-axis direction through the correction matrix M ori .

[0107] Optionally, Figure 13 is a structural schematic diagram of the corn tassel short semi-axis length provided by the present application, the tassel short semi-axis length is determined according to the length of the tassel main axis point cloud, the tassel main axis point cloud MainAxis pcThe length of the short half axis of the tassel is recorded as SemAxisL.

[0108] Optionally, Figure 14 The structural diagram of the length of the tassel is provided by the present application. The length of the tassel is determined according to the shortest distance from the highest point of the branch tip point cloud of the tassel to the base point of the main axis of the tassel. The length of the tassel point cloud Tpc ij is calculated. The calculation method of the length of the tassel is the shortest path distance from the tip point of the highest branch to the base point Basep of the main axis of the tassel. The length of the tassel is recorded as TasselL.

[0109] Optionally, the tassel curvature is determined according to the height of the tassel and the length of the tassel. The curvature of the tassel is calculated, for example, according to the quotient of the height of the tassel and the length of the tassel. The curvature of the tassel is recorded as TasselW.

[0110] Optionally, Figure 15 The structural diagram of the height of the tassel is provided by the present application. The height of the tassel is determined according to the height difference between the highest point of the branch tip point cloud of the tassel in the vertical direction and the height of the main axis point cloud of the tassel. The height of the tassel point cloud Tpc ij is calculated first. The height difference of the point cloud Tpc ij in the Z-axis direction is determined. The height of the tassel is recorded as TasselH.

[0111] Optionally, Figure 16 The structural diagram of the volume of the tassel is provided by the present application. The volume of the tassel is determined according to the convex hull volume surrounded by the main axis point cloud of the tassel and the branch point cloud of the tassel. The convex hull volume of the tassel point cloud Tpc ij is calculated. The surrounding volume of the tassel is obtained and recorded as TasselV.

[0112] Optionally, the length of the long half axis of the tassel is determined according to the difference between the length of the tassel and the length of the short half axis of the tassel. The difference between the length of the tassel and the length of the short half axis of the tassel TasselL-SemAxisL is the length of the long half axis of the tassel, which is recorded as LongAxisL.

[0113] Optionally, the length of the tassel branch is determined according to the shortest distance between the head node and the tail node of the tassel branch. The head node Start i and the tail node End p of the branch point cloud BranchP p are calculated first. The shortest path from the head node Start p to the tail node End p is established. The length of the shortest path is calculated to obtain the length of the tassel branch, which is recorded as Length i .

[0114] Optionally, Figure 17 is a structural diagram of the tassel branch angle provided by the application, the tassel branch angle is determined according to the included angle of the tassel main axis point cloud and the tassel branch point cloud, the application obtains the tassel branch angle by calculating the included angle of the branch point cloud BranchP i to the tassel main axis point cloud MainAxis pc , and the tassel branch angle is recorded as Angle i .

[0115] Optionally, the tassel branch bending degree is determined according to the straight line distance between the tassel branch head node and the tassel branch tail node and the tassel branch length, as shown in Figure 17 , the straight line distance Segment p between the head node Start p and the tail node End i is calculated, and the bending degree of the branch is Segment i / Length i , recorded as Bend i .

[0116] Optionally, the tassel dispersion degree is determined according to the average of all tassel branch bending degrees, the dispersion degree of the tassel is calculated, and the average of all tassel branch bending degrees is obtained by adding the bending degrees of each tassel branch and dividing by the number of branches, that is, the dispersion degree of the tassel, recorded as TasselDiv.

[0117] The application provides an automatic corn tassel DUS phenotype analysis algorithm, which comprises: multi-layer corn tassel multi-view image point cloud reconstruction, multi-layer multi-corn tassel sample point cloud segmentation, single corn tassel branch point cloud organ segmentation, corn tassel DUS phenotype analysis, and realizes automatic extraction of 11 DUS phenotype indexes including tassel length, tassel bending degree, tassel volume, tassel height, tassel dispersion degree, corn tassel short semi-axis length, corn tassel long semi-axis length, tassel branch number, tassel branch length, tassel branch angle and tassel branch bending degree.

[0118] The application provides a corn variety testing method, device, system, equipment and medium, by inputting the single tassel point cloud to the point cloud segmentation model, obtaining the single tassel branch tip point cloud, determining the single tassel main axis point cloud and the single tassel branch point cloud according to the single tassel branch tip point cloud and the single tassel point cloud; according to the single tassel main axis point cloud and the single tassel branch point cloud, the fine phenotype index data of the branch length, the branch angle, the branch bending degree and the semi-axis length are determined, so as to realize high-throughput collection of phenotype index data, and improve the fineness and accuracy of testing corn varieties.

[0119] Figure 2is a flowchart of a process for determining all tassel point clouds of a target corn provided by the present application, which determines all tassel point clouds of a target corn according to the multi-layer corn tassel multi-view images, comprising:

[0120] reconstructing the multi-layer corn tassel multi-view images by three-dimensional point clouds, and obtaining multi-layer corn tassel three-dimensional point clouds;

[0121] removing fixed column point clouds from the multi-layer corn tassel three-dimensional point clouds, and obtaining a tassel point cloud group;

[0122] processing the tassel point cloud group according to a clustering algorithm, and obtaining all tassel point clouds of a target corn.

[0123] In step 1011, the Structure from Motion (SFM) algorithm and the Multi View Stereo (MVS) algorithm are used to reconstruct the obtained multi-layer corn tassel multi-view images by three-dimensional point clouds, and obtain multi-layer corn tassel three-dimensional point clouds MTassel PC .

[0124] In step 1012, fixed column point clouds are removed from the multi-layer corn tassel three-dimensional point clouds, and a tassel point cloud group is obtained, Figure 10 is a structural diagram of the multi-layer corn tassel three-dimensional point clouds provided by the present application, as Figure 10 shown, the multi-layer corn tassel three-dimensional point clouds MTassel PC are composed of two parts, one part is fixed column point clouds Support pc , and the other part is a fixed sample branch unit point cloud group wherein Spc ij represents a fixed sample branch point cloud, and the tassel group point cloud wherein Tpc ij represents a tassel point cloud, m is the number of layers, and n is the number of tassels and tassel fixed sample branches of each layer.

[0125] Figure 11 is a structural diagram of the tassel point cloud group provided by the present application, in the process of point cloud denoising, the density of each point of the point cloud is calculated: for each point p i in the point cloud, the number of points t in the surrounding sphere with the point as the center and radius r is calculated, and the density of the point is: p pi =t / r, wherein the point density in the point cloud is less than p1, i.e. p pi < p1, the points are noise points, and then the column point cloud is found and removed. For the tassel point cloud group MTassel PCUsing the principle of the lowest point in the Z-direction, find the lowest point in the point cloud. Using this point as the center, and with a radius greater than twice the radius of the fixed column, construct a cylindrical bounding volume. The point cloud within this bounding volume is the Support. pc The point cloud group of tassels with fixed posts removed is marked as MTassel_S pc ,like Figure 11 The diagram shown is a structural schematic of the male ear dot cloud group after the fixed pillars have been removed.

[0126] In step 1013, Figure 12 This is a schematic diagram of the shortest path structure of the point cloud of two fixed sample branch units provided by the present invention, for the male ear point cloud group MTassel_S pc The Euclidean distance clustering algorithm is used to obtain m clusters. Where m represents the number of tassel sample layers. For each layer of point cloud MTassel_S pci Using the lowest point on the Z-axis as a constraint, the fixed-pattern branch unit point cloud Spc is clipped from the point cloud using a plane. i Based on the point cloud clusters closest to the fixed column, locate the fixed sample branch unit point cloud Spc. i Branch node group {T i}, in sequence (T1, T2), (T2, T3), ... (T u-1 T u ), ...(T n T1) is a pair of branch nodes. Determine the relationship between each pair of branch nodes (T1). u-1 T u In the point cloud MTassel_S pci In the middle, with branch node T u-1 and T u The point cloud MTassel_S within the sector α region pciu Is it connected? If it is connected, then it means T u-1 and T u The corresponding tassel samples exhibit point cloud intersections and should be disconnected. The method is as follows: Based on the point cloud shortest path algorithm, in the point cloud MTassel_S pciu Build node T in u-1 To node T u Shortest path {T u-1 T u-1,1 T u-1,2 , ..., T u-1,k1 , ..., T u,k , ..., T u,2 T u,1 T u}, among the shortest path nodes, select the node {T} located within the sector β. u-1,k1 , ..., Tu,k , wherein β is located in the middle of α and half of the region, and the calculation node {T u-1,k1 ,..., T u,k} each point adjacent segment angle, wherein the largest point P t is the disconnection point, and a disconnection ball is established with P t as the center and r2 as the radius, and the point cloud in the surrounding ball is removed from MTassel_S pci . The above disconnection intersection point cloud processing is sequentially performed on each branch pair to obtain the processed point cloud MTassel_S pci .

[0127] Figure 3 is a flowchart of the corn variety testing method provided by the present application, and before the tassel point cloud is input into the point cloud segmentation model, it further includes:

[0128] For any tassel point cloud sample, all point clouds in the region formed after extending a preset length from the tassel branch tip as the starting point to the tassel branch bottom end are marked as the first label, and all point clouds in other regions except the region are marked as the second label.

[0129] The tassel branch tip point cloud corresponding to the first label and the non-tip point cloud corresponding to the second label are used to construct a sample set, all sample sets corresponding to all tassel point cloud samples are trained, and a point cloud segmentation model is determined.

[0130] In step 201, those skilled in the art understand that the present application realizes the segmentation of the tassel point cloud sample through the point cloud segmentation model. For each corn tassel point cloud Tpc ij , a corn tassel branch tip point cloud deep learning point cloud segmentation model is first constructed, the corn tassel branch tip point cloud is segmented, the first label can be 0, and the second label can be 1. In the construction process of the branch tip point cloud deep learning training data set, the points of the 3cm length tip point cloud of each branch tip of the tassel point cloud sample are marked as 0, and the remaining non-tip point cloud points are marked as 1. All corn tassel point clouds are selected for the above marking operation to form a branch tip point cloud deep learning training data set.

[0131] In step 202, the branch tip point cloud segmentation deep learning network is constructed, a point cloud multi-layer deep learning network layer is constructed, which is a plurality of multi-scale sampling layers, a plurality of grouping layers, a plurality of local point cloud global feature extraction layers and a skip connection, a multi-layer convolution deep learning neural network is constructed, a sample set is constructed from the tassel branch tip point cloud corresponding to the first label and the non-tip point cloud corresponding to the second label, the branch tip point cloud segmentation deep learning network is trained according to the training sample set, a training network is formed, and a point cloud segmentation model is determined.

[0132] Figure 4 is a flowchart of determining a main axis point cloud of a male spike and a branch point cloud of the male spike provided by the present application. The main axis point cloud of the male spike and the branch point cloud of the male spike are determined according to the branch tip point cloud of the male spike and the male spike point cloud, and the method comprises the following steps:

[0133] calculating a gravity point of the male spike point cloud, establishing a surrounding cylinder with the gravity point as the center and a preset length as the radius, and determining a lowest point of the surrounding cylinder in the vertical direction as a main axis base point of the male spike point cloud;

[0134] determining all point cloud paths from the main axis base point of the male spike point cloud to each point cloud of the branch tip point cloud of the male spike;

[0135] determining all point clouds with the shortest point cloud path as the main axis point cloud of the male spike, and determining the point clouds in other point cloud paths except the main axis point cloud of the male spike as the branch point cloud of the male spike.

[0136] In step 1031, a gravity point of the male spike point cloud is calculated, a surrounding cylinder is established with the gravity point as the center and a preset length as the radius, and a lowest point of the surrounding cylinder in the vertical direction is determined as a main axis base point of the male spike point cloud. For the corn male spike branch tip point cloud Tip_pc ij , a Euclidean distance clustering algorithm is used to obtain a branch tip point cloud group wherein Tip i represents the branch tip point cloud of the male spike, and w represents the number of branches of the male spike. For the branch tip point cloud Tip i , a center point CP i of the point cloud is taken to form a branch tip point group For the corn male spike point cloud Tpc ij , a gravity point Gp of the corn male spike point cloud is calculated, a surrounding cylinder is established with the gravity point as the center and a radius r3, and a lowest point in the Z-axis direction among the point clouds in the cylinder is calculated as the main axis base point Basep of the corn male spike.

[0137] In step 1032, in the corn male spike point cloud Tpc ij , the shortest paths of the branch tip point CP i and the main axis base point Basep of the corn male spike are calculated respectively, the branch point cloud corresponding to the branch tip point is searched and expanded along the shortest path, and the point clouds of the two branches in the branch point cloud are counted to form the main axis point cloud MainAxis pc of the corn male spike, and the main axis point cloud is removed to obtain the segmented corn male spike branch point cloud wherein BranchP i is a branch point cloud.

[0138] In step 1033, through the automatic processing of steps 1031 to 1032 described above, the tassel point cloud Tpc of the corn is automatically segmented into the main axis point cloud MainAxis of the tassel ij pc and w tassel branch point clouds

[0139] Figure 5 The structure diagram of the corn variety testing system provided by the present application is shown in the figure. The present application provides a corn variety testing system, which comprises a corn variety testing device of the corn variety testing method as described above. Specifically, the corn variety testing method comprises the following steps: determining all tassel point clouds of a target corn according to multi-layer corn tassel multi-view images; the multi-layer corn tassel multi-view images are obtained by identifying a target corn electronic tag; for any tassel point cloud of the target corn, inputting the tassel point cloud into a point cloud segmentation model to obtain a tassel branch tip point cloud output by the point cloud segmentation model; determining a tassel main axis point cloud and a tassel branch point cloud according to the tassel branch tip point cloud and the tassel point cloud; determining tassel phenotype index data according to the tassel main axis point cloud and the tassel branch point cloud, so as to test the corn variety according to all tassel phenotype index data of the target corn; the tassel phenotype index data at least comprises tassel length, tassel curvature, tassel volume, tassel height, tassel dispersion, corn tassel short semi-axis length, corn tassel long semi-axis length, tassel branch number, tassel branch length, tassel branch angle and tassel branch curvature; the point cloud segmentation model is determined according to each tassel branch tip point cloud and non-tip point cloud in a plurality of groups of different types of tassel point cloud samples of corn varieties; the tassel branch tip point cloud is all point clouds in a region formed by extending a preset length from a tassel branch tip to a tassel branch bottom end.

[0140] The corn variety testing system further comprises a field corn tassel sampling device 1 for sampling corn tassels in a corn planting field. Those skilled in the art understand that a corn test field for carrying out DUS phenotype index research of corn tassels is designed and planted according to experimental requirements, and an RFID radio frequency chip is hung on a corn plant of a test variety for sampling unique number identification.

[0141] The corn variety testing system further comprises a multi-layer corn tassel fixing device 2 for fixing the sampled corn tassels and identifying a target corn electronic tag. An identifier of the multi-layer corn tassel fixing device 2 identifies a radio frequency chip of a fixing hole in the field corn tassel sampling device 1, so as to complete transfer of a target corn number identification.

[0142] ​The multi-view image acquisition device 3 is used for acquiring multi-layered tassel multi-view images of target corn, and after determining the number corresponding to the target corn of the to-be-tested variety, the multi-layered tassel multi-view images of the target corn are acquired, and finally, the acquired multi-layered tassel multi-view images are transmitted to the corn variety testing device to execute the corn variety testing method, so as to obtain the tassel DUS phenotype index of the corn variety and test the target corn variety.

[0143] The application constructs a corn tassel DUS phenotype automatic acquisition and analysis method, provides a convenient tool and system for corn tassel DUS trait investigators, builds an automatic and streamlined corn tassel DUS data acquisition electronic acquisition device from the source, including a field corn tassel electronic sampling table, a multi-layered corn tassel solid sample table and a multi-view image acquisition device and system, realizes high-throughput automatic acquisition of corn tassel DUS data and automatic management of data, and realizes multi-layered and multiple corn tassel sample point cloud segmentation, single corn tassel branch point cloud organ segmentation and corn tassel DUS phenotype analysis through an automatic corn tassel DUS phenotype analysis algorithm.

[0144] The application provides a corn variety testing method, device, system, equipment and medium, the single tassel point cloud is input into the point cloud segmentation model, the single tassel branch tip point cloud is acquired, the single tassel main axis point cloud and the single tassel branch point cloud are determined according to the single tassel branch tip point cloud and the single tassel point cloud, the branch length, the branch angle, the branch bending degree and the semi-major axis length are determined according to the single tassel main axis point cloud and the single tassel branch point cloud, so as to realize high-throughput acquisition of phenotype index data, and improve the fineness and accuracy of the tested corn variety.

[0145] Figure 6 It is a structural schematic view of the field corn tassel sampling equipment provided by the application, and the field corn tassel sampling equipment 1 comprises a plurality of multi-layered solid sample plates 11;

[0146] For each solid sample plate 11, a plurality of solid sample grooves 12 arranged in a matrix are arranged, each solid sample groove 12 is provided with a solid sample hole 13, the solid sample hole 13 comprises a hole outer ring 131, a hole inner ring 132 and a hole bottom 133, wherein the hole outer ring is made of a first material, the hole inner ring is made of a second material, the hole bottom is provided with an electronic chip for reading and writing an electronic tag, the electronic chip is in contact with a radio frequency chip hung on a field corn plant in a working state, and the elasticity of the second material is greater than that of the first material.

[0147] Further comprising a display panel 14, the display panel 14 is used for displaying the sampling results of the corn tassel in the solid sample hole and the electronic tag.

[0148] Optionally, the field corn tassel sampling device 1 is composed of a plurality of fixed sampling boards, which are fixed on a movable trolley for facilitating field sampling. Figure 6 As shown in the structure of a fixed sampling board of the plurality of fixed sampling boards, the fixed sampling boards are supported and fixed in layers by support frames. The fixed board has a length of 160 cm and a width of 100 cm. Fixed sampling grooves 12 are arranged on the fixed board at equal intervals, and the interval is 20 cm. Fixed sampling holes 13 are installed on the fixed sampling grooves. The outer ring 131 of the fixed sampling hole 13 is made of plastic material, the inner ring 132 is filled with sponge material, and the bottom 133 is placed with an RFID radio frequency chip. The fixed sampling hole is easy to place and insert into the fixed sampling groove. A display panel 14 is arranged on one side of the fixed sampling board 11. The display panel 14 is an LED liquid crystal display control panel. The LED liquid crystal panel is used to display whether a sample is inserted into the fixed sampling hole and sample number information. The RFID radio frequency chip corresponding to each fixed sampling hole 13 is electrically connected through a circuit board and the display panel 14. The circuit board and a power supply module are arranged at the bottom of the fixed sampling board. Figure 6 As shown in the structure of a fixed sampling board of the plurality of fixed sampling boards, the fixed sampling boards are supported and fixed in layers by support frames. The fixed board has a length of 160 cm and a width of 100 cm. Fixed sampling grooves 12 are arranged on the fixed board at equal intervals, and the interval is 20 cm. Fixed sampling holes 13 are installed on the fixed sampling grooves. The outer ring 131 of the fixed sampling hole 13 is made of plastic material, the inner ring 132 is filled with sponge material, and the bottom 133 is placed with an RFID radio frequency chip. The fixed sampling hole is easy to place and insert into the fixed sampling groove. A display panel 14 is arranged on one side of the fixed sampling board 11. The display panel 14 is an LED liquid crystal display control panel. The LED liquid crystal panel is used to display whether a sample is inserted into the fixed sampling hole and sample number information. The RFID radio frequency chip corresponding to each fixed sampling hole 13 is electrically connected through a circuit board and the display panel 14. The circuit board and a power supply module are arranged at the bottom of the fixed sampling board.

[0149] Optionally, during sampling, the number of the plant sample is read by an RFID radio frequency reader, and then the corn tassel sample is cut from the base of the corn tassel. The RFID radio frequency reader is contacted with the fixed sampling hole 13 to write the unique number of the corn tassel, and the corn tassel sample is inserted into the fixed sampling hole 13 to be fixed, thereby completing the sampling operation of one corn tassel sample. According to the above operation, the field sampling work of all sampling panels on the sampling trolley is completed, and the multi-layer corn tassel fixed sampling device of the indoor collection platform is transported to perform data collection.

[0150] Figure 7 is a structural schematic view of the multi-layer corn tassel fixed sampling device provided by the application. The multi-layer corn tassel fixed sampling device 2 comprises a fixed column 21 and a plurality of fixed sampling support units 22 arranged on the fixed column.

[0151] Each fixed sampling support unit 22 is provided with a fixed sampling branch unit 23 at the end away from the fixed column 21. The branch unit 23 comprises a sliding groove 231 and an identifier 232. The sliding groove 231 is provided with a fixed sampling groove 233. The identifier 232 is in contact with the fixed sampling hole in the working state to read and write the electronic tag.

[0152] Optionally, as shown in Figure 7As shown, the multi-layer corn tassel fixed sample platform is fixed in the center of the multi-view image acquisition device, and is used for synchronously acquiring a plurality of corn tassel samples. The fixed sample platform is composed of a fixed column 21 and a plurality of fixed sample support units 22. The fixed column is a cylindrical body with a height of 100 cm and a radius of 5 cm and is made of plastic. A plurality of fixed sample support units are installed on the fixed column. The first layer of fixed sample support units is fixed at a height of 20 cm of the fixed column, and the second layer of fixed sample support units is fixed at a height of 80 cm of the fixed column. Each fixed sample support unit 22 is provided with a fixed sample branch unit 23 at the end away from the fixed column 21. The branch unit is provided with 5 or 10 branch units which are distributed around the fixed sample support unit 22 at an average angle and arranged in a circular layout.

[0153] Figure 8 Figure 3 is a structural schematic diagram of the fixed sample branch unit provided by the application. The branch unit 23 includes a sliding groove 231 and an identifier 232. The length of the branch unit 23 is 50 cm, and the length of the sliding groove 231 is 20 cm. The fixed sample groove is used for fixing the electronic fixed sample hole and bearing the corn tassel. The fixed sample groove adjusts the placement position of the corn tassel through the sliding groove 231. The identifier 232 is connected with the fixed sample hole and is used for identifying the unique number of the tassel sample and transmitting the number to the multi-view image acquisition device.

[0154] Figure 9 Figure 4 is a structural schematic diagram of the multi-view image acquisition device provided by the application. The multi-view image acquisition device 3 includes a rotating shaft 31, a rotating arm 32 and an acquisition arm 33. The rotating shaft 31 is connected with the rotating arm 32, and the rotating arm 32 is fixedly connected with the acquisition arm 33.

[0155] The acquisition arm 33 includes a first acquisition arm 331 and a second acquisition arm 332. The first acquisition arm 331 is fixedly provided with a first image collector 3311, and the second acquisition arm 332 is fixedly provided with a second image collector 3321 and a third image collector 3322.

[0156] Optionally, the multi-view image acquisition device is composed of a data acquisition controller, the rotating shaft 31, the rotating arm 32 and the acquisition arm 33. The acquisition arm 33 is provided with a camera array unit. The rotating shaft 31 is controlled by a motor and drives the rotating arm 32 to rotate at a uniform speed around the center axis. The speed can be set to 90 seconds / revolution. The camera array unit is composed of three image collectors, i.e., the first image collector 3311, the second image collector 3321 and the third image collector 3322. The image collectors can be cameras which are distributed and installed on the rotating arm unit. The camera array is electrically connected with the data acquisition controller. The exposure time interval of the camera is set to 2 seconds. The multi-layer corn tassel fixed sample device is fixed in the center of the multi-view image acquisition device. The data acquisition controller is electrically connected with the multi-layer corn tassel fixed sample device.

[0157] Optionally, the first collecting arm 331 and the second collecting arm 332 are an integral structure, and the included angle formed by the first collecting arm 331 and the second collecting arm 332 is greater than a right angle, so that better image collection is achieved.

[0158] Figure 18 It is the third flowchart of the corn variety testing method provided by the application, and the field sampling and electronic marking of the corn tassel are carried out. In the working state, the solid sample hole with the corn tassel is taken out from the solid sample disc and inserted into the solid sample groove. After the placement of all the corn tassels is completed, the data acquisition system is driven to acquire data. The data acquisition controller identifies the unique number of each tassel sample on the multi-layer corn tassel solid sample device, drives the rotating shaft to rotate at a uniform speed, synchronously drives each camera unit of the camera array to take pictures at a fixed time, stores the collected multi-view images into the same folder, records the sample number of each tassel and the corresponding position node information of the multi-layer corn tassel solid sample device, takes out the corn tassel and the solid sample hole after completing a multi-view image data acquisition task of the multi-layer corn tassel, inserts the solid sample hole into the solid sample disc for repeated use, sequentially carries out multi-layer corn tassel point cloud three-dimensional reconstruction, multi-layer corn tassel sample point cloud segmentation and corn tassel branch segmentation based on deep learning, and finally realizes the calculation of the DUS phenotype index of the corn tassel.

[0159] Figure 19 It is a structural schematic diagram of the corn variety testing device provided by the application. The application provides a corn variety testing device, which comprises a first determination unit 1: determining all tassel point clouds of a target corn according to multi-view images of the multi-layer corn tassel, wherein the multi-view images of the multi-layer corn tassel are identified by a target corn electronic tag, and the multi-view images of the multi-layer corn tassel corresponding to the target corn are obtained. The working principle of the first determination unit 1 can refer to the foregoing step 101, and will not be repeated here.

[0160] The corn variety testing device further comprises an acquisition unit 2: for any tassel point cloud of the target corn, inputting the tassel point cloud into a point cloud segmentation model, and acquiring tassel branch tip point clouds output by the point cloud segmentation model. The working principle of the acquisition unit 2 can refer to the foregoing step 102, and will not be repeated here.

[0161] The corn variety testing device further comprises a second determination unit 3: determining tassel main shaft point clouds and tassel branch point clouds according to the tassel branch tip point clouds and the tassel point clouds. The working principle of the second determination unit 3 can refer to the foregoing step 103, and will not be repeated here.

[0162] The corn variety testing device also comprises a third determination unit 4: determining tassel phenotype index data according to the tassel main axis point cloud and the tassel branch point cloud, so as to test the corn variety according to all tassel phenotype index data of the target corn, and the working principle of the third determination unit 4 can refer to the foregoing step 104, which will not be described here.

[0163] The tassel phenotype index data at least comprises tassel length, tassel curvature, tassel volume, tassel height, tassel dispersion, corn tassel short semi-axis length, corn tassel long semi-axis length, tassel branch number, tassel branch length, tassel branch angle and tassel branch curvature.

[0164] The point cloud segmentation model is determined according to each tassel branch tip point cloud and non-tip point cloud in a plurality of groups of different kinds of corn variety tassel point cloud samples.

[0165] The tassel branch tip point cloud is all point clouds in a region formed after extending a preset length from a tassel branch tip to a tassel branch bottom end.

[0166] The present application provides a corn variety testing method, device, system, equipment and medium, by inputting the tassel point cloud to the point cloud segmentation model, obtaining the tassel branch tip point cloud, determining the tassel main axis point cloud and the tassel branch point cloud according to the tassel branch tip point cloud and the tassel point cloud, determining the refined phenotype index data of branch length, branch angle, branch curvature and semi-axis length according to the tassel main axis point cloud and the tassel branch point cloud, so as to realize high-throughput collection of phenotype index data, and improve the refinement and accuracy of testing corn variety.

[0167] Figure 20 It is a structural schematic diagram of the electronic equipment provided by the present application. Figure 20As shown, the electronic device can include a processor 110, a communications interface 120, a memory 130, and a communications bus 140, wherein the processor 110, the communications interface 120, and the memory 130 complete mutual communication through the communications bus 140. The processor 110 can call logical instructions in the memory 130 to execute a corn variety testing method, which includes: determining all tassel point clouds of a target corn according to a multi-layer corn tassel multi-view image, the multi-layer corn tassel multi-view image being determined by identifying a target corn electronic tag and acquiring a corresponding multi-layer corn tassel multi-view image of the target corn; for any tassel point cloud of the target corn, inputting the tassel point cloud into a point cloud segmentation model to obtain a tassel branch tip point cloud output by the point cloud segmentation model; determining a tassel main axis point cloud and a tassel branch point cloud according to the tassel branch tip point cloud and the tassel point cloud; determining tassel phenotype index data according to the tassel main axis point cloud and the tassel branch point cloud, to test a corn variety according to all tassel phenotype index data of the target corn; the tassel phenotype index data at least includes tassel length, tassel curvature, tassel volume, tassel height, tassel dispersion, corn tassel short semi-axis length, corn tassel long semi-axis length, tassel branch number, tassel branch length, tassel branch angle, and tassel branch curvature; the point cloud segmentation model is determined according to each tassel branch tip point cloud and non-tip point cloud in a plurality of groups of different types of corn tassel point cloud samples.

[0168] In addition, the logical instructions in the memory 930 described above can be implemented in the form of a software function unit and sold or used as an independent product, which can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the methods described in various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0169] In another aspect, the present application also provides a computer program product, which comprises a computer program, the computer program being stored on a non-transitory computer-readable storage medium, and the computer program being capable of executing a corn variety testing method provided by any of the above methods when executed by a processor, the method comprising: determining all tassel point clouds of a target corn according to multi-view images of multi-layer tassels of the target corn, the multi-view images of multi-layer tassels of the target corn being determined by identifying an electronic tag of the target corn; inputting any tassel point cloud of the target corn into a point cloud segmentation model to obtain a tassel branch tip point cloud output by the point cloud segmentation model; determining a tassel main axis point cloud and a tassel branch point cloud according to the tassel branch tip point cloud and the tassel point cloud; determining tassel phenotype index data according to the tassel main axis point cloud and the tassel branch point cloud, and testing a corn variety according to all tassel phenotype index data of the target corn; the tassel phenotype index data at least comprising tassel length, tassel curvature, tassel volume, tassel height, tassel dispersion, corn tassel short semi-axis length, corn tassel long semi-axis length, tassel branch number, tassel branch length, tassel branch angle, and tassel branch curvature; the point cloud segmentation model being determined according to each tassel branch tip point cloud and non-tip point cloud in a plurality of groups of different kinds of tassel point cloud samples of corn varieties; and the tassel branch tip point cloud being all point clouds in a region formed by extending a preset length from a tassel branch tip to a tassel branch bottom end.

[0170] In yet another aspect, the present application also provides a non-transitory computer-readable storage medium having stored thereon a computer program, which, when executed by a processor, implements the method for testing corn varieties provided above, which includes: determining all tassel point clouds of a target corn from a multi-layer corn tassel multi-view image, the multi-layer corn tassel multi-view image being determined by identifying an electronic tag of the target corn; for any tassel point cloud of the target corn, inputting the tassel point cloud into a point cloud segmentation model to obtain a tassel branch tip point cloud output by the point cloud segmentation model; determining a tassel main axis point cloud and a tassel branch point cloud according to the tassel branch tip point cloud and the tassel point cloud; determining tassel phenotype index data according to the tassel main axis point cloud and the tassel branch point cloud, so as to test corn varieties according to all tassel phenotype index data of the target corn; the tassel phenotype index data at least includes tassel length, tassel curvature, tassel volume, tassel height, tassel dispersion, corn tassel short semi-axis length, corn tassel long semi-axis length, tassel branch number, tassel branch length, tassel branch angle, and tassel branch curvature; the point cloud segmentation model is determined according to each tassel branch tip point cloud and non-tip point cloud in a plurality of groups of different types of corn tassel point cloud samples.

[0171] The device embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the embodiment scheme according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0172] From the above description of the embodiments, those skilled in the art can clearly understand that the embodiments can be realized by means of software plus necessary universal hardware platforms, and of course can also be realized by hardware. Based on such understanding, the above technical solutions, essentially or in terms of the contribution to the prior art, can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, or an optical disk, and includes a plurality of instructions to cause a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0173] It should be pointed out finally that the above embodiments are only used to illustrate the technical solutions of the present application, but not to limit the same; and although the present application has been described in detail with reference to the foregoing embodiments, it should be appreciated by those skilled in the art that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features thereof can be replaced equivalently; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for testing maize varieties, characterized in that, include: The point cloud of all single tassels of the target corn is determined based on the multi-layer maize tassel multi-view image. The multi-layer maize tassel multi-view image is determined by identifying the electronic tag of the target corn and obtaining the corresponding multi-layer maize tassel multi-view image. For any single tassel point cloud of the target maize, input the single tassel point cloud into the point cloud segmentation model, and obtain the single tassel branch tip point cloud output by the point cloud segmentation model. The point cloud of the main axis of the single tassel and the point cloud of the single tassel branch are determined based on the point cloud of the single tassel branch tip and the single tassel point cloud. The single tassel phenotypic index data are determined based on the single tassel main axis point cloud and the single tassel branch point cloud, so as to test maize varieties based on all single tassel phenotypic index data of the target maize. The single tassel phenotypic index data shall include at least the tassel length, tassel curvature, tassel volume, tassel height, tassel dispersion, maize tassel short semi-axis length, maize tassel long semi-axis length, number of tassel branches, tassel branch length, tassel branch angle, and tassel branch curvature. The point cloud segmentation model is determined by training on the point cloud at the tip of each tassel branch and the non-tip point cloud in multiple sets of tassel point cloud samples of different maize varieties. The point cloud at the tip of the male spike branch is the point cloud within the area formed by extending a preset length from the tip of the male spike branch to the bottom of the male spike branch.

2. The method for testing maize varieties according to claim 1, characterized in that, The step of determining all single tassel point clouds of the target maize based on the multi-layer maize tassel multi-view image includes: The multi-layered maize tassel multi-view image is reconstructed from the three-dimensional point cloud to obtain the multi-layered maize tassel three-dimensional point cloud; Remove the fixed column point cloud from the multi-layered three-dimensional point cloud of maize tassels to obtain the tassel point cloud group; The tassel point cloud group is processed using a clustering algorithm to obtain all single tassel point clouds of the target corn.

3. The method for testing maize varieties according to claim 1, characterized in that, Before inputting the single tassel point cloud to point cloud segmentation model, the following is also included: For any tassel point cloud sample, all point clouds within the area formed after extending a preset length from the tip of the tassel branch to the bottom of the tassel branch are labeled as the first label, and all point clouds in other areas besides the aforementioned area are labeled as the second label. A sample set is constructed using the tip point cloud of the male spike branch corresponding to the first label and the non-tip point cloud corresponding to the second label. All sample sets corresponding to all male spike point cloud samples are trained to determine the point cloud segmentation model.

4. The method for testing maize varieties according to claim 1, characterized in that, The step of determining the single tassel main axis point cloud and the single tassel branch point cloud based on the single tassel branch tip point cloud and the single tassel point cloud includes: Calculate the centroid of the single tassel point cloud, establish a confining cylinder with the centroid as the center and a preset length as the radius, and determine the lowest point of the confining cylinder in the vertical direction as the main axis base point of the single tassel point cloud. Determine all point cloud paths from the main axis base point of the single tassel point cloud to the tip point cloud of each single tassel branch; All point clouds with the shortest point cloud paths are identified as single tassel main axis point clouds, and the point clouds in other point cloud paths other than single tassel main axis point clouds are identified as single tassel branch point clouds.

5. The method for testing maize varieties according to claim 1, characterized in that, The determination of single tassel phenotypic index data based on the single tassel main axis point cloud and the single tassel branch point cloud includes: The direction of the single tassel point cloud is corrected to the vertical direction according to the correction matrix, which is determined by linear fitting of the key node sequence of the skeleton after skeleton shrinkage of the single tassel main axis point cloud. The length of the male spike is determined based on the shortest distance from the highest point of the point cloud at the tip of the single male spike branch to the base point of the main axis of the single male spike. The height of the male ear is determined based on the height difference between the highest point of the tip cloud of the single male ear branch in the vertical direction and the base point of the main axis of the single male ear. The curvature of the tassel is determined based on the height and length of a single tassel in corn. The volume of the male spike is determined based on the volume of the convex hull enclosed by the point cloud of the main axis of the single male spike and the point cloud of the branch of the single male spike. The length of the short semi-axis of the tassel is determined based on the length of the dot cloud of the main axis of the single tassel. The length of the long semi-axis of the tassel is determined based on the difference between the length of the tassel and the length of the short semi-axis of the tassel. The length of the male spike branch is determined by the shortest distance between the first node and the last node of the male spike branch. The branch angle of the male spike is determined based on the angle between the point cloud of the main axis of the single male spike and the point cloud of the branch of the single male spike. The curvature of the male spike branch is determined based on the straight-line distance between the first node and the last node of the male spike branch and the length of the male spike branch. The tassel dispersion is determined based on the mean of the curvature of all tassel branches.

6. A maize variety testing system, characterized in that, Includes a maize variety testing device based on the maize variety testing method according to any one of claims 1-5; Also includes: Field corn tassel sampling equipment is used to sample corn tassels in corn planting fields; Multi-layer corn tassel fixation device is used to fix the corn tassels after sampling and to identify the target corn electronic tag; Multi-view image acquisition equipment is used to acquire multi-view images of the tassels of corn corresponding to the target corn.

7. The maize variety testing system according to claim 6, characterized in that, The field corn tassel sampling equipment includes: a sampling vehicle, on which multiple layers of sample fixing plates are fixed; For each layer of the sample plate, multiple sample slots are arranged in a matrix. Each sample slot is provided with a sample hole. The sample hole includes an outer ring, an inner ring, and a bottom. The outer ring is made of a first material, the inner ring is made of a second material, and an electronic chip for reading and writing electronic tags is provided at the bottom. When the electronic chip is in working condition, it is in contact with the radio frequency chip suspended on the corn plant in the field. The elasticity of the second material is greater than that of the first material. It also includes a display panel for displaying the sampling results of the corn tassels in the sample well and the electronic tag.

8. The maize variety testing system according to claim 6, characterized in that, The multi-layer maize tassel fixation device includes: a fixed column, and multiple fixation support units disposed on the fixed column; Each sample-fixing support unit has a sample-fixing branch unit at its end away from the fixed column. The sample-fixing branch unit includes a chute and an identifier. The chute has a sample-fixing groove, and the identifier contacts the sample-fixing hole in the working state to read and write electronic tags.

9. The maize variety testing system according to claim 6, characterized in that, The multi-view image acquisition device includes: a rotating shaft, a rotating arm, and an acquisition arm, wherein the rotating shaft is connected to the rotating arm, and the rotating arm is fixedly connected to the acquisition arm; The acquisition arm includes a first acquisition arm and a second acquisition arm. A first image acquisition device is fixed on the first acquisition arm, and a second image acquisition device and a third image acquisition device are fixed on the second acquisition arm.

10. A corn variety testing device, characterized in that, include: First determining unit: Determine all single tassel point clouds of the target corn based on the multi-layer corn tassel multi-view image, wherein the multi-layer corn tassel multi-view image is determined by identifying the target corn electronic tag and obtaining the multi-layer corn tassel multi-view image corresponding to the target corn; Acquisition Unit: For any single tassel point cloud of the target maize, input the single tassel point cloud into the point cloud segmentation model, and acquire the single tassel branch tip point cloud output by the point cloud segmentation model; Second determining unit: Determine the single tassel main axis point cloud and the single tassel branch point cloud based on the single tassel branch tip point cloud and the single tassel point cloud. The third determining unit: determines the single tassel phenotypic index data based on the single tassel main axis point cloud and the single tassel branch point cloud, so as to test the maize variety based on all single tassel phenotypic index data of the target maize. The single tassel phenotypic index data shall include at least the tassel length, tassel curvature, tassel volume, tassel height, tassel dispersion, maize tassel short semi-axis length, maize tassel long semi-axis length, number of tassel branches, tassel branch length, tassel branch angle, and tassel branch curvature. The point cloud segmentation model is determined by training on the point cloud at the tip of each tassel branch and the non-tip point cloud in multiple sets of tassel point cloud samples of different maize varieties. The point cloud at the tip of the male spike branch is the point cloud within the area formed by extending a preset length from the tip of the male spike branch to the bottom of the male spike branch.

11. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the corn variety testing method as described in any one of claims 1-5.

12. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the corn variety testing method as described in any one of claims 1-5.

Images