A Method for Identifying Roll-Resistant Varieties of Regenerated Crops Based on UAV LiDAR Point Clouds

Abstract

This invention provides a method for identifying trampling-resistant varieties of regenerated crops based on UAV lidar point clouds. The method includes: acquiring three-dimensional point cloud data of a target monitoring area from multi-temporal global images of crops collected by a UAV lidar device; extracting trampled and untrapped row crop points from the three-dimensional point cloud data; extracting trampling resistance indices from the trampled and untrapped row crop points respectively; predicting the trampling resistance of the crop based on the trampling resistance indices corresponding to the trampled and untrapped row crop points; and determining whether the crop is a trampling-resistant variety based on its trampling resistance. This invention enables high-throughput automatic identification of trampling-resistant varieties of regenerated crops, which is beneficial for improving the screening efficiency of high-yielding varieties of regenerated crops.

Description

Technical Field

[0001] This invention relates to the field of rapid acquisition of attribute information of regenerated rice crops, and in particular to a method for identifying crush-resistant varieties of regenerated crops based on UAV lidar point clouds. Background Technology

[0002] Ratoon rice refers to a planting method where, after rice harvest, dormant buds on the rice stubble are allowed to grow into panicles through proper cultivation, resulting in another harvest. This method offers advantages such as increased yield, higher income, and improved efficiency. Full mechanization of ratoon rice production is an important way to further reduce labor intensity and costs, and promote the modernization and increased production capacity of ratoon rice.

[0003] However, during the first harvest of ratooning rice, the tracks of the harvester damage the rice stubble, affecting the growth of the ratooning season and hindering high and stable yields. At the same time, the lack of trampling-resistant varieties limits the improvement of yield and quality in machine-harvested ratooning rice. Therefore, screening for trampling-resistant varieties under field conditions can help increase yields in the ratooning season.

[0004] Currently, the commonly used method for screening roll-resistant varieties of ratooning rice is manual field sampling survey. This method has problems such as large human error, large labor input, low work efficiency, and small sample coverage, which cannot meet the needs of agricultural production. Summary of the Invention

[0005] This invention provides a method for identifying trampling-resistant varieties of regenerated crops based on UAV lidar point clouds, which solves the problems of the existing technology that uses manual field sampling and investigation to screen trampling-resistant varieties of regenerated rice, which is time-consuming, labor-intensive and inefficient, and achieves high-throughput identification of trampling-resistant varieties of regenerated rice.

[0006] This invention provides a method for identifying trampling-resistant varieties of regenerated crops based on UAV lidar point clouds, comprising:

[0007] Three-dimensional point cloud data of the target monitoring area is obtained from multi-temporal global images of crops collected by UAV lidar equipment, and compacted and uncompacted crop points are extracted from the three-dimensional point cloud data.

[0008] The rolling tolerance index of the crop is extracted from the rolled row points and the unrolled row points respectively;

[0009] The compaction resistance of the crop is predicted based on the compaction resistance index corresponding to the compacted and uncompacted rows, and the compaction resistance of the crop is used to determine whether the crop is a compaction-resistant variety.

[0010] According to the present invention, a method for identifying roll-resistant varieties of regenerated crops based on UAV lidar point clouds is provided. The roll-resistant index of the crop includes one or more of the following: canopy characteristics, stem inclination angle, stem resilience, and growth status of regenerated buds in the roll row.

[0011] The canopy traits include crop canopy height, density, and volume, as well as one or more of canopy coverage, three-dimensional canopy surface area, and three-dimensional canopy index.

[0012] According to the present invention, a method for identifying roll-resistant varieties of regenerated crops based on UAV lidar point clouds is provided, which extracts roll-resistant indices of crops from the rolled row points and the unrolled row points, respectively, including:

[0013] From the compacted row points and the uncompacted row points, select three-dimensional points whose height is greater than the maximum height of the compacted row points by a first preset ratio, and use the average height of the three-dimensional points as the crop canopy height;

[0014] From the compacted row points and the uncompacted row points, select three-dimensional points with a height greater than the maximum height of the compacted row points at a second preset ratio as the canopy point cloud, and calculate the canopy coverage based on the top view of the canopy point cloud;

[0015] The two-dimensional projection image of the canopy point cloud is converted into a spectrum map by two-dimensional Fourier transform, and the zero frequency component of the spectrum map is moved to the center of the spectrum. Gaussian fitting is used to fit the frequency and amplitude of the height value in the spectrum map and plot the curve. The three-dimensional canopy index 3DCI is obtained according to the area under the curve.

[0016] The canopy point cloud is voxelized, and a canopy triangular mesh model is reconstructed based on the voxelized canopy point cloud. The three-dimensional canopy surface area is obtained based on the canopy triangular mesh model.

[0017] According to the present invention, a method for identifying trampling-resistant varieties of regenerated crops based on UAV lidar point clouds is provided, wherein the three-dimensional canopy surface area is obtained from the canopy triangular mesh model using the following formula:

[0018]

[0019] Among them, S CS S represents the three-dimensional canopy surface area. CM S represents the surface area of ​​the canopy triangular mesh model. P The sampling area is represented by k, which is the surface area coefficient of the canopy triangular mesh model.

[0020] According to the present invention, a method for identifying trampling-resistant varieties of regenerated crops based on UAV lidar point clouds is provided, wherein the stem inclination angle is calculated using the following formula:

[0021]

[0022] Where θ is the inclination angle of the stem, H R H represents the height of the crop canopy corresponding to the compaction point of the crop row. N This indicates the height of the crop canopy corresponding to the non-compacted row crop point.

[0023] According to the present invention, a method for identifying crush-resistant varieties of regenerated crops based on UAV lidar point clouds is provided, wherein the stem resilience is calculated using the following formula:

[0024] RD i =θ t1 -θ t2

[0025] Wherein, the stem resilience calculated for the i-th time after crop harvest, θ t1 and θ t2 The stem inclination angles were obtained from sampling at times t1 and t2, respectively, during different growth stages.

[0026] According to the present invention, a method for identifying roll-resistant varieties of regenerated crops based on UAV lidar point clouds predicts the roll-resistantness of the crop using the following formula based on the roll-resistant index corresponding to the rolled and unrolled rows:

[0027]

[0028] Wherein, RI represents the crop's resistance to rolling, and RD... i The stem resilience is calculated at different times, where n is the number of times the stem resilience is calculated, and S is the number of times the stem resilience is calculated. csR and S csN H represents the canopy surface area of ​​the compacted row crop point and the non-compacted row crop point, respectively. R and H N , respectively, represent the crop canopy heights at the compacted row crop points and the non-compacted row crop points, with a, b, and c being weighting coefficients.

[0029] According to the present invention, a method for identifying trampling-resistant varieties of regenerated crops based on UAV lidar point clouds includes extracting trampled and non-trampled row crop points from the three-dimensional point cloud data, comprising:

[0030] After removing noise from the 3D point cloud data, the 3D point cloud data is divided into ground points and crop points;

[0031] The crop points are divided into compacted row crop points and uncompacted row crop points.

[0032] This invention also provides a device for identifying trampling-resistant varieties of regenerated crops based on UAV lidar point clouds, comprising:

[0033] The identification module is used to obtain three-dimensional point cloud data of the target monitoring area from the multi-temporal global image of crops collected by the UAV lidar equipment, and to extract compacted and uncompacted crop points from the three-dimensional point cloud data.

[0034] The extraction module is used to extract the rolling resistance index of crops from the rolled crop points and the unrolled crop points, respectively.

[0035] The identification module is used to predict the rolling resistance of the crop based on the rolling resistance index corresponding to the rolled and unrolled crop points, and to determine whether the crop is a rolling-resistant variety based on the rolling resistance of the crop.

[0036] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for identifying crush-resistant varieties of regenerated crops based on UAV lidar point clouds as described above.

[0037] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for identifying crush-resistant varieties of regenerated crops based on UAV lidar point clouds as described above.

[0038] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the method for identifying crush-resistant varieties of regenerated crops based on UAV lidar point clouds as described above.

[0039] The present invention provides a method for identifying roll-resistant varieties of regenerated crops based on UAV lidar point clouds. By using multi-temporal global images of crops collected by UAV lidar, a high-density point cloud can be generated. Based on this, roll-resistant indices of crops can be extracted over a large scale and with high throughput. Based on the extracted roll-resistant indices, roll-resistant varieties of regenerated crops can be automatically identified in a high-throughput manner. This method is beneficial to improving the screening efficiency of high-yielding varieties of regenerated crops and is suitable for large-scale and high-throughput evaluation of the roll-resistantness and field growth status of regenerated rice varieties. Attached Figure Description

[0040] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0041] Figure 1 This is one of the flowcharts of the method for identifying crush-resistant varieties of regenerated crops based on UAV lidar point clouds provided by the present invention;

[0042] Figure 2 This is the second flowchart of the method for identifying crush-resistant varieties of regenerated crops based on UAV lidar point clouds provided by the present invention.

[0043] Figure 3 This is a schematic diagram of the structure of the device for identifying crush-resistant varieties of regenerated crops based on UAV lidar point clouds provided by the present invention;

[0044] Figure 4 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0046] The following is combined with Figure 1 This invention describes a method for identifying trampling-resistant varieties of regenerated crops based on UAV lidar point clouds, comprising:

[0047] Step 101: Obtain three-dimensional point cloud data of the target monitoring area from the multi-temporal global image of crops collected by the UAV lidar equipment, and extract compacted and uncompacted crop points from the three-dimensional point cloud data;

[0048] The crop in this embodiment can be regenerated rice or other crops.

[0049] LiDAR is not easily affected by weather conditions, has strong penetrating power, and can accurately, quickly, efficiently, cost-effectively, and over a wide area extract information on the growth of ratooning rice populations. Therefore, this embodiment uses a drone equipped with a LiDAR, which can generate high-density point clouds and extract paddy field information over a large area with high throughput.

[0050] The UAV lidar measurement system used in this embodiment includes a multi-rotor UAV, a lidar sensor, a GPS (Global Positioning System) positioning system, and an inertial navigation system.

[0051] UAV lidar equipment can be used to collect multi-temporal global images of crops in regenerated rice fields, and the data can be preprocessed to obtain three-dimensional point cloud data of the target monitoring area.

[0052] Multi-temporal global images can include the day of the first harvest of ratooning rice, 1 day after harvest, 7 days after harvest, 21 days after harvest, the heading stage of the ratooning season, and the maturity stage.

[0053] Preprocessing of multi-temporal global crop images may include: generating LAS format point clouds from the collected point cloud data using ZhiTu software, and reconstructing them to obtain 3D point cloud data containing RGB images; drawing the target cell using ArcGIS software to obtain vector files of the region of interest; and using the ExtractLas tool in the arcpy library of ArcGIS to extract LAS point cloud data of a specified area from the LAS point cloud data.

[0054] The 3D point cloud data can be divided into ground points and crop points, and the crop points can be further divided into compacted crop points and uncompacted crop points.

[0055] The algorithm converts 3D point cloud data into 2D raster data, generating DSM (Digital Surface Model) and DEM (Digital Elevation Model) models. These models are then subtracted in ArcGIS to generate a crop canopy model (DHM). Coordinates of crop distribution points are extracted, and statistical values ​​representing the canopy top and ground elevation are extracted separately. Cluster analysis and machine learning models are used to categorize elevation values ​​into compacted and uncompacted rows, obtaining the geographic coordinates of the boundary vertices of these rows. The algorithm uses the `transform` function to obtain affine transformation matrices, converting the pixel coordinates of these vertices to their corresponding geographic coordinates. The algorithm batch-outputs Shapefiles (ESRI Shapefiles) corresponding to the geographic coordinates of each cell's vertices. Finally, the `clip_lidar_to_polygon` function completes the segmentation and saving of the 3D point clouds for compacted and uncompacted rows in the field.

[0056] After segmenting the point cloud of each cell, the orientation of different rows of compacted and uncompacted point clouds is automatically calibrated for subsequent automated trait processing. First, the cornerHarris function is used to extract the inflection points of the cell image, and four vertices are selected by combining the image edge values. Then, the getPerspectiveTransform function is used to perform a two-dimensional perspective transformation on the image, and the orientation calibration of all compacted and uncompacted point clouds within the target field is completed iteratively.

[0057] Step 102: Extract the crop's rolling tolerance index from the rolled row points and the unrolled row points respectively;

[0058] It can calculate crop canopy height and extract stem inclination angles and canopy surface areas at different stages of compaction. Combining the stem resilience of compacted rows after the first harvest with the canopy characteristics of the compacted rows in the regenerated season, it determines the crop's compaction tolerance. A complete flowchart is shown below. Figure 2 As shown.

[0059] Step 103: Predict the rolling resistance of the crop based on the rolling resistance index corresponding to the rolled and unrolled crop points, and determine whether the crop is a rolling-resistant variety based on the rolling resistance of the crop.

[0060] This embodiment uses multi-temporal global images of crops acquired by UAV lidar to generate high-density point clouds. Based on these, the crush resistance index of crops can be extracted over a large scale and with high throughput. Then, based on the extracted crush resistance index, the crush resistance varieties of regenerated crops can be automatically identified in high throughput, which is beneficial to improving the screening efficiency of high-yielding varieties of regenerated crops. It is suitable for large-scale and high-throughput evaluation of the crush resistance and field growth status of regenerated rice varieties.

[0061] Based on the above embodiments, the crop's rolling resistance index in this embodiment includes one or more of the following: crop canopy characteristics, stem tilt angle, stem resilience, and growth status of regenerated buds in the rolling row.

[0062] The canopy traits include crop canopy height, density, and volume, as well as one or more of canopy coverage, three-dimensional canopy surface area, and three-dimensional canopy index.

[0063] The stalk tilt angle is the angle between the stalk and the ground at different sampling times after being compacted by the combine harvester in the first season. The stalk resilience is the change in the angle between the stalk and the ground after the compaction row is subjected to combine harvester compaction.

[0064] Based on the above embodiments, this embodiment extracts the crop's rolling tolerance index from the rolled row points and the unrolled row points, respectively, including:

[0065] From the compacted row points and the uncompacted row points, select three-dimensional points whose height is greater than the maximum height of the compacted row points by a first preset ratio, and use the average height of the three-dimensional points as the crop canopy height;

[0066] From the compacted row points and the uncompacted row points, select three-dimensional points with a height greater than the maximum height of the compacted row points at a second preset ratio as the canopy point cloud, and calculate the canopy coverage based on the top view of the canopy point cloud;

[0067] The two-dimensional projection image of the canopy point cloud is converted into a spectrum map by two-dimensional Fourier transform, and the zero frequency component of the spectrum map is moved to the center of the spectrum. Gaussian fitting is used to fit the frequency and amplitude of the height value in the spectrum map and plot the curve. The three-dimensional canopy index 3DCI is obtained according to the area under the curve.

[0068] The canopy point cloud is voxelized, and a canopy triangular mesh model is reconstructed based on the voxelized canopy point cloud. The three-dimensional canopy surface area is obtained based on the canopy triangular mesh model.

[0069] For crop canopy height, three-dimensional points greater than 90% of the maximum canopy height of the plot can be selected, and the average height of these three-dimensional points can be used as the average canopy height of the plot to reduce the error caused by the difference in plant height in the plot.

[0070] For canopy cover characteristics, canopy point clouds can be extracted using 60% of the maximum canopy height as a threshold height. Then, based on the top-down view of the canopy point cloud, the `threshold_local` function is used to obtain a binary map of the corresponding cell canopy region, thereby calculating the cell-scale canopy cover. The canopy cover value ranges from 0 to 1; the closer the value is to 1, the higher the cell canopy cover.

[0071] The calculation of 3DCI traits involves converting the acquired cell image into the frequency domain (spectral map) using a two-dimensional Fourier transform and shifting the zero-frequency component to the center of the spectrum. Then, Gaussian fitting is used to fit the frequency and amplitude of the height value in the spectrum map and plot the curve. The area under the curve is obtained using the trapz function and normalized to obtain the 3DCI.

[0072] For the three-dimensional canopy surface area characteristics, the voxel_down_sample function is used to voxelize the canopy point cloud to reduce the point cloud density and then the key three-dimensional spatial points are calibrated; the creat_from_point_cloud_alpha_shape function is then used to reconstruct the canopy triangular mesh model; finally, the get_surface_area function is used to obtain the three-dimensional surface area of ​​the canopy within the cell.

[0073] Based on the above embodiments, this embodiment obtains the three-dimensional canopy surface area according to the canopy triangular mesh model using the following formula:

[0074]

[0075] Among them, S CS S represents the three-dimensional canopy surface area. CM S represents the surface area of ​​the canopy triangular mesh model. P The sampling area is represented by k, which is the surface area coefficient of the canopy triangular mesh model.

[0076] Due to the varying number of point clouds in different cells, the 3D mesh model of the canopy in cells with sparse point clouds may be incomplete. Therefore, the 3D canopy surface area is calculated using formula (1). The value of k can range from 0.6 to 0.7, which is close to one standard deviation.

[0077] Based on the above embodiments, the stem tilt angle is calculated using formula (2) in this embodiment:

[0078]

[0079] Where θ is the inclination angle of the stem, H R H represents the height of the crop canopy corresponding to the compaction point of the crop row. N This indicates the height of the crop canopy corresponding to the non-compacted row crop point.

[0080] Based on the above embodiments, the stalk resilience is calculated using formula (3) in this embodiment:

[0081] RD i =θ t1 -θ t2 (3)

[0082] Wherein, the stem resilience calculated for the i-th time after crop harvest, θ t1 and θ t2 The stem inclination angles were obtained from sampling at times t1 and t2, respectively, during different growth stages.

[0083] RD iThe rebound rate is measured at different growth stages after harvest, ranging from 0 to 90°, where 0 indicates no rebound and 90° indicates complete rebound.

[0084] Based on the above embodiments, this embodiment uses formula (4) to predict the compaction resistance of the crop according to the compaction resistance index corresponding to the compacted and uncompacted rows:

[0085]

[0086] Wherein, RI represents the crop's resistance to rolling, and RD... i The stem resilience is calculated at different times, where n is the number of times the stem resilience is calculated, and S is the number of times the stem resilience is calculated. csR and S csN H represents the canopy surface area of ​​the compacted row crop point and the non-compacted row crop point, respectively. R and H N , respectively, represent the crop canopy heights at the compacted row crop points and the non-compacted row crop points, with a, b, and c being weighting coefficients.

[0087] Rolling resistance (RI) refers to the growth status of a row after it has been rolled. The higher the rolling resistance, the more easily the variety is rolled.

[0088] Based on the above embodiments, this embodiment extracts compacted and uncompacted crop points from the three-dimensional point cloud data, including:

[0089] After removing noise from the 3D point cloud data, the 3D point cloud data is divided into ground points and crop points;

[0090] The crop points are divided into compacted row crop points and uncompacted row crop points.

[0091] Remove noise from 3D point cloud data while retaining ground and crop point data. A noise removal algorithm that removes statistical outliers can be used, with the specific steps as follows:

[0092] (1) For each point i in the 3D point cloud data, find all its neighboring points j and calculate the distance d from each point to its neighboring points. ij The distance parameters are modeled using a Gaussian distribution model d~N(μ,σ). For each point, the mean μ and standard deviation σ of the distance between it and its neighbors are calculated. These values ​​reflect the distribution characteristics of the distance between the point and its neighbors.

[0093] (2) Traverse all points and calculate the distance d between each point and its neighboring points. ij If the mean of its distance is greater than the specified confidence level of the Gaussian distribution, it is removed.

[0094] The following describes the device for identifying roll-resistant varieties of regenerated crops based on UAV lidar point clouds provided by the present invention. The device for identifying roll-resistant varieties of regenerated crops based on UAV lidar point clouds described below can be referred to in correspondence with the method for identifying roll-resistant varieties of regenerated crops based on UAV lidar point clouds described above.

[0095] like Figure 3 As shown, the device includes an identification module 301, an extraction module 302, and an authentication module 303, wherein:

[0096] The identification module 301 is used to obtain three-dimensional point cloud data of the target monitoring area from the multi-temporal global image of crops collected by the UAV lidar equipment, and extract compacted and uncompacted crop points from the three-dimensional point cloud data;

[0097] Extraction module 302 is used to extract the rolling resistance index of crops from the rolled crop points and the unrolled crop points respectively;

[0098] The identification module 303 is used to predict the rolling resistance of the crop based on the rolling resistance index corresponding to the rolled and unrolled crop points, and to determine whether the crop is a rolling-resistant variety based on the rolling resistance of the crop.

[0099] This embodiment uses multi-temporal global images of crops acquired by UAV lidar to generate high-density point clouds. Based on these, the crush resistance index of crops can be extracted over a large scale and with high throughput. Then, based on the extracted crush resistance index, the crush resistance varieties of regenerated crops can be automatically identified in high throughput, which is beneficial to improving the screening efficiency of high-yielding varieties of regenerated crops. It is suitable for large-scale and high-throughput evaluation of the crush resistance and field growth status of regenerated rice varieties.

[0100] Figure 4 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 4As shown, the electronic device may include a processor 410, a communication interface 420, a memory 430, and a communication bus 440, wherein the processor 410, the communication interface 420, and the memory 430 communicate with each other through the communication bus 440. The processor 410 can call logical instructions in the memory 430 to execute a method for identifying roll-resistant varieties of regenerated crops based on UAV lidar point clouds. This method includes: acquiring three-dimensional point cloud data of the target monitoring area from multi-temporal global images of crops collected by a UAV lidar device; extracting rolled and unrolled row points from the three-dimensional point cloud data; extracting roll-resistant indices of the crops from the rolled and unrolled row points respectively; predicting the roll-resistant properties of the crops based on the roll-resistant indices corresponding to the rolled and unrolled row points; and determining whether the crops are roll-resistant varieties based on the roll-resistant properties of the crops.

[0101] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0102] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the method for identifying roll-resistant varieties of regenerated crops based on UAV lidar point clouds provided by the above methods. The method includes: acquiring three-dimensional point cloud data of a target monitoring area from multi-temporal global images of crops collected by a UAV lidar device; extracting rolled and unrolled row points from the three-dimensional point cloud data; extracting roll-resistant indices of the crops from the rolled and unrolled row points respectively; predicting the roll-resistantness of the crops based on the roll-resistant indices corresponding to the rolled and unrolled row points; and determining whether the crops are roll-resistant varieties based on the roll-resistantness of the crops.

[0103] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements the method for identifying roll-resistant varieties of regenerated crops based on UAV lidar point clouds provided by the above methods. The method includes: acquiring three-dimensional point cloud data of a target monitoring area from multi-temporal global images of crops collected by a UAV lidar device; extracting rolled and unrolled row points from the three-dimensional point cloud data; extracting roll-resistant indices of the crops from the rolled and unrolled row points respectively; predicting the roll-resistantness of the crops based on the roll-resistant indices corresponding to the rolled and unrolled row points; and determining whether the crops are roll-resistant varieties based on the roll-resistantness of the crops.

[0104] The device embodiments described above are merely illustrative. The units described as crop separation components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0105] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0106] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying a regenerative crop variety resistant to rolling based on unmanned aerial vehicle (UAV) LiDAR point cloud, characterized in that, The method comprises the following steps: acquiring three-dimensional point cloud data of a target monitoring area from multi-temporal global images of crops collected by a UAV laser radar device, and extracting rolled row crop points and non-rolled row crop points from the three-dimensional point cloud data; extracting the rolling resistance index of the crops from the rolled row crop points and the non-rolled row crop points, respectively; predicting the rolling resistance of the crops according to the rolling resistance indexes corresponding to the rolled row crop points and the non-rolled row crop points, and determining whether the crops are rolling resistance varieties according to the rolling resistance of the crops; predicting the rolling resistance of the crops according to the rolling resistance indexes corresponding to the rolled row crop points and the non-rolled row crop points by the following formula: ； wherein, RD is the crush resistance of the crop, i S is the stem resilience calculated at different times, n is the number of times the stem resilience is calculated, and csR and S csN are the canopy surface area of the crush row crop point and the non-crush row crop point, respectively, R and H N are the crop canopy height of the crush row crop point and the non-crush row crop point, respectively, and a, b, and c are weight coefficients.

2. The method for identifying a regenerative crop roller- resistant variety based on UAV LiDAR point cloud according to claim 1, characterized in that, the rolling resistance index of the crops comprises one or more of the canopy characteristics, stem inclination angle, stem resilience and rolled row regenerative bud growth conditions of the crops; the canopy characteristics comprise one or more of the canopy height, density and volume of the crops, and the canopy coverage, three-dimensional canopy surface area and three-dimensional canopy index. 3.The method of identifying regenerative crop varieties based on UAV LiDAR point cloud for resistance to rolling pressure according to claim 2, wherein, extracting the rolling resistance index of the crops from the rolled row crop points and the non-rolled row crop points, respectively, comprises: selecting three-dimensional points with a height greater than the maximum height of the rolled row crop points by a first preset proportion from the rolled row crop points and the non-rolled row crop points, and taking the average height of the three-dimensional points as the canopy height of the crops; selecting three-dimensional points with a height greater than the maximum height of the rolled row crop points by a second preset proportion from the rolled row crop points and the non-rolled row crop points as a canopy point cloud, and calculating the canopy coverage according to the overhead view of the canopy point cloud; converting the two-dimensional projection image of the canopy point cloud into a frequency spectrum image by two-dimensional Fourier transform, moving the zero frequency component of the frequency spectrum image to the frequency spectrum center, fitting the frequency and amplitude of the height value in the frequency spectrum image using Gaussian fitting and drawing a curve, and obtaining a three-dimensional canopy index 3DCI according to the area under the curve; performing voxelization processing on the canopy point cloud, reconstructing a canopy triangular mesh model according to the voxelized canopy point cloud, and obtaining the three-dimensional canopy surface area according to the canopy triangular mesh model. 4.The method of identifying regenerative crop varieties based on UAV LiDAR point cloud for resistance to rolling according to claim 3, wherein, obtaining the three-dimensional canopy surface area according to the canopy triangular mesh model by the following formula: ； where S CS represents the surface area of the three-dimensional crown layer, S CM represents the surface area of the crown layer triangular mesh model, S P represents the sampling area, and k is the surface area coefficient of the crown layer triangular mesh model.

5. The method for identifying a regenerative crop roller- resistant variety based on UAV LiDAR point cloud according to claim 2, characterized in that, calculating the stem inclination angle by the following formula: ； wherein, is the stem inclination angle, H R represents the crop canopy height corresponding to the rolled row crop point, H N represents the crop canopy height corresponding to the non-rolled row crop point.

6. The method for identifying a regenerative crop roller- resistant variety based on UAV LiDAR point cloud according to claim 5, wherein, calculating the stem resilience by the following formula: ； wherein, is the stem resilience calculated for the i-th time after the crop is harvested, are the stem inclination angles obtained by sampling at the times t1 and t2, respectively, in the different growth periods.

7. The method for identifying the varieties of the regenerative crops with the resistance to the rolling pressure based on the unmanned aerial vehicle laser radar point cloud according to any one of claims 1-6, characterized in that, in the step of extracting the rolled row crop points and the non-rolled row crop points from the three-dimensional point cloud data, comprising: after removing the noise points in the three-dimensional point cloud data, dividing the three-dimensional point cloud data into ground points and crop points; dividing the crop points into the rolled row crop points and the non-rolled row crop points.

8. A device for identifying a regenerative crop roller-resistant variety based on unmanned aerial vehicle lidar point cloud, characterized in that, The system comprises: a recognition module configured to acquire three-dimensional point cloud data of a target monitoring area from multi-temporal global images of crops collected by a UAV laser radar device, and extract rolled row crop points and non-rolled row crop points from the three-dimensional point cloud data; an extraction module configured to extract the rolling resistance index of the crops from the rolled row crop points and the non-rolled row crop points, respectively; The identification module is configured to predict the crush resistance of the crop according to the crush resistance indexes corresponding to the crush row crop points and the non-crush row crop points, and determine whether the crop is a crush-resistant variety according to the crush resistance of the crop. The identification module specifically predicts the crush resistance of the crop according to the crush resistance indexes corresponding to the crush row crop points and the non-crush row crop points by the following formula: ； wherein, RD is the crush resistance of the crop, i S is the stem resilience calculated at different times, n is the number of times the stem resilience is calculated, and csR and S csN are the canopy surface area of the crush row crop point and the non-crush row crop point, respectively, R and H N are the crop canopy height of the crush row crop point and the non-crush row crop point, respectively, and a, b, and c are weight coefficients.

9. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor implements the method for identifying the crush-resistant variety of the regenerative crop based on the unmanned aerial vehicle laser radar point cloud according to any one of claims 1 to 7 when executing the program.

Images