Rice canopy height extraction method, device and equipment based on canopy interception compensation mechanism and medium

Abstract

The invention relates to a rice canopy height extraction method and device based on a canopy interception compensation mechanism, equipment and a medium. The method comprises the following steps: acquiring laser radar point cloud data returned from a rice field to construct a laser radar point cloud data set; projecting the laser radar point cloud data on a horizontal plane, dividing the horizontal plane according to grid units of a preset area threshold, and counting the proportion of non-ground points in each grid unit as a canopy laser interception rate; determining a preset compensation function, and correcting a difference value between the rice field digital surface model and the rice field digital elevation model according to the canopy laser interception rate and the compensation function so as to construct a rice canopy height extraction model; and determining the rice canopy height in the to-be-researched rice field area based on the rice canopy height extraction model. The precision of the rice canopy height extraction model can be remarkably improved, and manpower and material resources are greatly saved.

Description

Technical Field

[0001] The present application relates to the field of agricultural production, and in particular to a method for extracting rice canopy height based on a canopy interception compensation mechanism, a corresponding device, an electronic device and a computer-readable storage medium. Background Art

[0002] Thanks to the development of high-throughput sequencing technology, genomic research has made great progress. However, the lag in phenotypic information acquisition technology has made it impossible to synchronously match the data-rich genomic data, which has seriously hindered the breeding process and become the bottleneck of current rice breeding research. Rice is the main raw material for the textile industry and an important economic crop. Plant height, as an important phenotypic trait of rice, is an important indicator in yield prediction, biomass estimation, and lodging resistance. Therefore, it is of great significance to obtain rice crop height over a large area. Traditionally, crop height is obtained through field measurement, which is time-consuming and labor-intensive, so it is unrealistic to obtain height over a large area.

[0003] Passive optical remote sensing images based on drones have been widely used in the field of crop phenotypic monitoring, such as rice boll weight, chlorophyll content, leaf area index, etc. In addition, there are many studies that use visible light cameras on drones to obtain highly overlapping images, detect and match features on images based on the structure from motion (SfM) algorithm, and construct point cloud models to obtain the height of various crops (such as rice, corn, wheat, and soybeans). However, optical remote sensing data is prone to saturation in areas with high biomass or LAI (leaf area index), so it has great limitations in the study of predicting crop structural parameters such as vegetation height. This phenomenon can be alleviated to a certain extent by fusing millimeter-wave radar data, but the overall effect is still not ideal. In addition, digital cameras cannot penetrate the vegetation canopy to obtain ground height data, so accurate crop height estimation generally requires two flight missions, including collecting digital elevation model (DEM) data without vegetation cover and digital surface model (DSM) data with crop cover during the sowing period and after harvest, which greatly reduces the convenience of data acquisition. The root mean square error of this method in rice plant height monitoring is 8.4 cm, and this is when a laser leveler is used before sowing.

[0004] LiDAR has strong penetrating power. Through multiple echo reflections of laser pulses, it can simultaneously record the top coordinates of the ground and the target object. Early LiDARs were usually carried on helicopters or fixed-wing aircraft to measure the height of forest canopies. However, the flight altitude of airborne LiDARs is usually in the hundreds of meters, and the spatial resolution is limited, making it difficult to distinguish between soil and low crops. Ground-based LiDAR systems can obtain three-dimensional point cloud data with millimeter-level accuracy. There are many related studies in the direction of crop height extraction, mainly in the form of robot platforms, vehicle-mounted platforms, and ground fixed platforms. However, the data acquisition efficiency of ground fixed platforms is low, and ground mobile platforms must consider information such as the spacing between rows and plants of field crops and crop height, which is not suitable for land-intensive crops.

[0005] The emergence of UAV-mounted LiDAR systems fills the gap between high-precision but low-efficiency ground-based platforms and large-scale but less detailed manned platforms. Crop height estimation based on LiDAR can be obtained in different ways, such as the percentage method, the method based on the canopy height model (CHM), the method based on point cloud variables, etc. However, when the rice canopy density is extremely high, the error of plant height estimation will increase due to the lack of ground points, especially in the current "short, high-density, early-maturing" rice cultivation mode, the planting density reaches 20 plants / m 2 Even more, the error in extracting rice plant height reaches about 10cm, which makes it extremely difficult to match genomic data.

[0006] To sum up, the existing technology has great limitations in the research on predicting crop structural parameters such as rice vegetation height, the data acquisition efficiency of the ground fixed platform is low, and when the rice canopy density is extremely high, the error of plant height estimation will increase due to the lack of ground points. The applicant has made corresponding explorations to solve these problems. Summary of the invention

[0007] The purpose of the present application is to solve the above-mentioned problems and to provide a rice canopy height extraction method based on a canopy interception compensation mechanism, a corresponding device, an electronic device and a computer-readable storage medium.

[0008] In order to meet the various objectives of this application, this application adopts the following technical solutions:

[0009] A rice canopy height extraction method based on a canopy interception compensation mechanism is proposed to meet one of the purposes of this application, comprising:

[0010] In response to the rice canopy height extraction instruction, the laser radar point cloud data returned from the rice field is obtained to construct a laser radar point cloud data set, and a preset ground point extraction algorithm is called to extract ground points in the rice field from the laser radar point cloud data set;

[0011] Using the ground points in the rice paddy field and the non-ground points in the rice paddy field to construct a paddy field digital surface model, and using the ground points in the rice paddy field to construct a paddy field digital elevation model;

[0012] Projecting the laser radar point cloud data onto a horizontal plane, dividing the horizontal plane into grid units with a preset area threshold, and counting the proportion of non-ground points in each grid unit as a canopy laser interception rate, wherein the canopy laser interception rate represents the proportion of laser radar point cloud data that cannot reach the ground due to obstruction by the rice canopy;

[0013] Determine a preset compensation function, and correct the difference between the paddy field digital surface model and the paddy field digital elevation model according to the canopy laser interception rate and the compensation function to construct a rice canopy height extraction model;

[0014] The rice canopy height in the rice field area to be studied is determined based on the rice canopy height extraction model to complete the extraction of the rice canopy height.

[0015] Optionally, the step of calling a preset ground point extraction algorithm to extract ground points in the rice field from the laser radar point cloud data set includes:

[0016] In response to the ground point extraction instruction, a cubic bounding box is created by using a progressive irregular triangulated network encryption algorithm to wrap the laser radar point cloud data and divide it into multiple areas of n rows and m columns. Initial ground seed points are uniformly selected from the laser point cloud data according to the local minimum elevation principle, and the remaining points are points to be determined, wherein the ground point extraction algorithm is a progressive irregular triangulated network encryption algorithm, and the local minimum elevation principle indicates that the lowest elevation point of the geographical location is found in each area as a ground seed point;

[0017] Constructing an initial irregular triangulated network based on the initial ground seed points, and determining the projection coordinates of all the points to be determined on the horizontal plane to determine the triangular face to which the points to be determined belong in the irregular triangulated network;

[0018] For each triangular face to which the point to be determined belongs in the irregular triangulated network, the slope of the corresponding triangular face is calculated and determined, wherein the slope of the triangular face represents the elevation change rate inside the triangular face;

[0019] If the slope of the triangle surface is lower than the preset maximum slope threshold, calculate and determine a first distance from the point to be determined to the nearest triangle surface, and calculate and determine a first angle between a line between the point to be determined and its nearest triangle surface vertex and a normal vector of the triangle surface;

[0020] If the first angle is less than the preset maximum slope threshold and the first distance is less than the preset maximum height threshold, the point to be judged is regarded as a ground point and included in the ground point set. The above steps are iteratively performed until no new ground points are selected or the preset maximum number of iterations is reached, thereby completing the extraction of the ground points in the rice field from the lidar point cloud data set.

[0021] Optionally, after the step of calculating and determining the slope of the corresponding triangular face to which each of the to-be-determined points belongs in the irregular triangulated network, the method further includes:

[0022] If the slope of the triangular surface exceeds the preset maximum slope threshold, determine the mirror point of the point to be judged, and find the vertex with the largest elevation value in the triangular surface where the mirror point of the point to be judged is located, as the ground point in the rice field, wherein the mirror point represents the point that is symmetrical about a certain axis or plane of the point to be judged.

[0023] Optionally, after the step of projecting the laser radar point cloud data onto a horizontal plane, dividing the horizontal plane into grid units with a preset area threshold, and counting the proportion of non-ground points in each grid unit as the canopy laser interception rate, the method further comprises:

[0024] Detecting whether the canopy laser interception rate exceeds a preset interception rate threshold, if the canopy laser interception rate does not exceed the preset interception rate threshold, calculating the difference between the paddy field digital surface model and the paddy field digital elevation model, and using the difference as the rice canopy height of the rice field, the preset interception rate threshold is 98%;

[0025] If the canopy laser interception rate exceeds the preset interception rate threshold, continue to determine the preset compensation function, and correct the difference between the paddy field digital surface model and the paddy field digital elevation model according to the canopy laser interception rate and the compensation function.

[0026] Optionally, the step of constructing a rice canopy height extraction model includes:

[0027] Acquire a paddy field digital surface model and a paddy field digital elevation model, and calculate and determine the difference between the paddy field digital surface model and the paddy field digital elevation model;

[0028] Determine a preset compensation function, and calculate and determine the product between the compensation function and the canopy laser interception rate;

[0029] The rice canopy height extraction model is constructed based on the sum of the difference between the paddy field digital surface model and the paddy field digital elevation model and the product of the compensation function and the canopy laser interception rate.

[0030] Optionally, after the step of determining the rice canopy height in the rice field area to be studied based on the rice canopy height extraction model, the method further comprises:

[0031] Determine the rice canopy height in the rice field area to be studied, detect whether the rice canopy height in the rice field area to be studied falls within the canopy height threshold corresponding to each growth cycle of rice, and if so, determine and spray the corresponding fertilizer element based on the growth cycle corresponding to the rice canopy height to complete the extraction of the rice canopy height.

[0032] Optionally, the paddy field digital surface model is a DSM model, which includes rice on the paddy field ground and the corresponding elevation information of the ground, and the paddy field digital elevation model is a DEM model, which only includes the ground elevation information;

[0033] The rice canopy height extraction model is a piecewise function, and the expression of the rice canopy height extraction model is:

[0034]

[0035] Among them, the H LiDAR represents the difference between the paddy field digital surface model and the paddy field digital elevation model, and P represents the canopy laser interception rate.

[0036] A rice canopy height extraction device based on a canopy interception compensation mechanism is provided to meet another purpose of the present application, comprising:

[0037] A ground point extraction module is configured to respond to a rice canopy height extraction instruction, obtain laser radar point cloud data returned from a rice field to construct a laser radar point cloud data set, and call a preset ground point extraction algorithm to extract ground points in the rice field from the laser radar point cloud data set;

[0038] An initial model construction module is configured to construct a paddy field digital surface model using ground points in the paddy field and non-ground points in the paddy field, and to construct a paddy field digital elevation model using ground points in the paddy field;

[0039] a laser interception rate determination module, configured to project the laser radar point cloud data onto a horizontal plane, divide the horizontal plane into grid units of a preset area threshold, and count the proportion of non-ground points in each grid unit as a canopy laser interception rate, wherein the canopy laser interception rate represents the proportion of the laser radar point cloud data that cannot reach the ground due to obstruction by the rice canopy;

[0040] a height extraction model construction module, configured to determine a preset compensation function, and to correct the difference between the paddy field digital surface model and the paddy field digital elevation model according to the canopy laser interception rate and the compensation function, so as to construct a rice canopy height extraction model;

[0041] The canopy height extraction module is configured to determine the rice canopy height in the rice field area to be studied based on the rice canopy height extraction model to complete the extraction of the rice canopy height.

[0042] An electronic device provided to meet another purpose of the present application includes a central processing unit and a memory, wherein the central processing unit is used to call and run a computer program stored in the memory to execute the steps of the rice canopy height extraction method based on the canopy interception compensation mechanism described in the present application.

[0043] A computer-readable storage medium is provided to meet another purpose of the present application, which stores a computer program implemented according to the rice canopy height extraction method based on the canopy interception compensation mechanism in the form of computer-readable instructions. When the computer program is called and executed by a computer, the steps included in the corresponding method are executed.

[0044] Compared with the prior art, the present application aims at solving the problems that the prior art has great limitations in the research on predicting crop structural parameters such as rice vegetation height, the data acquisition efficiency of the ground fixed platform is low, and when the rice canopy density is extremely high, the error of plant height estimation will increase due to the lack of ground points. The present application includes but is not limited to the following beneficial effects:

[0045] First, the dense point cloud data obtained by the low-altitude UAV-mounted LiDAR includes rich canopy information and bare soil information, which has great advantages in extracting morphological features such as canopy height. Compared with the aerial image through the motion recovery structure algorithm (SFM) to obtain point cloud data to measure the height, the height obtained by LiDAR is more accurate, and the LiDAR point cloud is easier to obtain the bare soil DEM. Even in the case of extremely high canopy density, satisfactory accuracy can still be obtained through the compensation algorithm in this application, and only one flight is required - there is no need to collect the bare soil DEM before sowing;

[0046] Second, the PTD ground filtering algorithm has obvious advantages over CSF in the case of high canopy density, but the ground may be too high when the number of ground points is extremely low. Combining the canopy laser interception rate compensation to compensate the measurement results can improve this situation to a certain extent and achieve a higher estimation accuracy, providing a solution for obtaining the crop canopy height in a single flight using airborne LiDAR under high planting density.

[0047] Third, compared to the traditional method of collecting bare soil elevation before planting, the method used in this application is easier to use and has lower difficulty in data collection. The results of this application have important practical significance for improving the efficiency of rice breeding research, especially in breeding experiments involving thousands of samples. The data of each plot in traditional breeding experiments are usually determined by manual sampling, which consumes a lot of labor costs and the results are greatly affected by human factors. The method used in this application can generate the mean of the overall height of each plot, which is more representative than the results of manual sampling, and the data collection efficiency is doubled;

[0048] Fourthly, the canopy laser interception rate is highly correlated with the canopy density, so it can be applied to the quantitative inversion of leaf area index (LAI), biomass and other traits. In addition, the fusion application of LiDAR and spectral information will become an inevitable trend in the development of crop phenotypic monitoring related research. LiDAR can collect rich three-dimensional morphological data, while hyperspectral and multispectral can invert the content of certain necessary nutrients in crops, such as protein content, nitrogen content, chlorophyll content, etc. The two complement each other and lay a theoretical foundation for building a high-efficiency, high-throughput, and high-accuracy crop phenotyping acquisition platform;

[0049] Furthermore, even in the case of extremely high canopy density, the present application has achieved high accuracy by eliminating the need for traditional methods to collect bare soil DEM before sowing. The present application can significantly improve the accuracy of the rice canopy height extraction model, greatly saving manpower and material resources. The present application can replace part of the manual measurement work in crop breeding, cultivation and other research, and provides a novel solution for improving variety selection efficiency and optimizing cultivation patterns. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0051] Figure 1 It is a schematic flow chart of the method for extracting the height of a rice canopy based on a canopy interception compensation mechanism in an embodiment of the method for extracting the height of a rice canopy based on a canopy interception compensation mechanism of the present application;

[0052] Figure 2 A schematic diagram of the effect of de-redundancy of laser radar point cloud data in an embodiment of the present application;

[0053] Figure 3 A schematic diagram of constructing an initial irregular triangulated network based on the initial ground seed points selected in the embodiment of the present application;

[0054] Figure 4 This is a schematic diagram of using mirror points to determine ground points in a rice field when the slope of a triangular surface is higher than a preset maximum slope threshold in an embodiment of the present application;

[0055] Figure 5 A schematic diagram of a process for determining a ground point in a rice field when a slope of a triangular surface is lower than a preset maximum slope threshold in an embodiment of the present application;

[0056] Figure 6 This is a schematic diagram showing that all lasers will be intercepted by the rice canopy in the embodiment of the present application, so that the ground points cannot be acquired;

[0057] Figure 7 A schematic diagram of generating a canopy laser interception rate grid map of a study area in an embodiment of the present application;

[0058] Figure 8 A schematic diagram of the matching of CHM-PTD and CHM-CSF with ground measured samples in an embodiment of the present application;

[0059] Fig. 9 A flow chart showing the difference between the ground grids obtained by the two filtering algorithms PTD and CSF in the embodiment of the present application;

[0060] Fig.10 A schematic diagram showing the matching degree between CHM-PTD and data set P70 and the residual graph in the embodiment of the present application;

[0061] Fig.11 A schematic diagram of a flow chart showing the matching degree between CHM-PTD and the data set P90 and the residual graph in the embodiment of the present application;

[0062] Fig.12 A schematic diagram of the process of showing the matching degree between CHM-PTD and the data set P95 and the residual graph in the embodiment of the present application;

[0063] Fig.13 A schematic diagram of a flow chart showing the matching degree between CHM-PTD and the data set P97 and the residual graph in the embodiment of the present application;

[0064] Fig.14 A schematic diagram of a flow chart showing the matching degree between CHM-PTD and the data set P98 and the residual graph in the embodiment of the present application;

[0065] Fig.15 A schematic diagram of a flow chart showing the matching degree between CHM-PTD and the data set P99 and the residual graph in the embodiment of the present application;

[0066] Fig.16 This is a schematic diagram of a residual graph of a data set P90 in an embodiment of the present application;

[0067] Fig.17 It is a schematic diagram of the process in the embodiment of this application;

[0068] Fig.18 A schematic diagram showing the matching degree between the estimation result and the sampling result after the canopy laser interception rate compensation in the embodiment of the present application;

[0069] Fig.19 A schematic diagram of a plant height distribution diagram at a field scale in an embodiment of the present application;

[0070] Fig. 20 It is a principle block diagram of a rice canopy height extraction device based on a canopy interception compensation mechanism in an embodiment of the present application;

[0071] Fig.21 It is a schematic diagram of the structure of the computer device in the embodiment of the present application. DETAILED DESCRIPTION

[0072] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application, and cannot be interpreted as limiting the present application.

[0073] It will be understood by those skilled in the art that, unless expressly stated, the singular forms "one", "said", and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present application refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof. It should be understood that when we refer to an element as being "connected" or "coupled" to another element, it may be directly connected or coupled to the other element, or there may be an intermediate element. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The term "and / or" used herein includes all or any unit and all combinations of one or more associated listed items.

[0074] It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as those generally understood by those skilled in the art to which this application belongs. It should also be understood that terms such as those defined in general dictionaries should be understood to have meanings consistent with those in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless specifically defined as here.

[0075] It will be understood by those skilled in the art that the "client", "terminal" and "terminal device" used herein include both devices with wireless signal receivers, which are devices with only wireless signal receivers without transmission capabilities, and devices with receiving and transmitting hardware, which are devices with receiving and transmitting hardware capable of two-way communication on a two-way communication link. Such devices may include: cellular or other communication devices such as personal computers, tablet computers, which have single-line displays or multi-line displays or cellular or other communication devices without multi-line displays; PCS (Personal Communications Service, personal communication system), which can combine voice, data processing, fax and / or data communication capabilities; PDA (Personal Digital Assistant, personal digital assistant), which may include a radio frequency receiver, pager, Internet / intranet access, web browser, notepad, calendar and / or GPS (Global Positioning System, global positioning system) receiver; conventional laptop and / or palmtop computers or other devices, which have and / or include a conventional laptop and / or palmtop computer or other device with and / or including a radio frequency receiver. The "client", "terminal" and "terminal device" used herein may be portable, transportable, installed in a vehicle (air, sea and / or land), or suitable for and / or configured to run locally, and / or in a distributed form, at any other location on the earth and / or in space. The "client", "terminal" and "terminal device" used herein may also be a communication terminal, an Internet terminal, a music / video playing terminal, for example, a PDA, a MID (Mobile Internet Device) and / or a mobile phone with a music / video playing function, or a smart TV, a set-top box and other devices.

[0076] The hardware referred to by the names such as "server", "client", and "service node" in this application is essentially an electronic device with the equivalent capabilities of a personal computer. It is a hardware device with the necessary components revealed by the von Neumann principle, such as a central processing unit (including an arithmetic unit and a controller), a memory, an input device, and an output device. The computer program is stored in its memory, and the central processing unit calls the program stored in the external memory into the internal memory for execution, executes the instructions in the program, and interacts with the input and output devices to complete specific functions.

[0077] It should be pointed out that the concept of "server" referred to in this application can also be extended to the case of server clusters. According to the network deployment principle understood by those skilled in the art, the servers should be logically divided. In physical space, these servers can be independent of each other but can be called through interfaces, or integrated into a physical computer or a set of computer clusters. Those skilled in the art should understand this flexibility, and should not use it to restrict the implementation of the network deployment method of this application.

[0078] Unless expressly specified, one or more technical features of the present application can be deployed on a server for implementation and accessed by a client through a remote call to obtain an online service interface provided by the server, or can be directly deployed and run on a client for access.

[0079] The neural network models referenced or may be referenced in this application, unless expressly specified, can be deployed on a remote server and remotely called on the client, or can be deployed and directly called on a client with sufficient device capabilities. In some embodiments, when it runs on the client, its corresponding intelligence can be obtained through transfer learning to reduce the requirements for the client's hardware operating resources and avoid excessive occupation of the client's hardware operating resources.

[0080] Unless explicitly specified, the various data involved in this application can be stored remotely on a server or on a local terminal device, as long as it is suitable for being called by the technical solution of this application.

[0081] Those skilled in the art should be aware that, although the various methods of the present application are described based on the same concept and thus present commonality to each other, unless otherwise specified, these methods can be independently executed. Similarly, for each embodiment disclosed in the present application, they are all proposed based on the same inventive concept, therefore, concepts with the same expression, and concepts that are appropriately changed for convenience despite different expressions, should be understood as equivalent.

[0082] Unless the mutually exclusive relationship between the embodiments to be disclosed in the present application is explicitly stated, the relevant technical features involved in the embodiments can be cross-combined to flexibly construct new embodiments, as long as such combination does not deviate from the creative spirit of the present application and can meet the needs of the prior art or solve the problems of the prior art.

[0083] See also Figure 1 In one embodiment, the rice canopy height extraction method based on the canopy interception compensation mechanism of the present application includes:

[0084] Step S10, responding to the rice canopy height extraction instruction, acquiring the laser radar point cloud data returned from the rice field to construct a laser radar point cloud data set, and calling a preset ground point extraction algorithm to extract the ground points in the rice field from the laser radar point cloud data set;

[0085] The rice canopy height extraction system in the terminal device can respond to the rice canopy height extraction instruction, obtain the laser radar point cloud data returned by the LiDAR system in the drone from the rice field to construct a laser radar point cloud data set, and call a preset ground point extraction algorithm to extract the ground points in the rice field from the laser radar point cloud data set. The ground point extraction algorithm is a progressive irregular triangulated network encryption algorithm.

[0086] In some embodiments, the DJI M300 RTK aircraft equipped with the ZENMUSE L1 lidar can be used to collect lidar point cloud data. The aircraft integrates a differential positioning module and has strong anti-magnetic interference and precise positioning capabilities. The Zenmuse L1 integrates Livox lidar modules, high-precision inertial navigation, mapping cameras, three-axis gimbals and other modules, which can achieve all-weather, high-efficiency real-time 3D data acquisition and high-precision post-processing reconstruction in complex scenes.

[0087] In some embodiments, after obtaining the laser radar point cloud data returned from the rice field, the laser radar point cloud data is preprocessed. First, in order to avoid the influence of point cloud noise on the test results, the original point cloud is subjected to noise filtering to remove isolated points. Since the overlap rate of the flight strips is 10% during data collection, the point cloud density in the overlapping area is significantly increased, which will affect the subsequent calculation of the canopy laser interception rate, so it is necessary to delete the redundant points in the overlapping area. Usually, the measurement error of the laser point is proportional to the laser scanning angle. The closer to the edge of the strip, the greater the point cloud error. For the redundant part of the overlapping area of ​​the airborne laser radar flight strips, the de-redundancy function of the Lidar 360 software is used. According to the complete route information, the point cloud accuracy decreases with the increase of the scanning angle as the theoretical basis, thereby removing the point cloud with a larger route scanning angle in the overlapping area of ​​the flight strip, and retaining the points with higher quality close to the center of the flight strip, thereby eliminating the redundancy of the aerial photography data between the flight strips. The point cloud data collected by the two flight strips are colored red and green respectively. The effect of de-redundancy is as follows. Figure 2 shown.

[0088] In some embodiments, the ground point extraction algorithm includes a mature ground point extraction algorithm such as Progressive TIN Densification (PTD) algorithm and Cloth Simulation Filter (CSF) algorithm. Due to the high canopy density environment in this application, PTD (R 2=0.79) was significantly better than CSF (R 2 =0.29), it can be seen from the residual graph that the latter has a significant underestimation phenomenon in the estimation, and the difference in the ground grids obtained by the two filtering algorithms can be clearly seen from the cross-sectional view. It is preliminarily inferred that the "pointed" shape formed by the laser penetrating the canopy can easily pass through the gaps between the cloth particles, resulting in the result obtained by the gravity cloth algorithm filtering is not the ground but the lower surface of the canopy. Therefore, the Progressive TIN Densification (PTD) algorithm is more suitable for extracting ground points in the rice paddy field from the lidar point cloud data set, and the Progressive TIN Densification (PTD) algorithm is used as the ground point extraction algorithm of this application, which does not constitute a limitation on this application.

[0089] In some embodiments, the CSF algorithm is another popular ground point extraction method, which simulates the free fall process of cloth and identifies ground points by optimizing the tension and position of cloth. CSF can better extract ground information under complex terrain and vegetation occlusion.

[0090] In some embodiments, the step of calling a preset ground point extraction algorithm to extract ground points in the rice paddy field from the laser radar point cloud data set includes:

[0091] Step S101, using a progressive irregular triangulated network encryption algorithm to create a cubic bounding box to wrap the laser radar point cloud data and divide it into multiple areas of n rows and m columns, and evenly select initial ground seed points from the laser point cloud data according to the principle of local minimum elevation, and the remaining points are points to be determined, wherein the ground point extraction algorithm is a progressive irregular triangulated network encryption algorithm, and the local minimum elevation principle represents finding the lowest elevation point of the geographical location in each area as the ground seed point;

[0092] Step S102: construct an initial irregular triangulated network based on the initial ground seed points, and determine the projection coordinates of all the points to be determined on the horizontal plane to determine the triangular face to which the points to be determined belong in the irregular triangulated network;

[0093] Step S103, for each triangular face to which the to-be-determined point belongs in the irregular triangulated network, calculating and determining the corresponding triangular face slope, wherein the triangular face slope represents the elevation change rate inside the triangular face;

[0094] Step S104: if the slope of the triangle surface is lower than the preset maximum slope threshold, calculate and determine a first distance from the point to be determined to the nearest triangle surface, and calculate and determine a first angle between a line between the point to be determined and its nearest triangle surface vertex and a normal vector of the triangle surface;

[0095] Step S105: If the first angle is less than a preset maximum slope threshold and the first distance is less than a preset maximum height threshold, the point to be determined is regarded as a ground point and included in the ground point set. The above steps are iteratively performed until no new ground points are selected or the preset maximum number of iterations is reached, thereby completing the extraction of the ground points in the rice field from the lidar point cloud data set.

[0096] Furthermore, for each triangular face to which the to-be-determined point belongs in the irregular triangulated network, after the step of calculating and determining the slope of the corresponding triangular face, the method further comprises:

[0097] If the slope of the triangular surface exceeds the preset maximum slope threshold, determine the mirror point of the point to be judged, and find the vertex with the largest elevation value in the triangular surface where the mirror point of the point to be judged is located, as the ground point in the rice field, wherein the mirror point represents the point that is symmetrical about a certain axis or plane of the point to be judged.

[0098] Specifically, the progressive irregular triangulated network encryption algorithm is a classic ground filtering algorithm that achieved the best results in the filtering algorithm test conducted by ISPRS (International Society for Photogrammetry and Remote Sensing). The main idea of ​​the progressive irregular triangulated network encryption algorithm is to construct the initial irregular triangulated network (TIN) through the initial ground seed points, and then iteratively judge the remaining points. If they meet the requirements, the ground point set is added to make the TIN constructed in the next iteration more refined. The process is iteratively executed until no new ground points appear or the maximum number of iterations is reached. The specific steps include:

[0099] Step (1): Due to the multipath effect of the LiDAR system or the error of the laser ranging system, there is a probability that a laser point with an elevation much lower than other surrounding points will be formed. If it is not removed, when the ground seed points are selected using the above assumptions, the ground seed points obtained will contain these low points, thereby introducing errors to the subsequent filtering.

[0100] Step (2), define a cubic bounding box to wrap the point cloud data and divide it into multiple areas of n rows and m columns, select ground seed points evenly from the original point cloud according to the principle of local minimum elevation, and the remaining points are points to be determined, wherein the values ​​of n and m are calculated by the formula:

[0101]

[0102] n and m can be obtained from the above formula, where x and y are the length and width of the cube boundary, b max In the present application, it is set to the width of the ridge, and the ridge refers to the raised soil-buried strip formed during farming, which is used for planting crops or other purposes.

[0103] Step (3), please refer to Figure 3 , using the initial ground seed points selected in step (2) to construct an initial irregular triangulated network (TIN), selecting ground points from the points to be determined for iterative encryption, for all the points to be determined, determining the projection coordinates of the points to be determined on the horizontal plane to determine the triangular face to which the points to be determined belong in the irregular triangulated network; for each triangular face to which the points to be determined belong in the irregular triangulated network, calculating and determining the corresponding triangular face slope, wherein the triangular face slope represents the elevation change rate inside the triangular face;

[0104] For further information, see Figure 4 If the slope of the triangular surface exceeds the preset maximum slope threshold, determine the mirror point of the point to be determined, and find the vertex P with the largest elevation value in the triangular surface where the mirror point of the point to be determined P is located. h (x m ,y m , z m ), as the ground point in the rice field, wherein the mirror point represents a point symmetrical to the point to be determined about a certain axis or plane, and the calculation formula of each coordinate value of the mirror point of the point to be determined P includes:

[0105] x mirror =2x m -x P ,

[0106] y mirror =2y m -y p ,

[0107] z mirror =2z P ,

[0108] Among them, x mirror Represents the horizontal coordinate value of the mirror point of the point to be judged P, y mirror Indicates the vertical coordinate value of the mirror point of the point to be judged P, z mirror Indicates the vertical axis coordinate value of the mirror point of the point to be judged P, x p ,y p 、z p They respectively represent the horizontal axis coordinate value, the vertical axis coordinate value and the vertical axis coordinate value of the point P to be determined.

[0109] Specifically, if the calculated slope of the triangular surface exceeds the preset maximum slope threshold α1, it means that the terrain change in the area is too steep or abnormal. The mirror point is used to judge and deal with abnormal points or terrain features that do not meet expectations. At the same time, ensure that the classification of the point is consistent with the classification of the mirror point to ensure data consistency and reliability when dealing with abnormal or uncertain terrain data. Find the vertex P with the largest elevation value in the triangular surface where the mirror point of the point to be judged P is located h (x m ,y m , z m ), as a ground point in the rice paddy field.

[0110] For further information, see Figure 5 , if the slope of the triangular surface is lower than the preset maximum slope threshold α1, calculate and determine the first distance h from the point to be determined to the nearest triangular surface, and calculate and determine the first angle a between the line between the point to be determined and its nearest triangular surface vertex and the normal vector of the triangular surface; if the first angle a is less than the preset maximum slope threshold α1 and the first distance h is less than the preset maximum height threshold, the point to be determined is regarded as a ground point and included in the ground point set. The above steps are iteratively performed until no new ground points are selected or the preset maximum number of iterations is reached, thereby completing the extraction of the ground points in the rice field from the lidar point cloud data set.

[0111] Step S20, using the ground points in the rice field and the non-ground points in the rice field to construct a digital surface model of the rice field, and using the ground points in the rice field to construct a digital elevation model of the rice field;

[0112] After calling the preset ground point extraction algorithm to extract the ground points in the rice field from the lidar point cloud data set, it is not difficult to understand that the non-ground points in the rice field are also determined accordingly, and the ground points in the rice field and the non-ground points in the rice field are used to construct a digital surface model of the rice field, and the ground points in the rice field are used to construct a digital elevation model of the rice field; wherein the digital surface model of the rice field is a DSM model, which includes the rice on the ground of the rice field and the corresponding elevation information of the ground, and the digital elevation model of the rice field is a DEM model, which only includes ground elevation information.

[0113] In some embodiments, constructing DSM (Digital Surface Model) and DEM (Digital Elevation Model) models is a common task in Geographic Information System (GIS) and terrain analysis, especially in LiDAR point cloud data processing.

[0114] The DSM model is a digital surface model that reflects the elevation data and shape of rice plants and the ground on the ground of the rice paddy field. An initial surface model is constructed from the LiDAR point cloud data, including all ground points and non-ground points, to generate a digital surface model. This model shows the elevation data of rice plants and the ground on the ground of the rice paddy field.

[0115] The DEM model (Digital Elevation Model) is a digital elevation model that only contains ground elevation data, without considering rice plants on the ground of the rice field. The digital elevation model of the rice field is constructed using ground points in the rice field. This model focuses on the geometric shape and elevation changes of the rice ground terrain.

[0116] Step S30, projecting the laser radar point cloud data onto a horizontal plane, dividing the horizontal plane into grid units with a preset area threshold, and counting the proportion of non-ground points in each grid unit as a canopy laser interception rate, wherein the canopy laser interception rate represents the proportion of the laser radar point cloud data that cannot reach the ground due to obstruction by the rice canopy;

[0117] A digital surface model of the rice field is constructed using the ground points in the rice field and the non-ground points in the rice field. After the digital elevation model of the rice field is constructed using the ground points in the rice field, the lidar point cloud data is projected onto a horizontal plane, and the horizontal plane is divided into grid units according to a preset area threshold. The proportion of non-ground points in each grid unit is counted as the canopy laser interception rate, wherein the canopy laser interception rate represents the proportion of lidar point cloud data that cannot reach the ground due to obstruction by the rice canopy; wherein the preset area threshold is 1 square meter, and so on.

[0118] In some embodiments, when acquiring terrain data, optical radar technology is often unable to directly detect the ground due to obstruction by vegetation. Especially in areas with dense vegetation coverage, the laser beam may be completely intercepted by the vegetation canopy, resulting in the inability to correctly detect and record ground points in these areas. In this case, the difference between the generated digital surface model (DSM, which contains elevation information of vegetation and ground) and the digital elevation model (DEM, which only contains ground elevation information) may be large, thus affecting the accurate estimation of terrain features.

[0119] See also Figure 6When the gap between rice canopies is smaller than the laser density of the LiDAR system, all lasers will be intercepted by the rice canopy and the ground points cannot be obtained, which will cause the interpolated ground grid to be too high, affecting the estimation result. The canopy laser interception rate characterizes the rice canopy density, that is, it is used to measure the proportion of LiDAR point cloud data that cannot reach the ground due to the obstruction of the rice canopy. The canopy laser interception rate is positively correlated with the rice canopy density, that is, within a certain canopy density threshold range, the greater the canopy density, the greater the proportion of LiDAR point cloud data that cannot reach the ground due to the obstruction of the rice canopy; when the canopy density is smaller, the proportion of LiDAR point cloud data that cannot reach the ground due to the obstruction of the rice canopy is smaller. Therefore, the canopy laser interception rate is used to compensate the estimation result to obtain a more accurate estimation accuracy.

[0120] Further, after the step of projecting the laser radar point cloud data onto a horizontal plane, dividing the horizontal plane into grid units with a preset area threshold, and counting the proportion of non-ground points in each grid unit as the canopy laser interception rate, the method further comprises:

[0121] Step S301, detecting whether the canopy laser interception rate exceeds a preset interception rate threshold, if the canopy laser interception rate does not exceed the preset interception rate threshold, calculating the difference between the paddy field digital surface model and the paddy field digital elevation model, and taking it as the rice canopy height of the paddy field, wherein the preset interception rate threshold is 98%;

[0122] Step S302: if the canopy laser interception rate exceeds a preset interception rate threshold, continue to determine a preset compensation function, and correct the difference between the paddy field digital surface model and the paddy field digital elevation model according to the canopy laser interception rate and the compensation function.

[0123] In order to save computing resources and improve the calculation efficiency of the rice canopy height of the rice field, it is detected whether the canopy laser interception rate exceeds the preset interception rate threshold. If the canopy laser interception rate does not exceed the preset interception rate threshold, the difference between the rice field digital surface model and the rice field digital elevation model is calculated and used as the rice canopy height of the rice field.

[0124] Step S40, determining a preset compensation function, and correcting the difference between the paddy field digital surface model and the paddy field digital elevation model according to the canopy laser interception rate and the compensation function, so as to construct a rice canopy height extraction model;

[0125] The laser radar point cloud data is projected onto a horizontal plane, the horizontal plane is divided into grid units with a preset area threshold, the proportion of non-ground points in each grid unit is counted as the canopy laser interception rate, a preset compensation function is determined, and the difference between the paddy field digital surface model and the paddy field digital elevation model is corrected according to the canopy laser interception rate and the compensation function to construct a rice canopy height extraction model;

[0126] Specifically, the steps of constructing the rice canopy height extraction model include:

[0127] Step S401, obtaining a paddy field digital surface model and a paddy field digital elevation model, and calculating and determining a difference between the paddy field digital surface model and the paddy field digital elevation model;

[0128] Step S402, determining a preset compensation function, and calculating and determining the product between the compensation function and the canopy laser interception rate;

[0129] Step S403, constructing the rice canopy height extraction model based on the sum of the difference between the paddy field digital surface model and the paddy field digital elevation model and the product of the compensation function and the canopy laser interception rate.

[0130] More specifically, the laser radar point cloud data acquired by the LiDAR (optical radar) system is projected onto a horizontal plane, and the entire horizontal plane is divided into grids of 1 square meter. The proportion of non-ground points (laser points blocked by vegetation) in each grid unit is counted as an estimated value of the canopy laser interception rate P to determine the canopy laser interception rate P. The calculated canopy laser interception rate P is presented in the form of a raster map for terrain analysis and subsequent agricultural management or environmental planning decision support. A preset compensation function α is used to compare the paddy field digital surface model with the paddy field digital elevation model in combination with the canopy laser interception rate P. The difference between the models is corrected to obtain more accurate ground elevation information; a paddy field digital surface model and a paddy field digital elevation model are obtained, and the difference between the paddy field digital surface model and the paddy field digital elevation model is calculated and determined; a preset compensation function is determined, and the product between the compensation function and the canopy laser interception rate is calculated and determined; based on the sum of the difference between the paddy field digital surface model and the paddy field digital elevation model and the product between the compensation function and the canopy laser interception rate, the rice canopy height extraction model is constructed, and the expression of the rice canopy height extraction model is as follows:

[0131] H=H LiDAR +f(α)·P,

[0132] Wherein, H represents the height of rice canopy, and H LiDARrepresents the difference between the paddy field digital surface model and the paddy field digital elevation model, P represents the canopy laser interception rate, and f(α) represents the compensation function.

[0133] In some embodiments, the rice canopy height extraction model is a piecewise function, and the expression of the rice canopy height extraction model is:

[0134]

[0135] Wherein, H represents the height of rice canopy, and H LiDAR represents the difference between the paddy field digital surface model and the paddy field digital elevation model, and P represents the canopy laser interception rate.

[0136] In some embodiments, see Figure 7 The canopy laser interception rate represents the proportion of laser beams that are blocked by the vegetation canopy and cannot reach the ground during the LiDAR laser scanning process. This proportion can be calculated by analyzing the LiDAR data, especially by projecting the point cloud on the xoy horizontal plane and then dividing the entire area into a square grid of 1 square meter, and finally generating a canopy laser interception rate raster map of the study area.

[0137] Step S50: determining the rice canopy height in the rice field area to be studied based on the rice canopy height extraction model to complete the extraction of the rice canopy height.

[0138] After the rice canopy height extraction model is constructed, the rice canopy height in the rice field area to be studied is determined based on the rice canopy height extraction model to complete the extraction of the rice canopy height. The extracted rice canopy height data is used for subsequent analysis and applications, such as agricultural management, establishment of crop growth models, etc. These data can provide important spatial information support for agricultural decision-making and research.

[0139] In some embodiments, after the step of determining the rice canopy height in the rice field area to be studied based on the rice canopy height extraction model, the method includes:

[0140] Determine the rice canopy height in the rice field area to be studied, detect whether the rice canopy height in the rice field area to be studied falls within the canopy height threshold corresponding to each growth cycle of rice, and if so, determine and spray the corresponding fertilizer element based on the growth cycle corresponding to the rice canopy height to complete the extraction of the rice canopy height.

[0141] It can be seen from the above embodiments that the constructed rice canopy height extraction model is used to determine the rice canopy height in the rice field area to be studied. The rice canopy height varies with the growth stage, such as the vegetative growth stage, jointing stage, flowering stage and grain filling stage. The canopy height will be different, and the canopy height threshold can be determined according to agricultural research or growth model; the extracted canopy height is compared and classified with the predetermined canopy height threshold, and the growth cycle stage of the rice is currently determined according to the threshold range in which the canopy height falls, and the fertilizer element demand suitable for the growth stage is determined according to the detected rice growth cycle. For example, at different growth stages, the demand and proportion of nitrogen, phosphorus, potassium and other fertilizers for rice are different; according to relevant agricultural guidance or professional advice, formulate appropriate fertilizer application plans. According to the previous decision, implement corresponding fertilizer application measures; this may require precise fertilization in the field to ensure that rice can obtain appropriate nutritional support at each growth stage to promote healthy growth and high yield. Regularly monitor the growth of rice and changes in canopy height. Adjust fertilizer application strategies based on monitoring results to ensure that the nutritional needs of rice are met throughout the growth cycle.

[0142] Through the above steps, canopy height information can be effectively used to optimize fertilizer management of rice, thereby improving yield and quality. This refined management method is conducive to the sustainable development of agricultural production and the improvement of resource utilization efficiency.

[0143] It can be seen from the above embodiments that, compared with the prior art, the present application aims at the problems that the prior art has great limitations in the research on predicting crop structural parameters such as rice vegetation height, the data acquisition efficiency of the ground fixed platform is low, and when the rice canopy density is extremely high, the error of plant height estimation will increase due to the lack of ground points. The present application includes but is not limited to the following beneficial effects:

[0144] First, the dense point cloud data obtained by the low-altitude UAV-mounted LiDAR includes rich canopy information and bare soil information, which has great advantages in extracting morphological features such as canopy height. Compared with the aerial image through the motion recovery structure algorithm (SFM) to obtain point cloud data to measure the height, the height obtained by LiDAR is more accurate, and the LiDAR point cloud is easier to obtain the bare soil DEM. Even in the case of extremely high canopy density, satisfactory accuracy can still be obtained through the compensation algorithm in this application, and only one flight is required - there is no need to collect the bare soil DEM before sowing;

[0145] Second, the PTD ground filtering algorithm has obvious advantages over CSF in the case of high canopy density, but the ground may be too high when the number of ground points is extremely low. Combining the canopy laser interception rate compensation to compensate the measurement results can improve this situation to a certain extent and achieve a higher estimation accuracy, providing a solution for obtaining the crop canopy height in a single flight using airborne LiDAR under high planting density.

[0146] Third, compared to the traditional method of collecting bare soil elevation before planting, the method used in this application is easier to use and has lower difficulty in data collection. The results of this application have important practical significance for improving the efficiency of rice breeding research, especially in breeding experiments involving thousands of samples. The data of each plot in traditional breeding experiments are usually determined by manual sampling, which consumes a lot of labor costs and the results are greatly affected by human factors. The method used in this application can generate the mean of the overall height of each plot, which is more representative than the results of manual sampling, and the data collection efficiency is doubled;

[0147] Fourthly, the canopy laser interception rate is highly correlated with the canopy density, so it can be applied to the quantitative inversion of leaf area index (LAI), biomass and other traits. In addition, the fusion application of LiDAR and spectral information will become an inevitable trend in the development of crop phenotypic monitoring related research. LiDAR can collect rich three-dimensional morphological data, while hyperspectral and multispectral can invert the content of certain necessary nutrients in crops, such as protein content, nitrogen content, chlorophyll content, etc. The two complement each other and lay a theoretical foundation for building a high-efficiency, high-throughput, and high-accuracy crop phenotyping acquisition platform;

[0148] Furthermore, even in the case of extremely high canopy density, the present application has achieved high accuracy by eliminating the need for traditional methods to collect bare soil DEM before sowing. The present application can significantly improve the accuracy of the rice canopy height extraction model, greatly saving manpower and material resources. The present application can replace part of the manual measurement work in crop breeding, cultivation and other research, and provides a novel solution for improving variety selection efficiency and optimizing cultivation patterns.

[0149] Based on any of the above embodiments, this application uses the determination coefficient R 2 The performance of the plant height estimation model was evaluated by two statistical indicators: root mean square error (RMSE) and root mean square error (RMSE).

[0150] In order to explore the influence of canopy density on the accuracy of plant height estimation by LiDAR system, this application produced 6 data sets from the total sample according to the canopy laser interception rate classification, including data set P70, data set P90, data set P95, data set P97, data set P98 and data set P99, which represent different rice canopy densities respectively. The mean and standard deviation (Std) of each data set are shown in Table 1:

[0151] Table 1 Mean and standard deviation of each data set

[0152]

[0153] In some embodiments, the performance of two point cloud ground filtering algorithms, Progressive TINDensification (PTD) algorithm and Cloth Simulation Filter (CSF) algorithm, on the data collected by this application is tested. The non-ground points and ground points are rasterized along the Z axis to obtain DSM model and DEM model respectively. The difference between the two is subtracted to obtain CHM. The matching of CHM-PTD and CHM-CSF with the ground measured samples is shown in the figure. Figure 8 shown.

[0154] The results show that the PTD algorithm (R 2 =0.79) is significantly better than the CSF algorithm (R 2 =0.29), it can be seen from the residual diagram that the latter is significantly underestimated in the estimation. Fig. 9 As shown in the figure, the difference between the ground meshes obtained by the two filtering algorithms can be clearly seen from the cross-sectional view. It is preliminarily inferred that the "spike" formed by the laser penetrating the canopy can easily pass through the gaps between the cloth particles, resulting in the result obtained by the CSF algorithm filtering not being the ground but the lower surface of the canopy. Therefore, the subsequent application will be based on the ground points in the lidar point cloud data extracted by the Progressive TIN Densification (PTD) algorithm.

[0155] In some embodiments, Figures 10 to 15 The matching degree and residual graph of CHM-PTD with the six different canopy density datasets in Table 1 are shown. From the matching of dataset P99 and the residual graph, it can be seen that when the canopy interception rate is greater than 99%, the estimation accuracy of LiDAR decreases sharply (R 2=0.79, RMSE = 17.22), and there is an overall trend of being smaller, which directly reduces the estimation accuracy of the overall sample. Therefore, the f(α) compensation function is introduced to estimate and compensate for the positions where the laser canopy rate is greater than 99%, to compensate for the low CHM caused by the inability of the laser to penetrate the canopy.

[0156] In some embodiments, in order to find out the optimal compensation coefficient under different laser canopy interception rates, αP is added to the estimated value of the original plant height as compensation. Table 2 shows the rice canopy height extraction model and R before and after compensation. 2 Compared with RMSE, it can be seen that the compensation estimation method has significantly improved the estimation accuracy in datasets P98, P99 and the overall sample, while the accuracy improvement of the compensation method on other datasets is not obvious, and even negative optimization occurs in some datasets.

[0157] Table 2 Rice canopy height extraction model and R before and after compensation 2 and RMSE

[0158]

[0159] Therefore, it is not appropriate to use a global unified compensation coefficient, which will lead to negative optimization in areas with low canopy interception rate, increasing the estimation error. Therefore, it is a more reasonable solution to determine the compensation coefficient based on the canopy laser interception rate P. According to the results in Table 2, the accuracy difference before and after compensation is the largest in areas with canopy interception rate above 99%, but the P value has only a 1% difference, so the linear function may not be the best fitting method. Please refer to Fig.16 From the residual graph of data set P90, it can be seen that the estimation error tends to increase rapidly when P approaches 100%, which is very close to the trend of the power function on (0,1]. The trial and error method is used to determine the power of P in the rice canopy height extraction model. The results are shown in Fig.17 As shown, the optimal value is reached when the power of P is 100. 2 is 0.94. Finally, the rice canopy height extraction model is defined as a piecewise function, and its expression is as follows

[0160]

[0161] In some embodiments, see Fig.18 , Fig.18 The figure shows the matching degree between the estimated results and the sampling results after using the canopy laser interception rate compensation, and compares them with those before compensation. 2 And RMSE have been greatly improved, reaching 13.92% and 45.41% respectively.

[0162] In some embodiments, see Fig.19 , Fig.19 By using GIS technology to map the output values ​​of the rice canopy height extraction model to their corresponding geographical locations, a plant height distribution map at the field scale can be obtained, which is used to observe the distribution of height differences in the field.

[0163] As can be seen from the above embodiments, the dense point cloud data obtained by the low-altitude drone-mounted LiDAR includes rich canopy information and bare soil information, and has great advantages in extracting morphological features such as canopy height. Compared with the aerial image obtained by motion recovery structure algorithm (SFM) to obtain point cloud data to measure the height, the height obtained by LiDAR is more accurate, and the LiDAR point cloud is easier to obtain bare soil DEM. Even in the case of extremely high canopy density, satisfactory accuracy can still be obtained through the compensation algorithm in this application, and only one flight is required-no need to collect bare soil DEM before sowing.

[0164] Furthermore, the PTD ground filtering algorithm has obvious advantages over CSF in the case of high canopy density, but the ground may be too high when the number of ground points is extremely low. Combining the canopy laser interception rate compensation to compensate the measurement results can improve this situation to a certain extent and achieve a higher estimation accuracy (R 2 =0.90, RMSE =6.18), which provides a solution for obtaining crop canopy height in a single flight using airborne LiDAR under high planting density.

[0165] See also Fig. 20, a rice canopy height extraction device based on a canopy interception compensation mechanism is provided to meet one of the purposes of the present application, including a ground point extraction module 1100, an initial model construction module 1200, a laser interception rate determination module 1300, a height extraction model construction module 1400 and a canopy height extraction module 1500. Among them, the ground point extraction module 1100 is configured to respond to a rice canopy height extraction instruction, obtain the laser radar point cloud data returned from the rice paddy field to construct a laser radar point cloud data set, and call a preset ground point extraction algorithm to extract the ground points in the rice paddy field from the laser radar point cloud data set; the initial model construction module 1200 is configured to use the ground points in the rice paddy field and the non-ground points in the rice paddy field to construct a rice field digital surface model, and use the ground points in the rice paddy field to construct a rice field digital elevation model; the laser interception rate determination module 1300 is configured to project the laser radar point cloud data on a horizontal plane and divide it into grid units according to a preset area threshold. The horizontal plane counts the proportion of non-ground points in each grid unit as the canopy laser interception rate, wherein the canopy laser interception rate represents the proportion of lidar point cloud data that cannot reach the ground due to obstruction of the rice canopy; the height extraction model construction module 1400 is configured to determine a preset compensation function, and according to the canopy laser interception rate and the compensation function, correct the difference between the paddy field digital surface model and the paddy field digital elevation model to construct a rice canopy height extraction model; the canopy height extraction module 1500 is configured to determine the rice canopy height in the paddy field area to be studied based on the rice canopy height extraction model to complete the extraction of the rice canopy height.

[0166] Based on any embodiment of this application, please refer to Fig.21 Another embodiment of the present application further provides an electronic device, which can be implemented by a computer device, such as Fig.21 As shown, a schematic diagram of the internal structure of a computer device. The computer device includes a processor, a computer-readable storage medium, a memory, and a network interface connected via a system bus. Among them, the computer-readable storage medium of the computer device stores an operating system, a database, and computer-readable instructions. The database may store a control information sequence. When the computer-readable instructions are executed by the processor, the processor can implement a rice canopy height extraction method based on a canopy interception compensation mechanism. The processor of the computer device is used to provide computing and control capabilities to support the operation of the entire computer device. The memory of the computer device may store computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor can execute the rice canopy height extraction method based on a canopy interception compensation mechanism of the present application. The network interface of the computer device is used to connect and communicate with a terminal. Those skilled in the art will understand that Fig.21The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0167] In this embodiment, the processor is used to execute Fig. 20 The memory stores the program codes and various data required to execute the above modules or submodules. The network interface is used to transmit data between user terminals or servers. The memory in this embodiment stores the program codes and data required to execute all modules / submodules in the rice canopy height extraction device based on the canopy interception compensation mechanism of the present application, and the server can call the program codes and data of the server to execute the functions of all submodules.

[0168] The present application also provides a storage medium storing computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the rice canopy height extraction method based on the canopy interception compensation mechanism described in any embodiment of the present application.

[0169] The present application also provides a computer program product, including a computer program / instruction, which, when executed by one or more processors, implements the steps of the method for extracting rice canopy height based on the canopy interception compensation mechanism described in any embodiment of the present application.

[0170] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments of the present application can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a computer-readable storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0171] The above description is only a partial implementation method of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

[0172] In summary, this application has achieved high accuracy without the need for traditional methods to collect bare soil DEM before sowing. Even when the canopy density is extremely high, the results of this application can replace part of the manual measurement work in crop breeding, cultivation and other research, providing a novel solution for improving variety selection efficiency and optimizing cultivation patterns.

Claims

1. A rice canopy height extraction method based on canopy interception compensation mechanism, characterized in that: include: In response to the rice canopy height extraction instruction, the laser radar point cloud data returned from the rice field is obtained to construct a laser radar point cloud data set, and a preset ground point extraction algorithm is called to extract ground points in the rice field from the laser radar point cloud data set; Using the ground points in the rice paddy field and the non-ground points in the rice paddy field to construct a paddy field digital surface model, and using the ground points in the rice paddy field to construct a paddy field digital elevation model; Projecting the laser radar point cloud data onto a horizontal plane, dividing the horizontal plane into grid units with a preset area threshold, and counting the proportion of non-ground points in each grid unit as a canopy laser interception rate, wherein the canopy laser interception rate represents the proportion of laser radar point cloud data that cannot reach the ground due to obstruction by the rice canopy; Determine a preset compensation function, and correct the difference between the paddy field digital surface model and the paddy field digital elevation model according to the canopy laser interception rate and the compensation function to construct a rice canopy height extraction model; The rice canopy height in the rice field area to be studied is determined based on the rice canopy height extraction model to complete the extraction of the rice canopy height.

2. The method for extracting rice canopy height based on canopy interception compensation mechanism according to claim 1, characterized in that: The step of calling a preset ground point extraction algorithm to extract ground points in the rice field from the laser radar point cloud data set comprises: In response to the ground point extraction instruction, a cubic bounding box is created by using a progressive irregular triangulated network encryption algorithm to wrap the laser radar point cloud data and divide it into multiple areas of n rows and m columns. Initial ground seed points are uniformly selected from the laser point cloud data according to the local minimum elevation principle, and the remaining points are points to be determined, wherein the ground point extraction algorithm is a progressive irregular triangulated network encryption algorithm, and the local minimum elevation principle indicates that the lowest elevation point of the geographical location is found in each area as a ground seed point; Constructing an initial irregular triangulated network based on the initial ground seed points, and determining the projection coordinates of all the points to be determined on the horizontal plane to determine the triangular face to which the points to be determined belong in the irregular triangulated network; For each triangular face to which the point to be determined belongs in the irregular triangulated network, the slope of the corresponding triangular face is calculated and determined, wherein the slope of the triangular face represents the elevation change rate inside the triangular face; If the slope of the triangle surface is lower than the preset maximum slope threshold, calculate and determine a first distance from the point to be determined to the nearest triangle surface, and calculate and determine a first angle between a line between the point to be determined and its nearest triangle surface vertex and a normal vector of the triangle surface; If the first angle is less than the preset maximum slope threshold and the first distance is less than the preset maximum height threshold, the point to be judged is regarded as a ground point and included in the ground point set. The above steps are iteratively performed until no new ground points are selected or the preset maximum number of iterations is reached, thereby completing the extraction of the ground points in the rice field from the lidar point cloud data set.

3. The method for extracting rice canopy height based on canopy interception compensation mechanism according to claim 2, characterized in that: For each triangular face to which the to-be-determined point belongs in the irregular triangulated network, after the step of calculating and determining the slope of the corresponding triangular face, the method includes: If the slope of the triangular surface exceeds the preset maximum slope threshold, determine the mirror point of the point to be judged, and find the vertex with the largest elevation value in the triangular surface where the mirror point of the point to be judged is located, as the ground point in the rice field, wherein the mirror point represents the point that is symmetrical about a certain axis or plane of the point to be judged.

4. The method for extracting rice canopy height based on canopy interception compensation mechanism according to claim 1, characterized in that: After the step of projecting the laser radar point cloud data onto a horizontal plane, dividing the horizontal plane into grid units with a preset area threshold, and counting the proportion of non-ground points in each grid unit as the canopy laser interception rate, the method includes: Detecting whether the canopy laser interception rate exceeds a preset interception rate threshold, if the canopy laser interception rate does not exceed the preset interception rate threshold, calculating the difference between the paddy field digital surface model and the paddy field digital elevation model, and using the difference as the rice canopy height of the rice field, the preset interception rate threshold is 98%; If the canopy laser interception rate exceeds the preset interception rate threshold, continue to determine the preset compensation function, and correct the difference between the paddy field digital surface model and the paddy field digital elevation model according to the canopy laser interception rate and the compensation function.

5. The method for extracting rice canopy height based on canopy interception compensation mechanism according to claim 1, characterized in that: The steps to construct a rice canopy height extraction model include: Acquire a paddy field digital surface model and a paddy field digital elevation model, and calculate and determine the difference between the paddy field digital surface model and the paddy field digital elevation model; Determine a preset compensation function, and calculate and determine the product between the compensation function and the canopy laser interception rate; The rice canopy height extraction model is constructed based on the sum of the difference between the paddy field digital surface model and the paddy field digital elevation model and the product of the compensation function and the canopy laser interception rate.

6. The method for extracting rice canopy height based on canopy interception compensation mechanism according to claim 1, characterized in that: After the step of determining the rice canopy height in the rice field area to be studied based on the rice canopy height extraction model, the method includes: Determine the rice canopy height in the rice field area to be studied, detect whether the rice canopy height in the rice field area to be studied falls within the canopy height threshold corresponding to each growth cycle of rice, and if so, determine and spray the corresponding fertilizer element based on the growth cycle corresponding to the rice canopy height to complete the extraction of the rice canopy height.

7. The method for extracting rice canopy height based on canopy interception compensation mechanism according to any one of claim 5, characterized in that: The paddy field digital surface model is a DSM model, which includes the rice on the paddy field ground and the corresponding elevation information of the ground; the paddy field digital elevation model is a DEM model, which only includes the ground elevation information; The rice canopy height extraction model is a piecewise function, and the expression of the rice canopy height extraction model is: Among them, the H LiDAR represents the difference between the paddy field digital surface model and the paddy field digital elevation model, and P represents the canopy laser interception rate.

8. A rice canopy height extraction device based on canopy interception compensation mechanism, characterized in that: include: A ground point extraction module is configured to respond to a rice canopy height extraction instruction, obtain laser radar point cloud data returned from a rice field to construct a laser radar point cloud data set, and call a preset ground point extraction algorithm to extract ground points in the rice field from the laser radar point cloud data set; An initial model construction module is configured to construct a paddy field digital surface model using ground points in the paddy field and non-ground points in the paddy field, and to construct a paddy field digital elevation model using ground points in the paddy field; a laser interception rate determination module, configured to project the laser radar point cloud data onto a horizontal plane, divide the horizontal plane into grid units of a preset area threshold, and count the proportion of non-ground points in each grid unit as a canopy laser interception rate, wherein the canopy laser interception rate represents the proportion of the laser radar point cloud data that cannot reach the ground due to obstruction by the rice canopy; a height extraction model construction module, configured to determine a preset compensation function, and to correct the difference between the paddy field digital surface model and the paddy field digital elevation model according to the canopy laser interception rate and the compensation function, so as to construct a rice canopy height extraction model; The canopy height extraction module is configured to determine the rice canopy height in the rice field area to be studied based on the rice canopy height extraction model to complete the extraction of the rice canopy height.

9. An electronic device, comprising a central processing unit and a memory, characterized in that: The central processing unit is used to call and run the computer program stored in the memory to execute the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: It stores a computer program implemented according to the method described in any one of claims 1 to 7 in the form of computer-readable instructions, and when the computer program is called and executed by a computer, the steps included in the corresponding method are executed.

Images