A high-precision positioning method and system for densely planted and closed environments

Abstract

The present invention discloses a high-precision positioning method and system for use in a densely planted and closed environment, the method comprising: deploying a passive GNSS differential positioning system in a scene; recording the positioning data of the passive GNSS differential positioning system; calculating a positioning differential sequence; acquiring attitude information; performing differential analytical operations to obtain preliminary positioning correction data; performing multi-source heterogeneous fusion and filtering processing on the attitude information time series and the preliminary positioning correction data to obtain a high-precision positioning model; performing noise optimization and feature extraction processing on the original data sequence in the high-precision positioning model, and inputting the processing into the high-precision positioning model to obtain high-precision positioning data. The system comprises a UM982‑GNSS main positioning module, a ZED‑F9P auxiliary positioning module, a BT‑800S full-frequency antenna and a high-precision positioning processing module. By using the present invention, the positioning accuracy under complex conditions can be improved. The present invention can be widely used in the field of positioning.

Description

Technical Field

[0001] The present invention relates to the field of positioning, and in particular to a high-precision positioning method and system applied to a densely planted and enclosed environment. Background Art

[0002] With the continuous advancement of precision agriculture and smart agriculture technologies, the related fields of transportation equipment, automated sowing, etc. have made sufficient progress. However, for many economic crops in my country, such as citrus and lychees, although the equipment technologies such as picking and transportation have achieved obvious application results, these crops are often in densely planted conditions, which leads to positioning difficulties and further affects the full play of automation functions, making it difficult to apply new equipment smoothly. In response to this problem, based on the existing satellite positioning technology, how to overcome the impact of densely planted and closed environments on satellite positioning signals, thereby solving the problem of high-precision positioning of modern intelligent equipment under complex conditions, is of great significance for the development of more comprehensive smart agricultural intelligent equipment. Summary of the invention

[0003] In view of this, in order to solve the technical problem that the existing positioning method cannot overcome the influence of the densely planted and closed environment on the satellite positioning signal, thereby resulting in low positioning accuracy under complex conditions, the present invention proposes a high-precision positioning method applied to a densely planted and closed environment, the method comprising the following steps:

[0004] Deploy the passive GNSS differential positioning system in the scene;

[0005] Based on the passive GNSS differential positioning system, obtaining positioning information and attitude information;

[0006] Based on the positioning information and the posture information, multi-source heterogeneous fusion and filtering processing are performed to obtain a high-precision positioning model;

[0007] Perform noise optimization on the original data sequence in the high-precision positioning model to obtain the inverse transformation data sequence;

[0008] Feature extraction is performed based on the original data sequence and the inverse transformation data sequence, and the data is input into a high-precision positioning model to obtain high-precision positioning data.

[0009] In some embodiments, further comprising

[0010] Build a second passive GNSS differential positioning system and position prediction model;

[0011] Define error ranges and optimization rules;

[0012] Based on the error range, the optimization rule, the data of two sets of passive GNSS differential positioning systems and the position prediction model, the optimized positioning information and attitude information are obtained;

[0013] The optimization rule is:

[0014] Based on the position and speed at the previous moment, it is judged that the current position data of one of the passive GNSS differential positioning systems exceeds the error range. Combined with the preset coefficient, the data of the position prediction model and the data of the other passive GNSS differential positioning system are weighted and summed to obtain the optimized current position data.

[0015] The optimization rule formula is expressed as follows:

[0016]

[0017] Among them, I represents the positioning data of the passive GNSS differential positioning system whose position data is a suspected value; J represents the positioning data of another passive GNSS differential positioning system corresponding to I, E represents the data of the position prediction model, O represents the number of times exceeding the error range, and range represents the position interval constructed based on the position and speed at the previous moment.

[0018] Through this preferred step, better data is generated, which helps to output high-precision positioning data.

[0019] In some embodiments, it also includes:

[0020] The positioning data of the passive GNSS differential positioning system is recorded, and the positioning drift trajectory is drawn using a real-time incremental storage method.

[0021] Through this preferred step, users can see the trajectory data more intuitively and judge whether the data is reasonable.

[0022] In some embodiments, in the multi-source heterogeneous fusion, the posture fusion parameter is defined as in:

[0023]

[0024] Among them, g xk+1 represents the x-axis coordinate of the k+1th iteration in the geodetic coordinate system, g yk+1 represents the y-axis coordinate of the k+1th iteration in the geodetic coordinate system, A k represents the posture fusion parameter matrix of the kth iteration, g xk represents the x-axis coordinate of the kth iteration in the geodetic coordinate system, g yk represents the y-axis coordinate of the kth iteration in the geodetic coordinate system, ψk represents the heading angle of the kth iteration, v k represents the speed of the kth iteration, Ω k represents the yaw rate of the kth iteration, a k represents the acceleration of the kth iteration, r k represents the rolling angular rate at the kth iteration, f k represents the pitch angle of the kth iteration, and t represents the time.

[0025] In some embodiments, the iterative matrix expression of posture fusion in the high-precision positioning model is as follows:

[0026]

[0027] Where, wdt represents the instantaneous integral value of the yaw angular rate, q xk represents the first noise matrix, A k Represents the posture fusion parameters.

[0028] In some embodiments, the matrix expression of the filtering process in the high-precision positioning model is as follows:

[0029]

[0030] Among them, m k represents the measured parameter matrix, r zk represents the second noise matrix.

[0031] In some embodiments, the step of obtaining the inverse transformation data sequence based on noise optimization of the original data sequence in the high-precision positioning model specifically includes:

[0032] Construct a three-dimensional nonlinear filter based on a three-dimensional convolution kernel, a maximum pooling layer, and a secondary maximum pooling convolution kernel;

[0033] Based on the three-dimensional nonlinear filter output sensor data sequence, obtain a feature sequence;

[0034] The feature sequence is received based on the deconvolution kernel, and an inverse transformation data sequence is output.

[0035] In some embodiments, the step of performing feature extraction processing based on the original data sequence and the inverse transformation data sequence and inputting the features into a high-precision positioning model to obtain high-precision positioning data specifically includes:

[0036] Calculating the difference value between the inverse transformation data sequence and the original data sequence based on a radial basis neural network to form a difference value matrix;

[0037] Extracting eigenvalues ​​of the difference value matrix based on a sample kernel function to obtain noise eigenvalues;

[0038] The noise characteristic value is inputted back into the high-precision positioning model to form a comprehensive high-precision positioning parameter including noise characteristic correction.

[0039] The present invention also proposes a high-precision positioning system for use in a densely planted and closed environment, which is used to execute the high-precision positioning method for use in a densely planted and closed environment as described above, including a UM982-GNSS main positioning module, a ZED-F9P auxiliary positioning module, a BT-800S full-frequency antenna and a high-precision positioning processing module, wherein:

[0040] The high-precision positioning processing module includes an IDF-ESP-WROOM-32 differential analytical operation processing unit, an AWT601 inertial measurement unit, an E6B2 three-phase incremental encoder and a Jetson AGX Xavier core processing hub.

[0041] In some embodiments, the UM982-GNSS main positioning module, the ZED-F9P auxiliary positioning module and the BT-800S full-frequency antenna are electrically connected to the antenna respectively.

[0042] In some embodiments, the AWT601 inertial measurement unit and the E6B2 three-phase incremental encoder are respectively electrically connected to the Jetson AGX Xavier core processing hub.

[0043] Based on the above scheme, the present invention provides a high-precision positioning method and system for application in densely planted and closed environments. First, there is no need to deploy additional base stations, which has the advantages of low cost and easy maintenance. In addition, the present invention relies on its own basic satellite positioning module to achieve preliminary positioning, does not rely on other platforms, and mainly relies on its own situational awareness sensor. With the help of relevant fusion processing algorithms, high-precision positioning can be achieved. Finally, the basic modules and devices relied on by the present invention have the characteristics of high reliability, which is particularly suitable for positioning applications in complex environments such as densely planted and closed environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a flow chart of the steps of a high-precision positioning method applied to a densely planted and closed environment of the present invention;

[0045] Figure 2 This is a structural block diagram of a high-precision positioning system for use in a densely planted and closed environment according to the present invention;

[0046] Figure 3 It is a schematic diagram of the positioning trajectory of the method of the present invention in a densely planted and closed environment;

[0047] Figure 4 This is a schematic diagram of the comparison of the positioning trajectory of traditional RTK in a densely planted and closed environment; DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0049] It should be noted that, for the convenience of description, only the parts related to the relevant invention are shown in the drawings. In the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other.

[0050] It should be understood that the "system", "device", "unit" and / or "module" used in this application is a method for distinguishing different components, elements, parts, portions or assemblies at different levels. However, if other words can achieve the same purpose, the word can be replaced by other expressions.

[0051] Unless the context clearly indicates an exception, the words "a", "an", "a kind" and / or "the" do not refer to the singular, but also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of clearly identified steps and elements, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements. The elements defined by the sentence "includes a..." do not exclude the existence of other identical elements in the process, method, commodity or device that includes the elements.

[0052] In the description of the embodiments of the present application, "plurality" means two or more than two. The following terms "first" and "second" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features.

[0053] In addition, flow charts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, the various steps may be processed in reverse order or simultaneously. At the same time, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0054] Reference Figure 1 , is a flow chart of an optional example of a high-precision positioning method for densely planted and closed environments proposed by the present invention. The method can be applied to computer devices. The positioning method proposed in this embodiment may include but is not limited to the following steps:

[0055] Step S1, deploying a passive GNSS differential positioning system in a scene;

[0056] Step S2: acquiring positioning information and attitude information based on the passive GNSS differential positioning system;

[0057] Step S3, based on the positioning information and the posture information, multi-source heterogeneous fusion and filtering processing are performed to obtain a high-precision positioning model;

[0058] Step S4, performing noise optimization on the original data sequence in the high-precision positioning model to obtain an inverse transformation data sequence;

[0059] Step S5: Perform feature extraction processing based on the original data sequence and the inverse transformation data sequence, and input them into a high-precision positioning model to obtain high-precision positioning data.

[0060] In some feasible embodiments, step S1 further includes:

[0061] The positioning data of the passive GNSS differential positioning system is recorded, and the positioning drift trajectory is drawn using a real-time incremental storage method.

[0062] Specifically: After the passive GNSS differential positioning system consisting of the UM982-GNSS main positioning module, the ZED-F9P auxiliary positioning module, and the BT-800S full-frequency antenna is deployed at a designated densely planted point, it is kept stationary for 30 minutes, the approximate positioning data obtained during the period is recorded, and its positioning drift trajectory is plotted;

[0063] In some feasible embodiments, the step S2 further includes:

[0064] Build a second passive GNSS differential positioning system and position prediction model;

[0065] Define error ranges and optimization rules;

[0066] Based on the error range, the optimization rule, the data of two sets of passive GNSS differential positioning systems and the position prediction model, the optimized positioning information and attitude information are obtained;

[0067] The optimization rule is:

[0068] Based on the position and speed at the previous moment, the current position data of one of the passive GNSS differential positioning systems is judged to be a suspicious value. Combined with the preset coefficient, the data of the position prediction model and the data of the other passive GNSS differential positioning system are weighted and summed to obtain the optimized current position data.

[0069] Specifically, the formula is as follows:

[0070]

[0071] Among them, I represents the positioning data of the passive GNSS differential positioning system whose position data is a suspected value; J represents the positioning data of another passive GNSS differential positioning system corresponding to I, E represents the data of the position prediction model, O represents the number of times exceeding the error range, and range represents the position interval constructed based on the position and speed at the previous moment.

[0072] In addition, when there is no doubt value, any one of the passive GNSS differential positioning systems can be selected as accurate data, or the average value can be taken, which is not specifically limited here.

[0073] The location prediction model may be a neural network model trained in a common manner.

[0074] In some feasible embodiments, in step S2, specifically:

[0075] S2.1, obtain posture information;

[0076] The Jetson AGX Xavier core processing center is used to obtain the pitch angle, yaw angle, roll angle and speed information of the target carrier platform provided by the AWT601 inertial measurement unit and the E6B2 three-phase incremental encoder in real time, and an additional writing algorithm is used to establish a real-time attitude information time series;

[0077] S2.2, obtaining positioning information;

[0078] Based on the RTK algorithm, the data of the UM982-GNSS main positioning module and the ZED-F9P auxiliary positioning module are differentially calculated to generate positioning information.

[0079] Specifically, the Jetson AGX Xavier core processing center sends instructions to control the IDF-ESP-WROOM-32 differential analysis processing module through Wi-Fi communication, and uses the second-order smoothing filter method to process the positioning drift data. At the same time, the IDF-ESP-WROOM-32 differential analysis processing module uploads the calculated positioning drift data features to the global public NTRIP Caster platform through the Internet, thereby obtaining preliminary positioning correction data and generating positioning information.

[0080] In some feasible embodiments, in step S3, specifically:

[0081] The attitude information time series and positioning correction data obtained in S2 are used to establish a sensor data fusion matrix sequence and a sensor data filtering matrix sequence based on the geodetic coordinate system according to the multi-source heterogeneous fusion algorithm and the sensor filtering algorithm, thereby establishing a high-precision positioning model based on the geodetic coordinate system for the target carrier platform;

[0082] The multi-source heterogeneous fusion algorithm is used to fuse the posture data of the positioning target, and the posture parameter fusion matrix is ​​designed as Then the nonlinear equations of each parameter formula can be expressed as:

[0083]

[0084] Among them, g xk ,g yk Represents the coordinates of the geodetic coordinate system, ψ, v, Ω, a, r, f represent the heading angle, velocity, yaw rate, acceleration, roll angular velocity and pitch angle respectively. After measurement and data fusion by the integrated sensors, the omnidirectional fusion attitude information of the positioning target can be obtained, and the fusion model matrix expression formula is:

[0085]

[0086] Where, wdt represents the instantaneous integral value of the yaw angular rate, q xk represents the first noise matrix, A k Represents the posture fusion parameters.

[0087] The sensor filtering algorithm is used to process the noise of multi-source sensor data and construct the multi-source sensor measurement parameters into a parameter matrix Where x, y represent the satellite positioning data obtained by UM982-GNSS and ZED-F9P, ψ, v, a, r, f represent the sensor heading angle, velocity, acceleration, roll angular rate, and pitch angle information obtained by the sensor, respectively. The analytical method can be expressed as:

[0088] v k+1 =a k dt+v k

[0089] ψ k+1 =ψ k +Ω k dt

[0090] Ω k+1 =Ω k +q Ωk

[0091] a k+1 =a k +qak

[0092] r k+1 =r k +q rk

[0093] f k+1 =f k +q fk

[0094] Among them, q Ωk ,q ak ,q rk ,q fk is the corresponding sensor noise. After analytical processing, the multi-sensor comprehensive noise can be expressed by the matrix equation:

[0095]

[0096] Among them, r zk represents the second noise matrix.

[0097] Through this preferred embodiment, a sensor data fusion matrix sequence and a sensor data filtering sequence based on a geodetic coordinate system are established, thereby establishing a high-precision positioning model based on a geodetic coordinate system for a target carrier platform;

[0098] In some feasible embodiments, in step S4, it specifically includes:

[0099] Based on the three-dimensional nonlinear filter algorithm, the noise matrix in the original data sequence in the high-precision positioning model is deeply processed and optimized. The construction steps are as follows:

[0100] (1) Set the original sensor data sequence to S0;

[0101] (2) Establish 32 three-dimensional convolution kernels to form a primary feature extraction network, and the convolution kernel size is 3*3;

[0102] (3) Construct a maximum pooling layer to reduce some high-dimensional features, and the pooling convolution kernel size is 3*2;

[0103] (4) Construct 64 three-dimensional convolution kernels to form a secondary feature extraction module, and the convolution kernel size is 3*3;

[0104] (5) Add a secondary maximum pooling convolution kernel with a convolution kernel size of 3*2;

[0105] (6) A sensor data sequence after feature extraction obtained from the output of the three-dimensional nonlinear filter constructed in the above steps (1) to (5), denoted as Se1;

[0106] (7) Construct a 64-layer three-dimensional deconvolution kernel and receive the feature sequence Se1. The deconvolution kernel size is 3*3;

[0107] (8) Further, a 32-layer three-dimensional deconvolution kernel is constructed, and the deconvolution kernel size is 3*3;

[0108] (9) The inverse transformed data sequence is recorded as Se2, which is the sensor data sequence after nonlinear filtering.

[0109] In some feasible embodiments, in step S5, seeking specifically includes:

[0110] Calculating the difference value between the inverse transformation data sequence and the original data sequence based on a radial basis neural network to form a difference value matrix;

[0111] Extracting eigenvalues ​​of the difference value matrix based on a sample kernel function to obtain noise eigenvalues;

[0112] The kernel function formula is:

[0113]

[0114] Among them, k(a,b) is the calculated value of the kernel function of the difference value sequence sample, σ 2 Represents the one-dimensional variance of the sample. Its feature induction is based on the three-dimensional mutual reachability distance as the basis for judgment. Its calculation formula is:

[0115] d(a,b,c)=max{core k (a),core k (b),core k (c),d(a,b),d(a,c),d(b,c)}

[0116] Among them, a, b, and c represent the three dimensions of the sample respectively. Thus, the feature extraction of the difference value matrix sample and the acquisition of the target noise eigenvalue are converted into an optimization problem. The mutually reachable distance with the smallest distance and the largest number of samples is the eigenvalue being sought.

[0117] The noise characteristic value is inputted back into the high-precision positioning model to form a comprehensive high-precision positioning parameter including noise characteristic correction.

[0118] The noise characteristic value corresponds to the various error data mentioned above.

[0119] Through this preferred embodiment, on the basis of establishing a high-precision positioning model, a comprehensive optimization of the model noise is carried out to obtain the final high-precision positioning data, thereby achieving the purpose of high-precision positioning in a densely planted and closed environment.

[0120] The method and system were deployed in the field (balsam pear test field of Guangzhou National Modern Agricultural Science and Technology Innovation Center). Based on the same parameter settings, the original RTK positioning data in a densely planted and closed environment was compared with the positioning data performance of the method and system described in the patent of this invention by moving and recording real-time positioning data. The results are as follows: Figure 3 As shown, it can be seen that in a densely planted and closed environment, the positioning trajectory of the present invention is reasonable and clear, while the positioning trajectory obtained by the traditional RTK algorithm is messy and unreasonable.

[0121] like Figure 2 As shown, a high-precision positioning system for densely planted and closed environments includes a UM982-GNSS main positioning module, a ZED-F9P auxiliary positioning module, a BT-800S full-frequency antenna and a high-precision positioning processing module, wherein:

[0122] The high-precision positioning processing module includes an IDF-ESP-WROOM-32 differential analytical operation processing unit, an AWT601 inertial measurement unit, an E6B2 three-phase incremental encoder and a Jetson AGX Xavier core processing center, which incorporates a positioning data processing method in a densely planted and closed environment consisting of a multi-source heterogeneous fusion model, a sensor filtering algorithm, a three-dimensional nonlinear filter algorithm and a radial basis neural network.

[0123] The UM982-GNSS main positioning module, ZED-F9P auxiliary positioning module and BT-800S full-frequency antenna are used to provide basic satellite positioning data, and are electrically connected to the antenna respectively.

[0124] The AWT601 inertial measurement unit and E6B2 three-phase incremental encoder are used to provide real-time omnidirectional attitude information for the carrier platform and are electrically connected to the Jetson AGX Xavier core processing hub.

[0125] The IDF-ESP-WROOM-32 differential analysis operation processing module analyzes the NMEA differential data of the UM982-GNSS main positioning module and the ZED-F9P auxiliary positioning module respectively, and establishes a time-based differential positioning data sequence. The IDF-ESP-WROOM-32 differential analysis operation processing module carries the established time series data and obtains preliminary correction data through Wi-Fi access to the Internet.

[0126] The contents of the above method embodiments are all applicable to the present system embodiments. The functions specifically implemented by the present system embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0127] A storage medium stores processor-executable instructions, which are used to implement a high-precision positioning method for densely planted and closed environments as described above when executed by the processor.

[0128] The contents of the above method embodiments are all applicable to the present storage medium embodiments. The functions specifically implemented by the present storage medium embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0129] The above is a specific description of the preferred implementation of the present invention, but the invention is not limited to the embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. A high-precision positioning method applied to densely planted and closed environments, characterized in that: The following steps are involved: Deploy the passive GNSS differential positioning system in the scene; Based on the passive GNSS differential positioning system, obtaining positioning information and attitude information; Based on the positioning information and the posture information, multi-source heterogeneous fusion and filtering processing are performed to obtain a high-precision positioning model; Perform noise optimization on the original data sequence in the high-precision positioning model to obtain the inverse transformation data sequence; Performing feature extraction processing based on the original data sequence and the inverse transformation data sequence, and inputting the features into a high-precision positioning model to obtain high-precision positioning data; The step of obtaining the inverse transformation data sequence based on noise optimization of the original data sequence in the high-precision positioning model specifically includes: Construct a three-dimensional nonlinear filter based on a three-dimensional convolution kernel, a maximum pooling layer, and a secondary maximum pooling convolution kernel; Based on the three-dimensional nonlinear filter output sensor data sequence, obtain a feature sequence; The feature sequence is received based on the deconvolution kernel, and an inverse transformation data sequence is output.

2. According to claim 1, a high-precision positioning method applied to a densely planted and closed environment is characterized in that: Also includes: Build a second passive GNSS differential positioning system and position prediction model; Define error ranges and optimization rules; Based on the error range, the optimization rule, the data of two sets of passive GNSS differential positioning systems and the position prediction model, the optimized positioning information and attitude information are obtained; The optimization rule is: Based on the position and speed at the previous moment, it is judged that the current position data of one of the passive GNSS differential positioning systems exceeds the error range. Combined with the preset coefficient, the data of the position prediction model and the corresponding passive GNSS differential positioning system are weighted and summed to obtain the optimized current position data.

3. According to claim 2, a high-precision positioning method applied to a densely planted and closed environment is characterized in that: The optimization rule formula is expressed as follows: Among them, I represents the positioning data of the passive GNSS differential positioning system whose position data is a suspected value; J represents the positioning data of another passive GNSS differential positioning system corresponding to I, E represents the data of the position prediction model, O represents the number of times exceeding the error range, and range represents the position interval constructed based on the position and speed at the previous moment.

4. According to claim 1, a high-precision positioning method applied to a densely planted and closed environment is characterized in that: Also includes: The positioning data of the passive GNSS differential positioning system is recorded, and the positioning drift trajectory is drawn using a real-time incremental storage method.

5. According to claim 1, a high-precision positioning method applied to a densely planted and closed environment is characterized in that: In the multi-source heterogeneous fusion, the posture fusion parameter is defined as A k =[g xk ,g yk ,ψ k ,v k ,Ω k ,a k ,r k ,f k ] T ,in: Among them, g xk+1 represents the x-axis coordinate of the k+1th iteration in the geodetic coordinate system, g yk+1 represents the y-axis coordinate of the k+1th iteration in the geodetic coordinate system, A k represents the posture fusion parameter matrix of the kth iteration, g xk represents the x-axis coordinate of the kth iteration in the geodetic coordinate system, g yk represents the y-axis coordinate of the kth iteration in the geodetic coordinate system, ψ k represents the heading angle of the kth iteration, v k represents the speed of the kth iteration, Ω k represents the yaw rate of the kth iteration, a k represents the acceleration of the kth iteration, r k represents the rolling angular rate at the kth iteration, f k represents the pitch angle of the kth iteration, and t represents the time.

6. A high-precision positioning method for densely planted and closed environments according to claim 5, characterized in that: The iterative matrix expression of posture fusion in the high-precision positioning model is as follows: Where, wdt represents the instantaneous integral value of the yaw angular rate, q xk represents the first noise matrix, A k Represents the posture fusion parameters.

7. The high-precision positioning method for densely planted and closed environments according to claim 1, characterized in that: The matrix expression of the filtering process in the high-precision positioning model is as follows: Among them, m k represents the measured parameter matrix, r zk represents the second noise matrix.

8. The high-precision positioning method for densely planted and closed environments according to claim 1, characterized in that: The step of performing feature extraction processing based on the original data sequence and the inverse transformation data sequence and inputting the features into the high-precision positioning model to obtain high-precision positioning data specifically includes: Calculating the difference value between the inverse transformation data sequence and the original data sequence based on a radial basis neural network to form a difference value matrix; Extracting eigenvalues ​​of the difference value matrix based on a sample kernel function to obtain noise eigenvalues; The noise characteristic value is inputted back into the high-precision positioning model to form a comprehensive high-precision positioning parameter including noise characteristic correction.

9. A high-precision positioning system for use in densely planted environments, characterized in that: Used to execute the high-precision positioning method applied to a densely planted and closed environment as claimed in claim 1, comprising a UM982-GNSS main positioning module, a ZED-F9P auxiliary positioning module, a BT-800S full-frequency antenna and a high-precision positioning processing module, wherein: The high-precision positioning processing module includes an IDF-ESP-WROOM-32 differential analysis processing unit, an AWT601 inertial measurement unit, an E6B2 three-phase incremental encoder, and a JetsonAGX Xavier core processing hub; The UM982-GNSS main positioning module, the ZED-F9P auxiliary positioning module and the BT-800S full-frequency antenna are electrically connected to the antenna respectively; the AWT601 inertial measurement unit and the E6B2 three-phase incremental encoder are electrically connected to the JetsonAGXXavier core processing hub respectively.

Images