A tillage depth monitoring system and method based on cross-medium positioning and ground perception

Abstract

The application discloses a kind of based on cross-media positioning and surface perception's ploughing depth monitoring system and method, it is related to agricultural machinery technical field, including the following steps: step one, with rigid plane of being fixed to the working machine frame as reference, establish unified machine coordinate system;Step two, operating component relative positioning;Step three, dynamic surface datum solution based on regional imaging;Step four, real-time ploughing depth value fusion calculation and cyclic output.The application adopts the above-mentioned one based on cross-media positioning and surface perception's ploughing depth monitoring system and method, realizes the direct three-dimensional positioning of soil-entering operating component, eliminates mechanical structure error from source;Self-consistent relative measurement coordinate system is constructed, and the disturbance caused by vehicle attitude change is completely avoided;Regional imaging and statistical average method is used, and stable and reliable surface datum is obtained;Cross-media refraction correction is introduced, and positioning precision is improved.

Description

Technical Field

[0001] This invention relates to the field of agricultural machinery technology, and in particular to a tillage depth monitoring system and method based on cross-media positioning and surface sensing. Background Technology

[0002] Tillage depth is a key indicator for measuring the quality of farmland cultivation, significantly impacting soil physical and chemical properties, enhancing crop resistance, and increasing yield. Whether it's deep tillage to break up the plow pan, or plowing, sowing, and furrowing, real-time and accurate online monitoring of tillage depth is crucial for achieving precision agriculture and improving agricultural mechanization. Currently, existing tillage depth measurement technologies can be broadly categorized into two main approaches:

[0003] The first method is indirect measurement. This approach involves installing ultrasonic or laser ranging sensors at specific locations on the agricultural machinery frame to measure the relative distance between a part of the frame and the ground surface. This distance is then converted using pre-defined geometric relationships to obtain the tillage depth information. For example, related technologies (such as Chinese patents CN201621090104.7, CN201410143575.9, and CN202110245746.9) disclose schemes for indirectly obtaining tillage depth information using ranging sensors mounted on the frame. The measurement accuracy of this method depends to some extent on the assumption of rigidity in the machinery structure. However, in actual operation, the elastic deformation of the machinery, the wear and gaps of connecting parts, and the pitch and roll changes of tractors and other vehicles can all introduce uncertainties into calculations based on fixed geometric models, posing a challenge to the stability of the measurement results. Furthermore, the signal quality of these sensors may also be affected by common field environments such as dust, mud, and crop residues, placing higher demands on the reliability of the measurements.

[0004] The second type is mechanical contact measurement technology. This approach uses a moving mechanical component that directly contacts the soil to sense depth. For example, Chinese patent CN202411110879.5 discloses a technical solution that uses a sliding plate that follows a soil loosening shovel to measure spring compression to reflect tillage depth. In this type of mechanical contact method, the moving parts directly rub and collide with the soil, rocks, etc., which puts the wear resistance, impact resistance, and long-term stability of the components to the test under long-term high-intensity working conditions. At the same time, complex soil conditions and changes in friction force may also affect the linearity and consistency of its measurement response.

[0005] In summary, existing technologies still face certain challenges in achieving high-precision and high-stability real-time tillage depth measurement, especially in how to avoid the combined effects of mechanical structural deformation of various tillage implements, changes in vehicle posture, and complex field environments on the measurement from the perspective of measurement principles. There is still room for technological improvement. Summary of the Invention

[0006] The purpose of this invention is to provide a tillage depth monitoring system and method based on cross-media positioning and surface sensing. From the perspective of measurement principle, it effectively avoids the measurement uncertainty caused by the elastic deformation of machinery, component wear, and changes in vehicle attitude (pitch, tilt, etc.), and improves the stability and reliability of surface benchmark measurement in complex field environments such as surface soil and crop residue.

[0007] To achieve the above objectives, the present invention provides a tillage depth monitoring system based on cross-media positioning and surface sensing, including a relative positioning subsystem, a surface sensing subsystem, and a processing module, wherein the processing module is connected to the surface sensing subsystem and the relative positioning subsystem, respectively.

[0008] The relative positioning subsystem includes a UWB signal source built into the working end of the soil-entry component, and a UWB receiving anchor array fixedly installed on the frame of the tillage implement. The UWB receiving anchor array is located directly above the soil-entry component.

[0009] The surface sensing subsystem includes a depth camera or millimeter-wave radar mounted on the frame of the tillage implement, used to collect three-dimensional point clouds of the surface area to be tilled, and the field of view of the depth camera or millimeter-wave radar covers the area in front of the direction of travel of the soil-entry operation component.

[0010] The processing module is configured to execute the tillage depth monitoring method.

[0011] Preferably, the spatiotemporal alignment and tillage depth calculation steps performed by the processing module specifically include:

[0012] Get current job speed V And the longitudinal physical distance between the center of the field of view measurement area and the working end of the soil-penetrating component. L ;

[0013] Calculate the time delay Δ between the surface data and the measurement point to the work site. t : ;

[0014] At the current sampling time t k Retrieve time from data buffer t k -Δ t Corresponding historical surface reference height Z surface ( t k -Δ t );

[0015] Get the current time t kVertical height of the working end obtained by solution z tip ( t k The real-time tillage depth value after spatiotemporal consistency compensation is calculated using the following formula. D ( t k ): D ( t k )= Z surface ( t k -Δ t )- z tip ( t k ).

[0016] A method for monitoring tillage depth based on cross-media positioning and surface sensing includes the following steps:

[0017] Step 1: System initialization and coordinate system calibration. A unified machine coordinate system is established based on the rigid plane fixed to the machine tool frame.

[0018] Step 2: Relative positioning of the working components. Using the ultra-wideband receiving anchor array on the frame, the signal transmitted by the UWB source at the working end of the working components is received. Based on the time difference of arrival algorithm and combined with the cross-medium propagation refraction correction model, the first three-dimensional coordinates of the UWB source in the machine coordinate system are calculated.

[0019] Step 3: Dynamic surface benchmark calculation based on regional imaging. The three-dimensional point cloud data of the surface area to be cultivated in front of the soil entry operation component is obtained using a non-contact ranging device. The three-dimensional point cloud data is then converted to the coordinate system of the machine and the surface benchmark is calculated through statistical processing.

[0020] Step 4: Real-time tillage depth value fusion calculation and iterative output. Based on the current operating speed and the installation position relationship between the surface sensing subsystem and the operating components, spatiotemporal consistency compensation is performed on the surface benchmark calculated in Step 3; the vertical component of the first three-dimensional coordinate is calculated, and the difference between the height of the surface benchmark after spatiotemporal consistency compensation and the vertical component in the machine coordinate system is used to obtain the real-time tillage depth value. The real-time tillage depth value is used as the observation input, and the Kalman filter algorithm is used to perform temporal smoothing on the real-time tillage depth value to establish a state-space model of tillage depth change, and the smoothed tillage depth data is output.

[0021] Preferably, step one includes defining the coordinate system, calibrating the anchor point coordinates, and calibrating the extrinsic parameters of the depth camera.

[0022] Preferably, in step two, the UWB signal source is fixed at the connection between the end of the working component and the connecting handle. The working end is provided with a wave-transparent window, which is made of high-strength ceramic or composite material. The main lobe of the UWB signal source antenna faces the UWB receiving anchor array through the wave-transparent window.

[0023] Preferably, the UWB receiving anchor array includes at least four UWB receiving anchors arranged in a coplanar manner. The UWB receiving anchor array and the UWB signal source form a follow-up relative positioning system. The UWB signal source is a low-frequency chip with a center frequency of 3GHz to 4GHz.

[0024] Preferably, in step two, for deep soil operations or high moisture content conditions, the UWB source adopts a split radio frequency extension structure; the antenna of the UWB source is extended and arranged in the shallow soil area of ​​the upper middle part of the connecting handle of the soil entry operation component through radio frequency feed lines, and the computing unit of the UWB source is sealed inside the working end of the soil entry operation component.

[0025] Preferably, the cross-medium propagation refraction correction model in step two includes the following steps:

[0026] S21. Obtain soil physical parameters under the working environment;

[0027] S22. Calculate the propagation speed of dynamic signals in the soil medium based on soil physical parameters. v soil And determine the speed of signal propagation in the air. v air ;

[0028] The propagation speed of dynamic signals in the soil medium is calculated based on soil physical parameters, specifically including:

[0029] Using the Topp model based on real-time volumetric water content θ v Calculate the relative permittivity of soil : Then calculate the propagation speed in the soil. v soil ; in, c It is the speed of light in a vacuum.

[0030] S23. Using the surface height data acquired by the surface sensing subsystem, the soil-air interface is modeled as a plane. Snell's law constraint is introduced, and the UWB signal is propagated from the UWB source to the... i The path model for each anchor point is defined as the distance between the soil segment and the air segment. d air,i Based on the segmented propagation path, a propagation time model is established: ;

[0031] in,P tip Let be the three-dimensional coordinates of the end of the operation to be determined;

[0032] S24. Introduce a refraction point at the interface to satisfy the Snell's law constraint:

[0033] ;

[0034] in, θ soil , θ air These are the angles between the propagation directions of the soil segment and the air segment relative to the interface normal, respectively.

[0035] S25, To measure the arrival time difference Δ t i1,meas The objective function is the sum of squared residuals of the differences between the theoretical propagation time and the segmented propagation time. J ( P tip ), and perform nonlinear optimization to solve:

[0036] ;

[0037] The first three-dimensional coordinates are obtained by solving for the minimum value of the objective function. P tip .

[0038] Preferably, for the split RF extension structure, the three-dimensional coordinates of the working end to be determined in S23 to S25 are replaced with the three-dimensional coordinates of the UWB antenna center for calculation.

[0039] The specific steps for calculating the first three-dimensional coordinates are as follows: First, the position of the UWB antenna center in the machine coordinate system is calculated using the nonlinear optimization solution. Then, the three-dimensional coordinates of the working end are calculated using the known geometric dimensions of the connecting handle and the fixed geometric offset of the UWB antenna relative to the working end.

[0040] Preferably, the non-contact ranging device in step three is a depth camera or millimeter-wave radar; the statistical processing specifically includes: filtering and denoising the acquired three-dimensional point cloud data to remove outliers; statistically analyzing the vertical height values ​​of the filtered effective point cloud data, and selecting one of the arithmetic mean, median, or Gaussian distribution fitting peak value as the ground reference height.

[0041] Therefore, the above-mentioned tillage depth monitoring system and method based on cross-media positioning and surface sensing has the following beneficial effects:

[0042] (1) Direct three-dimensional positioning of the soil-entry operation component was achieved, eliminating mechanical structure errors at the source. By embedding an ultra-wideband signal source into the working end of the soil-entry operation component, direct, penetrating three-dimensional positioning of the soil-entry tool was achieved for the first time. Since the three-dimensional coordinates of the working component itself are directly measured, any elastic deformation, component wear, or linkage clearance generated by the tool during operation will not affect the accuracy of the positioning results.

[0043] (2) A self-consistent relative measurement coordinate system was constructed, completely avoiding the interference caused by changes in vehicle attitude. By deploying an array of UWB receiving anchor points on the frame components directly above the working parts, a dynamic relative positioning system with constant geometric relationships was formed.

[0044] (3) A stable and reliable ground reference was obtained by using regional imaging and statistical averaging. A depth camera or millimeter-wave radar was used to perform regional imaging of the ground directly above the working component. By performing statistical averaging of the point cloud in this area, a smooth ground reference height was obtained.

[0045] (4) The introduction of cross-medium refraction correction improves the positioning accuracy. A refraction correction model based on Snell's law is established, and the wave velocity is dynamically corrected by soil moisture content. This effectively solves the ranging error caused by changes in the medium and ensures high-precision monitoring under different soil environments.

[0046] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0047] Figure 1 This is a schematic diagram illustrating an application scenario of a tillage depth monitoring system based on cross-media positioning and surface sensing according to the present invention.

[0048] Figure 2 This is a side view of the overall application of a tillage depth monitoring system based on cross-media positioning and surface sensing according to the present invention.

[0049] Figure 3 This is a schematic diagram showing the location of the UWB receiving anchor array of the present invention installed below the rack;

[0050] Figure 4 This is a schematic diagram showing the UWB signal source of the present invention embedded at the end of the underground working component;

[0051] Figure 5 This is a schematic diagram illustrating the principle of the present invention for accurately measuring tillage depth deviation;

[0052] Figure 6 This is a schematic diagram illustrating the principle of the real-time calculation steps for the three-dimensional coordinates of the working end based on UWB-TDOA in this invention.

[0053] Figure 7 This is a schematic diagram illustrating the principle of the dynamic surface benchmark solution steps based on regional imaging in this invention.

[0054] Figure 8 This is the technical route for implementing the present invention;

[0055] Figure Labels

[0056] 1. UWB receiver anchor array; 2. Frame; 3. Tillage mechanism assembly; 301. Working end; 302. Connecting handle; 4. Depth camera; 5. UWB signal source; 6. Transparent window; 7. Soil topsoil; 8. Farmland surface; 9. Gravel; 10. Crop; 11. Filter; 12. Dust and debris. Detailed Implementation

[0057] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0058] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0059] Example

[0060] Please see Figures 1-5 This invention provides a tillage depth monitoring system based on cross-media positioning and surface sensing. By constructing a self-consistent relative measurement coordinate system that is unaffected by vehicle attitude, it enables real-time, high-precision monitoring of farmland tillage depth, including but not limited to deep loosening, plowing, and ditching.

[0061] Figure 1 This is a schematic diagram illustrating the overall application of the system of the present invention in a deep tillage operation scenario. Figure 2This is a side view of the scenario. The description uses subtilizing as an example, but those skilled in the art will understand that this scenario can also be replaced by plowing, sowing, or ditching. The subtilizing implement mainly includes a frame 2 and multiple tillage mechanism assemblies 3 fixed to it. Each tillage mechanism assembly 3 typically includes a subtilizing frame elastically connected to the frame 2, and a soil-penetrating component fixed to the lower end of the subtilizing frame. The soil-penetrating component consists of a connecting handle 302 (such as a subtilizing shovel handle or plowshare) and an operating end 301 (such as a subtilizing shovel tip or plowshare tip). During operation, the operating end 301 penetrates below the topsoil layer 7.

[0062] The tillage depth measurement system of this invention mainly consists of two major measurement subsystems and a central data fusion processing unit. The two major measurement subsystems are the direct positioning subsystem for the working components and the dynamic surface benchmark calculation subsystem. The central data fusion processing unit is not shown separately in the figures; it is typically an embedded industrial computer or controller, which can be installed on a rack or integrated into the cockpit via a host computer.

[0063] The core task of the direct positioning subsystem for the working component is to acquire the three-dimensional spatial coordinates of the working end piece 301 in real time and with high accuracy. For example... Figure 4 As shown, a miniature UWB signal source 5 (ultra-wideband pulse signal source) is directly embedded or integrated into the end of the working component using a wear-resistant and impact-resistant packaging process. The UWB signal source 5 is firmly fixed at the connection between the working end 301 and the connecting handle 302. To solve the problem of signal shielding by metal components, the UWB signal source is not completely enclosed within the metal components, but is installed within a high-strength ceramic or composite material wave-transparent window 6 in the working end 301, with the antenna main lobe pointing towards the receiving array above the rack.

[0064] The application of this invention is not limited to deep tillage shovels. For plowing operations, the UWB source can be embedded in the tip of the plowshare or the lower back of the plowshare; for seeding operations, the UWB source can be integrated into the bottom of the furrow opener. The built-in source positioning scheme of this invention can be used for any operating component that requires soil penetration.

[0065] In terms of signal frequency selection, the system preferentially uses the UWB band with a lower center frequency, taking advantage of the low attenuation of magnetic field components in soil and rock media to achieve stable communication in deep soil layers. UWB signals have nanosecond-level time resolution and excellent penetration, effectively penetrating a certain thickness of soil media, ensuring that the signal can still be effectively received even after the working end 301 is completely submerged in the soil.

[0066] Considering that soil is a complex and destructive medium, its high water content and physicochemical properties can cause attenuation, multipath effects, and refraction interference to electromagnetic wave propagation. Therefore, this embodiment adopts a series of optimized designs:

[0067] (1) Low-frequency communication: Low-frequency chips with a center frequency of 3GHz-4GHz are selected to avoid the high-frequency band with more severe attenuation and to take advantage of its stronger diffraction and penetration capabilities.

[0068] (2) Dynamic wave velocity correction and refraction model: A dynamic wave velocity correction mechanism is introduced, and the dielectric constant and wave velocity are corrected by using the Topp model in combination with real-time soil moisture parameters; and the soil-air interface is modeled, and Snell's law constraint is introduced to correct the cross-medium propagation path. See step two below for details.

[0069] (3) Medium window encapsulation: A high-strength ceramic or composite material transparent window is opened at the working end 301, which not only uses the front metal structure to protect against impact, but also ensures that the signal is transmitted upward.

[0070] (4) Split-type RF extension structure (for ultra-deep tillage or high-humidity conditions): A split design is adopted, in which the UWB antenna is extended to the upper part of the connecting handle (i.e., shallow soil layer or near the ground surface) through the RF feed line, while only the computing unit is sealed at the working end. At this time, the system uses the known geometry of the connecting handle and the fixed offset of the RF antenna relative to the working end to calculate the position of the working end.

[0071] Corresponding to the built-in UWB source 5, such as Figure 3 As shown, a UWB receiver anchor array 1 is fixedly installed on the frame component directly above the working end 301. This array consists of four UWB receiver anchors (which can be labeled as follows in the algorithm). A 1, A 2, A 3, A 4) A receiving array with a constant geometric configuration is formed. This array, together with the UWB source 5 built into the working end, constitutes a dynamic and self-consistent relative positioning coordinate system. During operation, the UWB source 5 periodically transmits pulse signals. After receiving the signals, the UWB receiving anchor array 1 records its precise signal arrival timestamps. By analyzing the differences in these timestamps (i.e., Time Difference of Arrival TDOA), the central processing unit can calculate the precise three-dimensional coordinates of the UWB source 5 relative to the UWB receiving anchor array 1 in real time, thereby obtaining the position of the working end 301 in the machine's local coordinate system.

[0072] The core task of the dynamic surface reference calculation subsystem is to accurately obtain the surface reference height directly above the working end 301. This invention employs a robust measurement method of "regional imaging + statistical averaging". A vertically downward-facing depth camera 4 (e.g., a time-of-flight (ToF) camera or millimeter-wave radar) is installed at an appropriate position on the gantry 2. The field of view of the depth camera 4 is adjusted to completely cover the surface area that the working end 301 will pass through in its forward direction. The central processing unit filters and denoises the acquired regional point cloud data and calculates its statistical characteristic values, such as the median, as the instantaneous "surface reference height".

[0073] The overall workflow of the system is as follows: During operation, the UWB source 5 built into the workpiece end continuously performs radio ranging with the UWB receiving anchor array 1 on the rack. The central processing unit calculates the three-dimensional coordinates (especially the Z-axis coordinates) of the workpiece end 301 in real time based on this. At the same time, the depth camera 4 continuously performs three-dimensional scanning of the ground surface area directly in front of the workpiece end and calculates the ground surface reference height. Finally, the central processing unit subtracts the two Z-axis coordinates obtained in the same machine coordinate system (and combines spatiotemporal consistency compensation) to obtain the closest value to the actual tillage depth at the current moment.

[0074] The core advantage of this invention lies in its unique measurement paradigm, which fundamentally avoids several major sources of error in existing technologies:

[0075] Avoid mechanical structural deformation and wear errors: as shown in the attached document. Figure 5 As shown, during deep tillage, the significant soil resistance causes an unexpected lift of the tillage mechanism assembly 3 relative to the frame 2. Furthermore, wear and increased clearance in the connecting rod pins due to prolonged operation also introduce similar nonlinear and unpredictable geometric position changes. Traditional "indirect calculation" methods measure the frame height but cannot detect this height change, leading to significant errors in tillage depth calculation. However, this invention places the UWB signal source 5 directly at the working end 301, directly measuring the position of the working end itself. Therefore, regardless of deformation or wear of the tillage mechanism assembly 3, the UWB subsystem can accurately capture the true three-dimensional coordinates of the working end, eliminating all systematic errors introduced by changes in the mechanical structure at the source.

[0076] Mitigating interference from vehicle attitude changes: When a tractor travels in the field, pitching, tilting, and bouncing are inevitable. These attitude changes constantly alter the absolute spatial orientation of the entire implement. Relying on external references (such as GPS-RTK) or assuming a constant frame attitude introduces significant errors. The "working end + frame + array" relative positioning system constructed in this invention, consisting of a UWB source 5 and a UWB receiving anchor array 1, calculates the working end coordinates relative to the frame anchor array. Since both the working end and the anchor array are fixed to the implement, they move as a whole with the tractor, and their relative geometric relationship is unaffected by tractor attitude changes. Similarly, the depth camera 4 is also fixed to the implement, and its measured ground height is relative to the implement's coordinate system. Therefore, the entire measurement system is self-consistent, completely immune to vehicle attitude changes, ensuring highly stable measurement results even in severely bumpy working environments.

[0077] This invention innovatively combines UWB direct positioning technology with regional imaging three-dimensional sensing technology to construct a direct, relative, and self-consistent measurement framework, enabling high-precision and robust real-time monitoring of the key parameter of tillage depth, and providing strong technical support for the implementation of precision agriculture.

[0078] The specific implementation steps and algorithm of the tillage depth measurement method:

[0079] Step 1: System initialization and coordinate system calibration.

[0080] To ensure that the measurement data is calculated under a unified benchmark, a local coordinate system, also known as the machine coordinate system, is first established and fixed to the machine.

[0081] (1) Define the coordinate system: take the geometric center of the UWB receiving anchor array 1 fixed on the rack 2 as the origin of the coordinate system. O= [0, 0, 0] T Let the direction of the machine's movement be defined as the positive X-axis, the horizontal rightward direction as the positive Y-axis, and the vertical upward direction as the positive Z-axis, forming a right-handed Cartesian coordinate system. This coordinate system moves along with the agricultural machinery.

[0082] (2) Anchor point coordinate calibration: After the system is installed, it is necessary to accurately measure and fix the four receiving anchor points in UWB receiving anchor point array 1. A 1 , A 2 , A 3 , A 4 The three-dimensional coordinates in the aforementioned machine coordinate system. These coordinates are known and constant. For example: A 1= [ x 1 ,y 1 ,z 1 ] T ; A 2 = [ x 2 ,y 2 ,z 2 ] T ; A 3 = [ x 3 ,y 3 ,z 3 ] T ; A 4 = [ x 4 ,y 4 ,z 4 ] T Since anchor point arrays are typically designed to be coplanar, their z The coordinate values ​​may be very close or equal, and these coordinates are used as known constants in subsequent solutions.

[0083] (3) Depth camera 4 extrinsic parameter calibration: Using the standard hand-eye calibration procedure, determine the extrinsic parameters of the coordinate system of depth camera 4 relative to the machine coordinate system. These extrinsic parameters include the rotation matrix. R cam Translation vector T cam This parameter is used to locate any 3D point in the depth camera's 4-coordinate system. P cam Transform to the machine coordinate system to obtain its corresponding coordinates. P impl The conversion relationship is as follows:

[0084] P impl = R cam P cam +T cam ;

[0085] Step 2: Real-time calculation of three-dimensional coordinates at the working end based on UWB-TDOA.

[0086] This step can determine the three-dimensional coordinates of the UWB signal source 5, which is built into the working end 301, in the machine coordinate system in real time and with high accuracy, denoted as:

[0087] P tip =[ x tip ,y tip ,z tip ] T ;

[0088] like Figure 6 As shown in the figure, this diagram illustrates in detail the principle and process of real-time calculation of three-dimensional coordinates at the work end based on UWB-TDOA.

[0089] (1) Basic Principles and Equations of TDOA: The Time Difference of Arrival (TDOA) positioning principle is adopted. The signal transmitted by UWB source 5 is received by each anchor point in UWB receiving anchor array 1. The system measures and obtains the signal arrival time of the first anchor point. i Anchor point and reference anchor point A The time difference between 1 and 1 is denoted as Δ. t i1,meas .

[0090] Considering that the UWB source 5 is located inside the soil medium while the anchor array 1 is located in the air, the signal propagation needs to cross the soil-air interface. Therefore, its propagation path no longer satisfies the straight-line propagation assumption under a single medium, but is refracted at the interface and forms a broken path.

[0091] The soil-air interface is defined as a plane in the machine coordinate system:

[0092] Π:z = Z interface ;

[0093] in, Z interface The ground reference height Z can be calculated from step three. surface It can be given directly, or its short-time moving average can be used as a stable estimate.

[0094] For any anchor point A i The signal propagation path can be decomposed into a soil segment and an air segment, and a refraction point is introduced at the interface.

[0095] Q i = [ x Qi ,yQi ,Z interface ];

[0096] It should be noted here that, due to the signal source to each different anchor point A i The propagation path is unique, therefore its refraction position at the soil-air interface is also unique, hence the use of subscripted values. Q i To provide a precise representation. (See appendix) Figure 6 For the sake of simplicity and clarity in the illustration, only a representative refraction point is schematically marked as [symbol missing]. Q The image Q This should be understood as referring to any specific refraction point. Q i Generalized representation of .

[0097] The propagation time can then be expressed as:

[0098] ;

[0099] in:

[0100] ;

[0101] v soil and v air These represent the propagation speeds of UWB signals in soil and air media, respectively.

[0102] Soil-air cross-medium refraction must satisfy Snell's law constraint:

[0103] ;

[0104] in θ soil , θ air These are the angles between the propagation directions of the soil segment and the air segment relative to the interface normal, respectively.

[0105] Therefore, it is possible to establish information about unknown coordinates. P tip The theoretical equations of TDOA:

[0106] Δ t i1 (P tip )=t i (P tip )-t 1(P tip ),i= 2, 3, 4;

[0107] With the measured value Δ t i1,meas This corresponds to the nonlinear constraint relationship for cross-medium refraction correction:

[0108] ;

[0109] (2) Coordinate calculation: minimization of time residual and iterative solution: due to the actual measurement time difference Δ t i1,meas Including noise, this problem needs to be constructed as a nonlinear optimization problem. This is achieved by solving the objective function. J ( P tip Find the minimum value of ) to obtain P tip The optimal estimate.

[0110] The objective function is defined as the sum of squared residuals between the measurement time difference and the theoretical time difference of the transmedium refraction model:

[0111] ;

[0112] The objective function is solved using iterative optimization algorithms such as the Gauss-Newton method or the Levenberg-Marquardt method. The initial values ​​for the iterations can be provided by the solution results from the previous time step to ensure the continuity of the solution and fast convergence.

[0113] In each iteration, for a given candidate P tip Refraction point Q i The piecewise distance can be determined by searching for points on the interface plane Π that satisfy Snell's law and minimize propagation time. d soil,i and d air,i Then, the theoretical TDOA value is calculated.

[0114] The final solution P tip Z-axis component z tip This refers to the vertical height of the end of the deep tillage operation at the current moment.

[0115] (3) Preferred embodiment: Adaptive activation strategy for propagation speed and refraction correction. To improve positioning accuracy, the propagation speed can be dynamically corrected. The propagation speed of a signal in soil is determined by the relative permittivity of the soil, which is related to physical parameters such as moisture content and temperature. Therefore, by integrating auxiliary sensors such as soil temperature and humidity onto the equipment, acquiring soil parameters in real time, and dynamically correcting the propagation speed based on existing physical models, the positioning accuracy under different soil conditions can be significantly improved.

[0116] Preferably, the relative permittivity of the soil can be calculated using the Topp model:

[0117] ;

[0118] in θ v Real-time volumetric water content; based on ε r The soil propagation speed can be obtained:

[0119] ;

[0120] in c The speed of light in a vacuum can be approximated as the speed of air propagation. v air ≈c .

[0121] Furthermore, the refraction correction mechanism can be set to adaptively enabled: when there is a large offset between the anchor point array and the working end in the horizontal projection, the machine posture is significantly tilted, or the soil moisture content is high, resulting in a significant change in the dielectric constant, the cross-medium refraction path model is enabled; under the condition that the anchor point is approximately above the working end and is nearly perpendicularly incident, it can degenerate into an equivalent linear propagation model to reduce the amount of computation and achieve a balance between accuracy and real-time performance.

[0122] (4) Technical Effects: By introducing a soil-air cross-medium refraction correction model, this step expands the UWB-TDOA positioning from the traditional calculation of "single-medium straight-line distance difference" to the solution of "segmented propagation time difference," which can effectively avoid false depth errors caused by abrupt changes in the medium, thereby significantly improving the three-dimensional coordinates at the working end (especially...). z tip The real-time calculation accuracy and stability under different soil moisture contents and complex attitude conditions.

[0123] Step 3: Dynamic surface benchmark solution based on regional imaging.

[0124] This step calculates a stable and reliable ground reference height from the regional point cloud data acquired by depth camera 4, denoted as . Z surface The specific processing flow and principles are as follows: Figure 7 As shown.

[0125] (1) Point cloud acquisition and coordinate transformation: Depth camera 4 captures the 3D point cloud of the surface region within its field of view. For example... Figure 7 As shown in (a), when the depth camera 4 acquires surface images, the surface environment is often quite complex. The farmland surface 8 may contain interfering elements such as crop residue 10, gravel 9, or clods of soil. Therefore, the system needs to identify different surface conditions (such as flat, gravelly, or straw-covered surfaces) and accordingly employ adaptive statistical algorithms in subsequent steps. Then, using the extrinsic parameters calibrated in step one (…),… R cam ，T cam This transforms all point cloud data from the camera coordinate system to a unified machine coordinate system.

[0126] (2) Point cloud preprocessing: To eliminate outliers introduced by dust, water droplets, and sensor noise in the working environment, the converted point cloud set is filtered. For example Figure 7 As shown in (b), during the data collection process, there is significant dust in the field. By setting filter 11 (preferably using a statistical outlier removal algorithm), noise such as dust and debris 12 suspended in the air can be effectively eliminated, which can significantly improve the accuracy of subsequent surface benchmark testing. This algorithm identifies and removes statistically significant outliers by calculating the distance distribution between each point and its neighboring points, thereby obtaining a cleaner set of valid points.

[0127] (3) Calculation of ground surface reference height: Statistical analysis is performed on the Z-axis coordinates of all points in the preprocessed effective point cloud to obtain a robust value that can represent the overall ground surface height of the area. Z surface .like Figure 7 As shown in (c), after the above processing, the system successfully extracted and calculated the reference surface for the final leveled ground. Z surface To address different surface conditions, one or more of the following statistical methods can be used:

[0128] Arithmetic mean: Calculates the average of the Z coordinates of all valid points, applicable to working conditions where the ground surface is relatively flat.

[0129] Median: The median of the Z coordinates of all valid points is taken after sorting them. This method has excellent robustness to sporadic extreme highs or lows (such as small stones or small pits).

[0130] Truncated average: After sorting, a certain percentage of the first and last data points are removed, and then the average of the remaining data is calculated. It is a compromise between the mean and the median, and can effectively avoid the influence of marginal extreme values.

[0131] Gaussian distribution fitting: Calculate the histogram of the Z-coordinates of all valid points, fit it using a Gaussian function, and take the mean of the fitted Gaussian distribution. μ As the ground surface elevation, this method effectively smooths out random fluctuations caused by minute undulations in the ground surface.

[0132] Step 4: Real-time tillage depth value fusion calculation and smooth output.

[0133] Tillage depth calculation: Under the same machine coordinate system, the surface reference height obtained in step three is... Z surface Vertical height of the working end obtained in step two z tip By performing fusion calculations, the instantaneous tillage depth value at the current moment can be obtained. D .

[0134] Considering that the detection area of ​​depth camera 4 is located in front of the travel direction of the working end 301, the surface data measured by depth camera 4 at time t is not the true surface height of the working end at the same time. Therefore, if it is directly used... D =Z surface -z tip This will lead to spatial inconsistencies between the surface benchmark and the location of the work site, resulting in systematic errors due to spatiotemporal mismatch. To address this, this step introduces a spatiotemporal consistency compensation mechanism based on work speed to align the surface data with the work site data in terms of spatial location.

[0135] (1) Spatiotemporal consistency calibration: Since the center of the measurement area of ​​depth camera 4 is located in front of the working end 301, let the longitudinal distance be . L The system obtains the operating speed of agricultural machinery in real time. V The time delay between the camera measuring the ground surface and the actual arrival of the working end at that ground surface position is calculated as follows:

[0136] ;

[0137] The system establishes a circular data buffer and timestamps the surface reference height sequence output by depth camera 4. Z surface ( t )} is cached and stored. At the current sampling time t k The system reads from the buffer. t k - Δ t The corresponding surface reference height at that time Z surface ( t k- Δ t When the buffer sampling time is... t k - Δ t When they are not completely identical, nearest neighbor search or linear interpolation can be used to obtain the result. Z surface ( t k - Δ t This ensures the continuity and real-time nature of the computation.

[0138] (2) Instantaneous tillage depth fusion calculation (spatiotemporal alignment difference): at the current sampling time t k Instantaneous tillage depth value D ( t k The calculation is corrected as follows:

[0139] D ( t k )= Z surface ( t k - Δ t )- z tip ( t k );

[0140] in: Z surface ( t k - Δ t ) Call the buffer Δ t The ground reference height stored seconds ago indicates the operating end at time [time missing]. t k The actual historical ground elevation corresponding to the physical location reached; z tip ( t k ): Vertical height of the working end obtained from real-time calculation in step two.

[0141] According to the coordinate system definition (Z-axis upward is positive), the ground surface height Z surface Typically negative or close to zero, while the working end height z tip For smaller negative values, therefore the calculation result D ( t kThe value is a positive value, which intuitively represents the depth of cultivation.

[0142] (3) Preferred Implementation: Low-Speed ​​Anomaly Handling and Output Stability Guarantee. To avoid Δ caused by excessively low operating speed. t If the value is too large or the calculation is unstable, when detected... V < V min When Δ = 0.2 m / s, the system can preferentially use Δ t Set to 0, or maintain the Δ value corresponding to the most recent effective velocity. t This ensures stable output of tillage depth calculations.

[0143] (4) Time-domain filtering and smoothing output: In order to suppress sensor noise and calculation jitter, smooth and continuous tillage depth data is provided to the control system or user interface. This can be achieved by processing the temporally and spatially aligned tillage depth time series. D ( t k Apply time-domain filters. For example, a Kalman filter can be used. By establishing a state-space model describing tillage depth and its rate of change, the Kalman filter can fuse the model's predictions with new measurements to recursively provide the optimal estimate of the current tillage depth. This process effectively filters out high-frequency noise and outputs a smooth data stream that reflects the true trend of tillage depth change, thereby improving the final reliability of the measurement results.

[0144] Please see the appendix Figure 8 The technical approach of this invention mainly includes two parallel data links and a final data fusion process, the specific logic of which is as follows:

[0145] (1) System initialization stage: First, a unified machine coordinate system is established, and the position calibration of the UWB receiving anchor array and the external parameters (rotation and translation matrix) calibration of depth camera 4 are completed, so as to provide a unified spatial reference for subsequent multi-source data fusion.

[0146] (2) Dual-link parallel processing stage:

[0147] Left-side link (operating component positioning): A UWB signal source built into the operating end transmits a signal, which is received by an anchor array on the frame, and the time difference of arrival (TDOA) is acquired. The system dynamically corrects the signal velocity based on real-time soil parameters (such as moisture content), calculates the precise three-dimensional coordinates of the operating end in the machine coordinate system using a nonlinear optimization algorithm, and extracts its vertical height component. z tip .

[0148] Right-side link (surface benchmark sensing): Depth camera 4 or radar acquires a 3D point cloud of the surface ahead of the operation. The raw point cloud is then transformed to the machine's coordinate system, followed by filtering and noise reduction to remove outliers (such as dust or splashed soil). Subsequently, statistical methods (such as taking the median or fitting the peak value of a Gaussian curve) are used to calculate a robust surface benchmark height from the regional point cloud. Z surface .

[0149] (3) Integration and output stage: Under a unified timestamp, the surface reference height is... Z surface Vertical height of the working end z tip The instantaneous tillage depth value is obtained by differential calculation. Finally, the Kalman filter algorithm is introduced to smooth the tillage depth sequence in the time domain, filtering out random fluctuations and outputting stable and high-precision real-time tillage depth data.

[0150] This technology roadmap clearly demonstrates how the present invention, through hardware and software collaboration, avoids mechanical deformation and attitude interference in principle, thereby achieving highly reliable tillage depth measurement.

[0151] To further illustrate the present invention, the tillage depth measurement method of the present invention will be quantitatively described below in conjunction with a specific deep tillage operation scenario.

[0152] Step 1: System initialization and coordinate system calibration.

[0153] Establishment of the coordinate system for the equipment: The origin is the geometric center of the UWB receiver anchor array 1 mounted on the frame directly above the deep tillage mechanism assembly. O ( 0,0,0 The direction of machine movement is defined as the positive X-axis, horizontal to the right as the positive Y-axis, and vertical upward as the positive Z-axis. The unit is meters (m).

[0154] UWB anchor point coordinate calibration: align the four UWB receiver anchor points. A 1, A 2, A 3, A 4. It is mounted in a rectangular configuration on a rigid plane below the frame. Its coordinates in the machine coordinate system are accurately measured and fixed using tools such as a laser rangefinder. The calibration results are as follows:

[0155] A 1 =[ - 0.4,0.3,0] T ; A 2 =[0.4,0.3,0] T ; A 3=[0.4,-0.3,0] T ; A 4 =[-0.4,-0.3,0] T .

[0156] These coordinates are used as known constants in subsequent calculations.

[0157] Extrinsic parameter calibration of depth camera 4: Mount a ToF depth camera 4 on the frame, ensuring its field of view covers an area of ​​approximately 0.5m × 0.5m directly in front of the working end. Using hand-eye calibration (such as the checkerboard calibration method), obtain the transformation relationship (rotation matrix) from the depth camera 4 coordinate system to the machine coordinate system. R cam Translation vector T cam In this embodiment, for the sake of simplicity, it is assumed that the camera optical axis is parallel to the machine. Z The axes are parallel, with no rotation, only translation. The calibration results are as follows:

[0158] R cam = I ;

[0159] T cam =[0.2,0,-0.1] T ;

[0160] This means that the origin of the camera coordinate system is located at (0.2, 0, -0.1) below the machine coordinate system.

[0161] Preferably, to achieve subsequent "spatiotemporal consistency compensation", the longitudinal distance between the center of the measurement area of ​​the depth camera 4 and the working end 301 in the direction of the machine's movement needs to be further defined as follows: L In this embodiment, we take... L =0.5m. The system can also acquire the operating speed in real time via wheel speed sensors, GNSS, or the implement controller's CAN bus. V It is used for time delay compensation calculation.

[0162] Step 2: Real-time calculation of the three-dimensional coordinates of the working end based on medium refraction correction and UWB-TDOA.

[0163] At a certain sampling time tk :

[0164] (1) Acquisition of TDOA measurement values: The UWB signal source 5, built into the deep tillage operation end 301, transmits signals, which are received by the UWB receiving anchor array 1. The central processing unit measures the distance of each anchor point relative to the reference anchor point. A The Time Difference of Arrival (TDOA) measurements for 1 are as follows (unit: nanoseconds, ns):

[0165] Δ t 21,meas =-1.52ns; Δ t 31,meas =-1.15ns; Δ t 41,meas =2.08ns;

[0166] (2) Coordinate calculation:

[0167] Propagation speed setting: Based on the soil moisture content and temperature measured by auxiliary sensors, the Topp model is used to dynamically correct the relative permittivity of the soil and calculate the propagation speed in the soil. v soil air propagation speed is taken v air ≈ c .

[0168] Unlike the traditional equivalent wave velocity assumption, this embodiment considers the refraction of the UWB signal at the soil-air interface when it propagates from the soil to the air. Therefore, a segmented propagation model of "soil segment + air segment" is required.

[0169] Interface establishment: Representing the soil-air interface as a plane. Π : z = Z interface .in Z interface The surface reference height can be obtained from step three. Z surface Given, that is, in this embodiment, it can be approximately taken as Zi nterface ≈ Z surface .

[0170] Transmedia propagation time difference model: For the first i Anchor points A i The signal propagation time is expressed as:

[0171] ;

[0172] in d soil,i and d air,i These represent the segmented paths of the signal in the soil and air layers, with refraction points introduced at the interfaces. It satisfies the constraints of Snell's Law:

[0173] ;

[0174] Therefore, a theoretical time difference equation is established:

[0175] Δ t i1 ( P tip )= t i ( P tip )- t 1 ( P tip );

[0176] Construct the objective function: Let the coordinates of the end point of the operation to be determined be... P tip =[ x tip , y tip , z tip ] T The objective function is the sum of squared residuals between the measured time difference and the theoretical time difference across the medium.

[0177] ;

[0178] Iterative optimization solution: The Levenberg-Marquardt algorithm is used to optimize the objective function. J ( P tip The solution is obtained through iterative optimization. The solution result at the previous time step is used. P tip ( t k-1 This is used as the initial value for this iteration. After several iterations, the algorithm converges, obtaining the optimal estimate of the coordinates of the work end at the current moment:

[0179] P tip =[0.018,-0.025,-0.452] T ;

[0180] Extracting the height of the working end: Extracting the Z-axis component from the solution results, the vertical height of the subsoil working end 301 in the machine coordinate system is obtained.

[0181] z tip =-0.452m;

[0182] Note: By introducing a cross-medium refraction correction model, the TDOA solution is upgraded from the traditional "single-medium straight-line distance difference" to a "segmented propagation time difference" solution, which can effectively reduce spurious depth errors caused by abrupt changes in the medium and improve accuracy. z tip The accuracy and stability of the solution.

[0183] Step 3: Dynamic surface benchmark solution based on regional imaging.

[0184] At the same sampling time t k :

[0185] (1) Point cloud acquisition and coordinate transformation: Depth camera 4 captured 3D point clouds of the surface area within its field of view, obtaining a total of 12,540 raw point cloud data. Using the extrinsic parameters calibrated in step one ( R cam , T cam This transforms all point clouds from the camera coordinate system to the machine coordinate system.

[0186] (2) Point cloud preprocessing: The point cloud was filtered using a statistical outlier removal algorithm. The number of neighboring points was set to 50, and the standard deviation threshold was set to 1.0. This process removed 98 outliers caused by dust or sensor noise, leaving 12,442 valid point clouds.

[0187] (3) Calculation of ground reference height: Statistical analysis of the Z-axis coordinates of 12,442 valid point clouds.

[0188] Analysis: The surface contains some debris (causing local points to appear higher) and shallow grooves from vehicle traffic (causing local points to appear lower). Using the arithmetic mean, the calculated mean Z-coordinate is -0.091m, a value influenced by both high and low points. To obtain a more robust surface benchmark, this embodiment preferably uses the median method.

[0189] Calculation: Sort the Z coordinate values ​​of all 12442 valid points and take the value at the middle position (6221st position).

[0190] Result: The obtained surface reference height is:

[0191] Z surface =-0.101m;

[0192] This median effectively ignores the influence of extreme values ​​such as stubble, and more accurately reflects the overall surface elevation of the area.

[0193] Preferably, the Z output in this step surface It can be used as an estimate of the height of the soil-air interface, i.e. Z interface ≈ Z surface This is used for the cross-medium propagation refraction correction model in step two.

[0194] Step 4: Real-time tillage depth value fusion calculation and smooth output based on spatiotemporal compensation.

[0195] (1) Time Delay Compensation Calculation: Since the detection area of ​​depth camera 4 is located in front of the travel direction of the working end 301, the surface data measured by depth camera 4 at time t is not the actual surface height of the working end at the same time. To achieve spatiotemporal consistency, this embodiment introduces time delay compensation: the system acquires the working speed in real time. V And based on the longitudinal distance between the center of the camera measurement area and the end of the work area. L Calculate latency:

[0196] ;

[0197] (2) Establish a circular buffer: The system establishes a circular data buffer to cache the surface reference height sequence according to the timestamp. Z surface ( t At the current sampling time t k Call the buffer t k - Δ t Corresponding historical surface reference height Z surface ( t k - Δ t If necessary, the nearest neighbor or linear interpolation method should be used to obtain this value to ensure continuous and stable fusion calculation.

[0198] Instantaneous tillage depth calculation: After spatiotemporal alignment, the tillage depth calculation is corrected as follows:

[0199] D ( t k )= Z surface ( t k - Δ t )- z tip ( t k );

[0200] In this embodiment, take L =0.5m, real-time operating speed V =2.0m / s, then Δ t =0.25s. Read from buffer. Corresponding ground reference height:

[0201] Z surface ( t k- Δ t ) = -0.101m;

[0202] Substitute the working end height obtained in step two z tip ( t k ) = -0.452m, therefore:

[0203] D ( t k )=(-0.101m)-(-0.452m)=0.351m;

[0204] That is, the instantaneous tillage depth at the current moment is 35.1 cm.

[0205] Time-domain filtering and smoothing: This instantaneous tillage depth value D ( t k The measured value (35.1 cm after spatiotemporal compensation) is used as the input to the Kalman filter. The smoothed tillage depth output by the filter at the previous time step was 34.8 cm. Based on the established state-space model (predicting the trend of tillage depth change), the Kalman filter fuses the predicted value with the current observation value, outputting the optimal estimated smoothed tillage depth value for the current time step.

[0206] D smooth =35.0cm;

[0207] The smoothed value is sent to the display terminal in the cab, or used in subsequent automatic tillage depth control systems.

[0208] Explanation: By introducing Δ t = L / V The time delay compensation and ring buffer mechanism ensure that the ground benchmark and the height of the working end are strictly corresponding in space, effectively solving the "looking east and hitting west" error; combined with the Kalman filter output, it can provide a stable and high-precision real-time tillage depth data stream.

[0209] Therefore, this invention employs a tillage depth monitoring system and method based on cross-media positioning and surface sensing, achieving direct three-dimensional positioning of the soil-entry operation components and eliminating mechanical structural errors at the source. By embedding an ultra-wideband signal source into the working end of the soil-entry operation component, direct, penetrating three-dimensional positioning of the soil-entry tool is achieved for the first time. This is fundamentally different from the indirect method of measuring the frame position and then performing geometric calculations in existing technologies. Because the three-dimensional coordinates of the operation component itself are directly measured, any elastic deformation, component wear, or linkage clearance generated during operation will not affect the accuracy of the positioning results.

[0210] A self-consistent relative measurement coordinate system is constructed, completely avoiding interference caused by changes in vehicle attitude. This invention establishes a dynamic, geometrically constant relative positioning system by arranging an array of UWB receiving anchor points on the frame components directly above the working component. The coordinates of the working component calculated by this system are relative to the frame anchor point array, not absolute coordinates relative to the ground or the vehicle. Therefore, regardless of the degree of pitch, roll, or bounce of the tractor or other vehicle, the measurement results of this relative coordinate system will not be affected.

[0211] A stable and reliable surface benchmark was obtained by employing regional imaging and statistical averaging methods. Addressing the problem that traditional single-point ranging methods are susceptible to interference from minor surface undulations, small clods of soil, or pits, this invention innovatively uses a depth camera or millimeter-wave radar to perform regional imaging of the surface directly above the working component, acquiring three-dimensional point cloud data within a specific area. By statistically averaging the point cloud data of this area, a statistically significant and smooth surface benchmark height is obtained.

[0212] The introduction of cross-medium refraction correction improves positioning accuracy. Addressing the cross-medium propagation characteristics of UWB signals, which require penetration through soil to reach the air, a refraction correction model based on Snell's law was established. Combined with dynamic correction of wave velocity based on soil moisture content, this effectively solves the ranging error caused by changes in the medium, ensuring high-precision monitoring under different soil environments.

[0213] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A tillage depth monitoring system based on cross-media positioning and surface sensing, characterized in that: It includes a relative positioning subsystem, a surface sensing subsystem, and a processing module, with the processing module connected to both the surface sensing subsystem and the relative positioning subsystem. The relative positioning subsystem includes a UWB signal source built into the working end of the soil-entry component and a UWB receiving anchor array fixedly installed on the frame of the tillage implement. The UWB receiving anchor array is located directly above the soil-entry component. The surface sensing subsystem includes a depth camera or millimeter-wave radar mounted on the frame of the tillage implement, the field of view of which covers the area in front of the direction of travel of the soil-penetrating component; The processing module is configured to execute the tillage depth monitoring method.

2. The tillage depth monitoring system based on cross-media positioning and surface sensing according to claim 1, characterized in that: The spatiotemporal alignment and tillage depth calculation steps performed by the processing module specifically include: Get current job speed V The longitudinal physical distance between the center of the field of view measurement area and the working end of the soil-penetrating component L ; Calculate the time delay Δ between the surface data and the measurement point to the work site. t : ; At the current sampling time t k Retrieve time from data buffer t k -Δ t Corresponding historical surface reference height Z surface ( t k -Δ t ); Get the current time t k Vertical height of the working end obtained by solution z tip ( t k The real-time tillage depth value after spatiotemporal consistency compensation is calculated using the following formula. D ( t k ): D ( t k )= Z surface ( t k -Δ t )- z tip ( t k ).

3. A method for monitoring tillage depth based on cross-media positioning and surface sensing, employing the tillage depth monitoring system based on cross-media positioning and surface sensing as described in any one of claims 1-2, characterized in that, Includes the following steps: Step 1: System initialization and coordinate system calibration. A unified machine coordinate system is established based on the rigid plane fixed to the machine tool frame. Step 2: Relative positioning of the working components. Using the ultra-wideband receiving anchor array on the frame, the signal transmitted by the UWB source at the working end of the working components is received. Based on the time difference of arrival algorithm and combined with the cross-medium propagation refraction correction model, the first three-dimensional coordinates of the UWB source in the machine coordinate system are calculated. Step 3: Dynamic surface benchmark calculation based on regional imaging. The three-dimensional point cloud data of the surface area to be cultivated in front of the soil entry operation component is obtained using a non-contact ranging device. The three-dimensional point cloud data is then converted to the coordinate system of the machine and the surface benchmark is calculated through statistical processing. Step 4: Real-time tillage depth value fusion calculation and cyclic output. Based on the current operation speed and the installation position relationship between the surface sensing subsystem and the operation components, spatiotemporal consistency compensation is performed on the surface benchmark calculated in Step 3. The vertical component of the first three-dimensional coordinate is calculated, and the difference between it and the ground reference height after spatiotemporal compensation in the machine coordinate system is used to obtain the real-time tillage depth value. The real-time tillage depth value is used as the observation input, and the Kalman filter algorithm is used to smooth the real-time tillage depth value in the time domain to establish a state-space model of tillage depth change, and the smoothed tillage depth data is output.

4. The method for monitoring tillage depth based on cross-media positioning and surface sensing according to claim 3, characterized in that: Step one includes defining the coordinate system, calibrating the anchor point coordinates, and calibrating the extrinsic parameters of the depth camera.

5. A method for monitoring tillage depth based on cross-media positioning and surface sensing according to claim 3, characterized in that: In step two, the UWB signal source is fixed at the connection between the end of the working component and the connecting handle. The working end is provided with a wave-transparent window, which is made of high-strength ceramic or composite material. The main lobe of the UWB signal source antenna faces the UWB receiving anchor array through the wave-transparent window.

6. The method for monitoring tillage depth based on cross-media positioning and surface sensing according to claim 5, characterized in that: The UWB receiver anchor array contains at least four UWB receiver anchors arranged in a coplanar manner. The UWB receiver anchor array and the UWB signal source form a dynamic relative positioning system. The UWB signal source is a low-frequency chip with a center frequency of 3GHz to 4GHz.

7. A method for monitoring tillage depth based on cross-media positioning and surface sensing according to claim 6, characterized in that: In step two, for deep soil operations or high moisture content conditions, the UWB source adopts a split radio frequency extension structure; the antenna of the UWB source is extended and arranged in the shallow soil area of ​​the middle and upper part of the connecting handle of the soil entry operation component through radio frequency feed lines, and the computing unit of the UWB source is sealed inside the working end of the soil entry operation component.

8. A method for monitoring tillage depth based on cross-media positioning and surface sensing according to claim 7, characterized in that, Step two of the cross-medium propagation refraction correction model includes the following steps: S21. Obtain soil physical parameters under the working environment; S22. Calculate the propagation speed of dynamic signals in the soil medium based on soil physical parameters. v soil And determine the speed of signal propagation in the air. v air ; The propagation speed of dynamic signals in the soil medium is calculated based on soil physical parameters, specifically including: Using the Topp model based on real-time volumetric water content θ v Calculate the relative permittivity of soil : Then calculate the propagation speed in the soil. v soil ; in, c It is the speed of light in a vacuum; S23. Using the surface height data acquired by the surface sensing subsystem, the soil-air interface is modeled as a plane. Snell's law constraint is introduced, and the UWB signal is propagated from the UWB source to the... i The path model for each anchor point is defined as the distance between the soil segment and the air segment. d air,i Based on the segmented propagation path, a propagation time model is established: ; in, P tip Let be the three-dimensional coordinates of the end of the operation to be determined; S24. Introduce a refraction point at the interface to satisfy the Snell's law constraint: ； in, θ soil , θ air These are the angles between the propagation directions of the soil segment and the air segment relative to the interface normal, respectively. S25, To measure the arrival time difference Δ t i1,meas The objective function is the sum of squared residuals of the differences between the theoretical propagation time and the segmented propagation time. J ( P tip ), and perform nonlinear optimization to solve: ； The first three-dimensional coordinates are obtained by solving for the minimum value of the objective function. P tip .

9. A method for monitoring tillage depth based on cross-media positioning and surface sensing according to claim 8, characterized in that: For the split RF extension structure, the three-dimensional coordinates of the working end to be determined in S23 to S25 are replaced with the three-dimensional coordinates of the UWB antenna center for calculation. The specific steps for calculating the first three-dimensional coordinates are as follows: First, the position of the UWB antenna center in the machine coordinate system is calculated using the nonlinear optimization solution. Then, the three-dimensional coordinates of the working end are calculated using the known geometric dimensions of the connecting handle and the fixed geometric offset of the UWB antenna relative to the working end.

10. A method for monitoring tillage depth based on cross-media positioning and surface sensing according to claim 3, characterized in that: In step three, the non-contact ranging device is a depth camera or millimeter-wave radar; the statistical processing specifically includes: filtering and denoising the acquired 3D point cloud data to remove outliers; statistically analyzing the vertical height values ​​of the filtered effective point cloud data, and selecting one of the arithmetic mean, median, or Gaussian distribution fitted peak value as the ground reference height.

Images