Pseudo-ginseng digging shovel digging control system based on electrical impedance tomography

Abstract

The invention provides a pseudo-ginseng digging shovel digging control system based on electrical impedance imaging, which relates to the field of intelligent digging and comprises an electrical impedance imaging component (10), a controllable digging component (20), an industrial computer, a PID (Proportion Integration Differentiation) controller (30) and a high-pressure air pump (40), the electrical impedance imaging assembly (10) is arranged in front of the controllable digging assembly (20) and comprises an electrode plate (11), an electrical impedance undercarriage (12), an electrode bar (13), an electrode slice (14) and an air injection cleaning head (15), and the controllable digging assembly (20) comprises a digging support (21), a digging shovel (22), an angle control hydraulic rod (23) and a depth control undercarriage (24); and the industrial computer, the PID controller (30) and the high-pressure air pump (40) are arranged behind the controllable excavating assembly (20). By means of the system, precise and lossless excavation of pseudo-ginseng can be achieved, and damage in the excavation process is effectively avoided.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent excavation, in particular to a ginseng digging spade excavation control system based on electrical impedance imaging. BACKGROUND

[0002] Ginseng is a common traditional Chinese medicine, and its core value is concentrated in dissipating blood stasis, stopping bleeding, reducing swelling and relieving pain, and it has high edible and medicinal value. At present, the planting of ginseng mainly adopts high ridge cultivation, physical rain avoidance control of diseases in the rainy season, and shade planting method of using natural shade under the forest or setting up sunshade shed; the existing ginseng excavation is mainly the excavation method combined with manual and mechanical, for example: manual use of hoe or digging spade to loosen the soil at 15-25 cm around the plant, and gradually dig along the growth direction of the root to the soil depth of 30-40 cm, and then the ginseng is excavated and harvested by extracting the main root, manual excavation can maximize avoid damage to the rhizome of ginseng in the excavation process and avoid the decrease of ginseng quality, but the excavation time is long, the efficiency is low, and the best opportunity for ginseng excavation is easily missed, and the labor cost of harvesting is high; mechanical excavation mainly uses traction harvester, self-propelled harvester and the like to excavate ginseng, which has short excavation time, high efficiency and no need for excessive participation of manual work, but the existing traction harvester and self-propelled harvester can only determine the approximate position of ginseng tuber, and cannot accurately lock the position and contour of ginseng rhizome under the soil, and at the same time, the differences in the growth process of ginseng (including the contour and size of tuber) lead to the inevitable damage of ginseng tuber by the digging spade in the excavation process, thereby affecting the quality of ginseng after excavation and the subsequent economic value. SUMMARY

[0003] In view of the problems existing in the prior art, the purpose of the present application is to provide a ginseng digging spade excavation control system based on electrical impedance imaging, which utilizes the significant difference in electrical characteristics between ginseng tuber and surrounding soil, stones and other media, constructs the electrical impedance image of the excavation area in real time, thereby accurately identifying the spatial position, size and the like of ginseng tuber; at the same time, the system automatically adjusts the excavation depth and angle of the digging spade through PID (Proportional-Integral-Derivative) control algorithm, realizes accurate and non-destructive excavation of ginseng, avoids damage in the excavation process, and significantly improves the intelligent level and economic benefit of harvesting operation.

[0004] The purpose of the present application is achieved by the following technical solutions: A kind of three seven excavating shovel excavating control system based on electrical impedance imaging, including electrical impedance imaging component, controllable excavating component, industrial computer and PID controller and high-pressure air pump;Electrical impedance imaging component is arranged in front of controllable excavating component, including electrode plate, electrical impedance landing gear, electrode stick, electrode sheet and jet cleaning head, two electrical impedance landing gears are symmetrically arranged in electrode plate end face and one end of electrical impedance landing gear is rotatably connected with electrode plate end face, the other end is connected with excavator chassis, two groups of jet cleaning head are symmetrically arranged in electrode plate bottom surface and the outer circle of two groups of jet cleaning head evenly distribute electrode stick, electrode sheet is arranged in the inside of electrode stick and along its vertical direction;Controllable excavating component includes excavating support, excavating shovel, angle control hydraulic rod and depth control landing gear, the longitudinal section of excavating support is U-shaped structure and multiple excavating shovels are evenly arranged in the groove end face of excavating support, angle control hydraulic rod and depth control landing gear are rotatably arranged on the two sides of excavating support and depth control landing gear is located on the downside of angle control hydraulic rod, one end of angle control hydraulic rod and depth control landing gear, which is away from excavating support, is connected with excavator chassis;Industrial computer and PID controller, high-pressure air pump are arranged in the rear of controllable excavating component.

[0005] Further optimization based on the above scheme, the specific steps of the excavating control system for excavating include: Step S1, system initialization and soil background imaging: before the start of excavating operation, first calibrate electrical impedance imaging component;Then, place electrical impedance imaging component into the soil of the area to be excavated, generate three-dimensional electrical conductivity background image of the area through soil background characteristics adaptive learning, which is used to reflect the electrical conductivity distribution of the current soil; Step S2, target recognition and positioning: excavator slowly advances forward, controllable excavating component controls excavating shovel to shallowly enter the soil, electrical impedance imaging component continuously collects data and images;Because the water content, tissue density of ginseng rhizome and the surrounding soil, stone exist significant difference, leading to its electrical conductivity is obviously different;In real-time acquisition process, by differentiating the real-time generated electrical impedance image and background image, the electrical conductivity abnormal area caused by ginseng rhizome is highlighted;Subsequently, image segmentation and morphological analysis algorithm is used to accurately locate the target area representing ginseng rhizome in the image, and calculate its centroid coordinates, contour size and buried depth; Step S3, excavating and posture control: industrial computer sets expected excavating depth according to target recognition result, so as to avoid the front end of excavating shovel directly hitting ginseng, and adjusts the pitch angle of excavating shovel; According to the position and size of ginseng rhizome in the image, the system plans the optimal excavating trajectory for avoiding ginseng rhizome;PID controller adjusts angle control hydraulic rod according to the deviation between planned trajectory and real-time angle of excavating shovel, to ensure that excavating shovel bypasses ginseng rhizome and excavates from the side below ginseng rhizome; Step S4, Excavation Completed and System Reset: After the excavation operation is completed, the entire system is reset; the electrical impedance imaging component and the controllable excavation component are raised. Based on the study of the adhesion characteristics of the excavation shovel and electrode rod to the soil, the vibration system set in the controllable excavation component is used to make the soil fall off the surface of the excavation shovel. At the same time, the high-pressure gas sprayed by the jet cleaning head removes the soil attached to the surface of the electrode plate and electrode rod, realizing the self-cleaning of the equipment.

[0006] Based on further optimization of the above scheme, step S1 specifically includes: Step S11, System Calibration: To ensure the accuracy and precision of the imaging, before the excavation operation begins, the electrical impedance imaging component is calibrated using a hemispherical test tank filled with a salt solution of known conductivity (e.g., 0.1 S / m). The measurement module of the electrical impedance imaging component is immersed in the hemispherical test tank, and the calibration program is run. The system automatically collects data and compares the collected data with the theoretical voltage value generated based on the finite element model to generate a correction matrix. This matrix is ​​used to compensate for system errors caused by factors such as inconsistent electrode contact impedance and wire impedance. Specifically: ; In the formula: V cor This represents the corrected voltage vector; V mea This represents the raw, unprocessed sampled voltage vector; G Represents the gain calibration matrix; O Indicates the bias calibration vector; For the i One test channel ( i =1,2,…, M Its component form is: ; In the formula: G i Indicates the first i The gain correction factor for each test channel compensates for the inconsistency in amplification between different measurement channels; O i Indicates the first i The DC bias error of each test channel compensates for the systematic offset caused by circuit common-mode voltage, contact potential difference, etc. For the gain calibration matrix G With bias calibration vector O The calibration is achieved through the acquisition of theoretical data from multiple channels. V ide Compared with actual measurement data V real Relationship acquisition, for the first i One channel: ; In the formula: Mean() represents the mean function; Var() represents the variance function; Cov() represents the covariance function; Step S12, soil background characteristics adaptive learning: Step S121, by moving the excavator to the starting position of the to-be-worked land, placing the electrical impedance imaging assembly into the soil, making the electrode sheet fully enter the surface soil, and according to the geometry of the excavator electrode, discretizing the soil area to be imaged into a finite element grid (such as discretizing the soil area into tetrahedral or hexahedral elements by using COMSOL, ANSYS, etc., and the conductivity of each grid element is an unknown to be solved, which is the basis for spatial discretization of subsequent numerical calculation) ; Step S122, by using the mode of "adjacent excitation-opposite side measurement", a full electrode pair scan is quickly completed, and background data containing a group of independent voltage measurement values are collected, that is, V bg , N represents the number of electrodes; and the conductivity distribution (initially assuming that the soil is a homogeneous medium) is initialized; Then, the current conductivity is estimated, and the following partial differential equation (current conservation equation, describing the relationship between conductivity and potential) is iteratively solved by the finite element method: ; In the formula: represents the nabla operator; represents the conductivity distribution of the k th iteration, represents the potential distribution corresponding to the k th iteration; The simulated voltage vector is obtained by solving: , wherein, F() represents the forward operator (defined by the finite element model, describing the mapping relationship from conductivity to voltage); Step S123, calculating the sensitivity matrix: The sensitivity matrix (i.e. Jacobian matrix J k ) describes the degree of influence of conductivity change on voltage: ; In the formula: represents the voltage measurement value of the i th grid element, represents the conductivity of the j th grid element;​ Step S124, since the conductivity inverse problem is an ill-posed problem (non-unique solution or sensitive to noise), a Tikhonov regularization constraint is introduced, and the objective function is: ; In the formula: L represents a regularization operator matrix (such as a Laplace operator, used to constrain the smoothness of the conductivity distribution); represents a regularization parameter (used to balance the "data fitting accuracy" and "smoothness of the solution"); , the derivative is taken to obtain the core of the optimization direction: ; In the formula: represents the transpose of the Jacobian matrix, represents the transpose of the regularization operator; Then, the Gauss-Newton linearization is performed, and the objective function is Taylor expanded at : ; In the formula: represents the conductivity update direction; An iterative solver (such as the conjugate gradient method CG) is used: ; In order to avoid iteration divergence, a damping step ( ) is introduced to ensure that the objective function is reduced: ; In the formula: represents a relaxation factor (generally in the range of 0-1, usually 0.3); The conductivity update vector and the optimal step are output; Step S125, conductivity update: ; Convergence conditions: Conductivity change threshold convergence condition: ; Objective function change threshold convergence condition: ; Voltage residual threshold convergence condition: ; In the formula: respectively represent the conductivity change threshold, the objective function change threshold, and the voltage residual threshold (and ); ​If any one of the above convergence conditions is met, the iteration is stopped; otherwise, let , return to step S122 for iteration until the requirement is met; Step S126: Perform three-dimensional Gaussian smoothing on the converged to eliminate local fluctuations: ; In the formula: represents the smoothing coefficient; represents the continuous three-dimensional space of the soil area to be imaged, r represents the node coordinates of the finite element grid, represents the position vector of the integral variable; The final output is a three-dimensional conductivity background image: .

[0007] Based on the further optimization of the above scheme, the step S2 is specifically: Step S21, by continuously collecting data, the boundary voltage data set of the soil volume in front of the excavating shovel is obtained in real time V rt ; for each frame of data, the image reconstruction module takes as the prior information and the initial value of iteration, and quickly reconstructs the real-time conductivity image ; obtain the normalized difference image of the real-time conductivity image and the background image: ; Through this difference and normalization processing, the artifacts caused by the non-uniformity of the soil itself can be greatly suppressed, and the significant conductivity changes caused by the target (ginseng or stone) newly entering the detection field are highlighted; ginseng stems usually show highly highlighted positive anomaly regions due to their high water content; Step S22, pre-process the image data, and for each voxel v xel in the image, define a three-dimensional neighborhood window v xel centered at W v (e.g. a 3x3x3 or 5x5x5 cube) to obtain the filtered new voxel value: ; In the formula: Median{} represents the median of all numerical values in the set; represents the three-dimensional neighborhood window centered at voxel (x, y, z); Step S23, perform iterative self-organizing data analysis by filtering the image Divided into three main categories, including background category C back Suspected Panax notoginseng C pn Similar to suspected rocks / cavities C stone , C back : Voxels with values ​​close to 0 C pn : Voxels with significantly positive values C stone : Voxels with significantly negative values; The clustering process iteratively performs "splits" (when the variance of one category is too large) and "merges" (when the centers of two categories are too close) until the clustering results converge; in the t-th iteration, the... j Categories C j Cluster center u j The update is as follows: ; In the formula: Indicates the first j The new center values ​​for each category in the next iteration; Indicates the first j The total number of voxels included in each category in the current iteration; Step S24: From the segmented suspected Panax notoginseng species C pn In this process, the centroid coordinates, contour size, and burial depth of the corresponding target are obtained; centroid coordinates ( x c,n ,y c,n ,z c,n ): ; In the formula: Indicate candidate target O n The number of voxels contained therein; Outline size, i.e. volume size Vol(O n ) : ; In the formula: V xel Represents the volume of a single physical voxel; Burial depth: ; wherein: Dep top , Dep cen , Dep bot respectively represent candidate target O n top buried depth, centroid buried depth, bottom buried depth.

[0008] Further optimization based on the above scheme, in step S3, the step of planning the excavation trajectory and completing the excavation (i.e. according to the position and size of the ginseng tuber in the image, the system plans the optimal avoidance excavation trajectory; the PID controller adjusts the angle of the hydraulic rod according to the deviation of the planned trajectory and the real-time angle of the excavation shovel, to ensure that the excavation shovel bypasses the ginseng tuber and excavates from below it) is as follows: Step S31, according to the position and size of the ginseng tuber in the image, the system plans the optimal avoidance excavation trajectory: According to the tuber voxel set obtained by EIT image segmentation O n , the centroid coordinates, volume, and bottom buried depth are calculated, and the principal component analysis (PCA) is used to obtain the tuber main axis direction vector n r as the reference direction of the excavation trajectory; According to the tuber volume, the equivalent envelope radius is defined: ; and the avoidance safety radius is obtained: ; wherein: T represents the resolution of the T system in the vertical direction, which is determined by the maximum detection depth of the system and the number of inversion grid divisions; The excavation starting point is: ; The starting point is located below the tuber, so that the excavation shovel "lifts" the tuber along the main axis direction, and the corresponding initial pitch angle is: ; wherein: n x0 ,n y0 ,n z0 ) represents the main axis direction vector of the corresponding point; In order to make the excavation shovel automatically reduce the pitch angle change rate when approaching the tuber, so as to avoid impact, a smooth parameter curve is used as the excavation trajectory : ; ; wherein: Dep 0 represents the starting depth of the trajectory; z(t) represents the current depth of the excavating shovel; represents the adjustment coefficient; Step S32, after the target recognition and positioning module successfully obtains the position and size of the ginseng stem and successfully plans the optimal excavating trajectory for avoiding the ginseng stem, the execution phase of the excavating operation is entered immediately; the core task of this phase is to convert the theoretical expected trajectory generated by the trajectory planning module into the precise and smooth physical movement of the excavating shovel execution mechanism through high-precision PID closed-loop servo control; The double-channel independent PID control strategy is adopted to perform parallel and real-time closed-loop control on the two key degrees of freedom, i.e., the excavating depth (h) Depth ) and the pitch angle (a Angle ) of the excavating shovel; Depth control loop: ; ; wherein: K p,d , K i,d , K d,d represent the proportional gain, the integral gain and the differential gain, respectively; D step1 (t) represents the expected excavating shovel depth, D act1 (t) represents the actual excavating shovel depth; Angle control loop: ; ; wherein: K p,a , K i,a , K d,a represent the proportional gain, the integral gain and the differential gain, respectively; D step2 (t) represents the expected pitch angle, D act2 (t) represents the actual pitch angle; Step S33, in the industrial computer (IPC), the control program is executed at a fixed time interval t(i.e., the sampling period) is executed cyclically; therefore, the above continuous-time PID formula is transformed into a discrete-time difference equation: ; In the formula: U [ T [] indicates the control output at the T-th sampling time; K p , K i , K d These represent proportional gain, integral gain, and derivative gain, respectively. e [ T ] represents the error at the Tth sampling time.

[0009] Based on further optimization of the above scheme, step S4 specifically includes: Step S41: After the excavator and imaging system complete their work and lifting, the system enters the automated self-cleaning stage; the controller drives the vibration motor to reach the vibration frequency according to the preset cleaning mode in the program. f and vibration amplitude A Vibration frequency f and vibration amplitude A The calculation method is as follows: The physical condition for soil to detach from the surface of the excavator shovel is: the maximum inertial force generated by vibration. F ine-max Greater than the adhesion between the soil and the shovel surface F adhe (The maximum adhesion between the soil and the shovel surface is obtained through pre-calibration): ; Maximum inertial force F ine-max Determined by vibration parameters: ; In the formula: m soil Indicates soil quality; therefore: ; With vibration frequency f and vibration amplitude A Vibration state and vibration time t vib Through high-frequency vibration, large clumps of soil (especially clay soil) adhering to the inner arc surface of the excavator are detached due to inertial force, achieving macroscopic and large-area coarse cleaning. Step S42, after the sensing system completes the sensing work, in order to ensure the measurement accuracy and service life of the equipment, the system presets a cleaning mode of the electrical impedance imaging component in the program of the controller; since vibration cannot effectively remove the small, wet soil particle thin layer attached to the edge of the electrode sheet, the gap and the surface, these residues will seriously affect the accuracy of EIT measurement, therefore, the electrode array needs to be finely cleaned in a non-contact manner through the impact force and blowing effect of high-speed airflow, specifically: The physical condition for the small soil particles to be stripped from the electrode surface is the impact force generated by the high-pressure airflow F jet Must be greater than the micro adhesion between the fine particles and the electrode surface F adhf (Obtained by prior calibration of the micro adhesion between the fine particles and the electrode surface): ; The impact force of high-pressure airflow on the surface is obtained by fluid mechanics: ; In the formula: H d The thrust coefficient is represented by C D P p The nozzle outlet pressure is represented by P A noz The nozzle outlet area is represented by A

[0010] The technical effects possessed by the scheme of the present application are as follows: The present application is based on the three-seven excavation of electrical impedance imaging technology (Electrical Impedance Tomography, EIT), which utilizes the significant difference in electrical characteristics between the three-seven rhizome and the surrounding soil, stone and other media, and through differential operation, image segmentation and other methods, accurately locates the centroid coordinates, contour size and burial depth of the three-seven rhizome, thereby effectively distinguishing the target from impurities, avoiding damage to the three-seven rhizome buried in the soil during the excavation process, and also avoiding the problems of misexcavation or failure to excavate the three-seven rhizome due to soil, stone and other interference. At the same time, the present application combines the PID (Proportional-Integral-Derivative) control algorithm, adjusts the excavation depth and pitch angle independently through the PID double closed-loop control strategy, plans the optimal avoidance trajectory, excavates from the side below the rhizome, and thus realizes the accurate and non-destructive excavation of the three-seven rhizome.

[0011] This invention features a fully automated system that significantly improves excavation efficiency compared to manual methods, while effectively reducing human intervention costs and preventing losses due to human error. Furthermore, the combination of vibration-based soil removal and high-pressure air jet cleaning quickly removes soil adhering to the excavator shovel and electrode surfaces, ensuring the accuracy and lifespan of subsequent measurements. Attached Figure Description

[0012] Figure 1 This is a schematic diagram of the structure of the 37-inch excavator shovel excavation control system in an embodiment of the present invention.

[0013] Figure 2 This is a schematic diagram of the electrical impedance imaging component of the 37-inch excavator shovel excavation control system in an embodiment of the present invention.

[0014] Figure 3 This is a schematic diagram of the controllable excavation component of the 37-inch excavator shovel excavation control system in an embodiment of the present invention.

[0015] Figure 4 This is a schematic diagram of the PID control structure of the 37-inch excavator shovel excavation control system in an embodiment of the present invention.

[0016] Figure 5 This is a control flowchart of the 37-inch excavator shovel excavation control system in an embodiment of the present invention.

[0017] Among them, 10, electrical impedance imaging component; 11, electrode plate; 12, electrical impedance landing gear; 13, electrode rod; 14, electrode sheet; 15, jet cleaning head; 20, controllable excavation component; 21, excavation support; 22, excavation shovel; 23, angle control hydraulic rod; 24, depth control landing gear; 30, industrial computer and PID controller; 40, high-pressure air pump. Detailed Implementation

[0018] The technical solutions in the embodiments of the present invention will be clearly and completely described below. In the following description, specific details such as specific system structures and technologies are presented for illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention.

[0019] Example 1: A digging control system for a 37mm digging shovel based on electrical impedance imaging includes an electrical impedance imaging component 10, a controllable digging component 20, an industrial computer and PID controller 30, and a high-pressure air pump 40, such as... Figure 1 As shown. The electrical impedance imaging component 10 is arranged in front of the controllable excavation component 20 to achieve target identification and positioning before excavation, thereby adjusting the excavation attitude of the excavator shovel in real time to achieve the optimal excavation trajectory; the electrical impedance imaging component 10 includes an electrode plate 11, an electrical impedance landing gear 12, an electrode rod 13, an electrode sheet 14, and an air jet cleaning head 15, as shown. Figure 2As shown, two impedance landing gears 12 are symmetrically arranged on the end face of the electrode plate 11, with one end of the impedance landing gear 12 rotatably connected to the end face of the electrode plate 11 and the other end connected to the excavator chassis. Two sets of jet cleaning heads 15 are symmetrically arranged on the bottom face of the electrode plate 11, and electrode rods 13 are evenly distributed around the outer ring of the two sets of jet cleaning heads 15. Electrode plates 15 are arranged inside the electrode rods 13 and along their vertical direction (e.g., Figure 2 As shown, three layers of electrode plates 15 are arranged vertically on the inner side of an electrode rod 13. The jet cleaning head 15 is connected to the high-pressure air pump 40 through an air guide tube, thereby realizing the jetting of high-pressure gas onto the electrode plates 15 to achieve the self-cleaning purpose of the electrical impedance imaging component 10.

[0020] The controllable excavation assembly 20 includes an excavation support 21, an excavation shovel 22, an angle control hydraulic rod 23, and a depth control landing gear 24. The longitudinal section of the excavation support 21 is a U-shaped structure (e.g., Figure 3 (As shown) and its groove end face is evenly provided with multiple digging shovels 22, the digging shovels 22 are connected to the digging support 21 by bolts and are used for digging; the digging support 21 is rotatably provided with an angle control hydraulic rod 23 and a depth control landing gear 24 on both sides respectively, and the depth control landing gear 24 is located below the angle control hydraulic rod 23 (as shown). Figure 3 As shown, the angle control hydraulic rod 23 is hinged and installed on both sides of the excavation support 21 to control the digging angle of the excavator shovel 22; the depth control landing gear 24 is hinged and installed on both sides of the excavation support 21 to control the digging depth of the excavator shovel 22. The ends of the angle control hydraulic rod 23 and the depth control landing gear 24 away from the excavation support 21 are connected to the excavator chassis; the industrial computer, PID controller 30, and high-pressure air pump 40 are located behind the controllable excavation assembly 20.

[0021] Example 2: As another preferred embodiment of the present invention, the specific steps for excavation using the excavation control system implemented in Example 1 include: Step S1, System Initialization and Soil Background Imaging: Before the excavation operation begins, the electrical impedance imaging component is first calibrated, specifically as follows: Step S11, System Calibration: To ensure the accuracy and precision of the imaging, before the excavation operation begins, the electrical impedance imaging component is calibrated using a hemispherical test tank filled with a salt solution of known conductivity (e.g., 0.1 S / m). The measurement module of the electrical impedance imaging component is immersed in the hemispherical test tank, and the calibration program is run. The system automatically collects data and compares the collected data with the theoretical voltage value generated based on the finite element model to generate a correction matrix. This matrix is ​​used to compensate for system errors caused by factors such as inconsistent electrode contact impedance and wire impedance. Specifically: ; In the formula: V cor This represents the corrected voltage vector; V mea This represents the raw, unprocessed sampled voltage vector; G Represents the gain calibration matrix; O Indicates the bias calibration vector; For the i One test channel ( i =1,2,…, M Its component form is: ; In the formula: G i Indicates the first i The gain correction factor for each test channel compensates for the inconsistency in amplification between different measurement channels; O i Indicates the first i The DC bias error of each test channel compensates for the systematic offset caused by circuit common-mode voltage, contact potential difference, etc. For the gain calibration matrix G With bias calibration vector O The calibration is achieved through the acquisition of theoretical data from multiple channels. V ide Compared with actual measurement data V real Relationship acquisition, for the first i One channel: ; In the formula: Mean() Represents the mean function; Var() Represents the variance function; Cov() This represents the covariance function.

[0022] Step S12, Soil Background Characteristic Adaptive Learning: Then (through the control of the impedance landing gear), the impedance imaging component is placed into the soil of the area to be excavated. A weak, high-frequency safe excitation current is applied to the soil using arrayed electrode plates, and the voltage response between the electrodes is measured. Based on the collected boundary voltage data, the industrial computer generates a three-dimensional conductivity background image of the area through soil background characteristic adaptive learning, reflecting the current soil conductivity distribution. Specifically: Step S121: By moving the excavator to the starting position of the plot to be worked, the electrical impedance imaging component is placed inside the soil, so that its electrode plate is completely inserted into the surface soil. According to the geometry of the excavator electrode, the soil area to be imaged is discretized into a finite element mesh (such as using COMSOL, ANSYS, etc. to discretize the soil area into tetrahedral or hexahedral elements, where the conductivity of each mesh element is the unknown quantity to be solved, and this mesh is the spatial discretization basis for subsequent numerical calculations). Step S122: Using the "adjacent excitation-opposite edge measurement" mode, quickly complete one full electrode pair scan and acquire a set of data containing... Background data for each independent voltage measurement value, i.e. V bg N represents the number of electrodes; and the conductivity distribution is initialized. (Assuming the soil was initially a homogeneous medium); Then, regarding the current conductivity Estimate the value and iteratively solve the following partial differential equation (current conservation equation, describing the relationship between conductivity and potential) using the finite element method: ; In the formula: Represents the nabla operator; Indicates the first k Conductivity distribution in the next iteration Indicates the first k The potential distribution corresponding to the next iteration; The simulated voltage vector is obtained by solving: ,in, F() This represents the forward modeling operator (defined through a finite element model, describing the mapping relationship from conductivity to voltage). Step S123: Calculate the sensitivity matrix: Through the sensitivity matrix (i.e., the Jacobian matrix) J k Describe the degree to which changes in conductivity affect voltage: ; In the formula: Indicates the first i Voltage measurement value of each grid cell. Indicates the first j The conductivity of each grid cell; Step S124: Since the inverse conductivity problem is ill-posed (the solution is not unique or is sensitive to noise), Tikhonov regularization constraints are introduced, and the objective function is: ; In the formula: LThis represents a regularization operator matrix (such as the Laplace operator, used to constrain the smoothness of the conductivity distribution). This represents the regularization parameter (used to balance "data fitting accuracy" and "solution smoothness"). right Finding the derivative is the core of determining the optimization direction: ; In the formula: This represents the transpose of the Jacobian matrix. This represents the transpose of the regularization operator; Next, Gaussian-Newton linearization is performed, and the objective function is linearized in... Taylor's expansion: ; In the formula: Indicates the direction of conductivity update; Using iterative solvers (such as the conjugate gradient method CG): ; To avoid iterative divergence, a damping step size is introduced. ( ), ensuring the objective function decreases: ; In the formula: This represents the relaxation factor (generally in the range of 0 to 1, usually taken as 0.3). Output conductivity update vector with optimal step size ; Step S125, Conductivity Update: ; Convergence condition: Convergence condition for threshold of change in conductivity: ; Convergence condition for threshold of change in objective function: ; Voltage residual threshold convergence condition: ; In the formula: These represent the threshold values ​​for the change in conductivity, the threshold value for the change in the objective function, and the threshold value for the voltage residual (and...). ); If any of the above convergence conditions are met, then stop the iteration; otherwise, let Return to step S122 and iterate until the requirements are met; Step S126: After convergence Perform 3D Gaussian smoothing to eliminate local fluctuations: ; In the formula: Indicates the smoothing coefficient; This represents the continuous three-dimensional space of the soil region to be imaged. r Represents the node coordinates of the finite element mesh. The position vector representing the integral variable; The final output is a three-dimensional conductivity background image: .

[0023] Step S2, Target Recognition and Localization: The excavator slowly advances forward, the controllable excavation component controls the excavation shovel to shallowly penetrate the soil, and the electrical impedance imaging component continuously collects data and images; due to the significant differences in water content and tissue density between the Panax notoginseng tuber and the surrounding soil and rocks, its electrical conductivity is significantly different; during real-time acquisition, the abnormal electrical conductivity areas caused by the Panax notoginseng tuber are highlighted by performing a difference operation between the real-time generated electrical impedance image and the background image; subsequently, image segmentation and morphological analysis algorithms are used to accurately locate the target area representing the Panax notoginseng tuber in the image, and calculate its centroid coordinates, contour size, and burial depth; specifically: Step S21: By continuously collecting data, obtain the boundary voltage dataset of the soil volume in front of the excavator shovel in real time. V rt For each frame of data, the image reconstruction module uses... Using prior information and initial values ​​for iteration, real-time conductivity images can be quickly reconstructed. ; Obtain the normalized difference image between the real-time conductivity image and the background image: ; This differential and normalization process can significantly suppress artifacts caused by soil heterogeneity, highlighting the significant changes in electrical conductivity caused by newly entered targets (Panax notoginseng or rocks); Panax notoginseng tubers, due to their high water content, typically exhibit… Highlighted positive abnormal areas; Step S22, for Image data is preprocessed for each voxel in the image. v xel Define a v xel 3D neighborhood window centered W v (For example, a 3×3×3 or 5×5×5 cube), obtain the new voxel values ​​after filtering: ; In the formula: Median{} This indicates that the median of all values ​​in the set is taken. Represents a three-dimensional neighborhood window centered on the voxel (x,y,z); Step S23: Perform iterative self-organizing data analysis by analyzing the filtered image. Divided into three main categories, including background category C back Suspected Panax notoginseng C pn Similar to suspected rocks / cavities C stone , C back : Voxels with values ​​close to 0 C pn : Voxels with significantly positive values C stone : Voxels with significantly negative values; The clustering process iteratively performs "splits" (when the variance of one category is too large) and "merges" (when the centers of two categories are too close) until the clustering results converge; in the t-th iteration, the... j Categories C j Cluster center u j The update is as follows: ; In the formula: Indicates the first j The new center values ​​for each category in the next iteration; Indicates the first j The total number of voxels included in each category in the current iteration; Step S24: From the segmented suspected Panax notoginseng species C pn In this process, the centroid coordinates, contour size, and burial depth of the corresponding target are obtained; centroid coordinates ( x c,n ,y c,n ,z c,n ): ; In the formula: Indicate candidate target O n The number of voxels contained therein; Outline size, i.e. volume size Vol(O n ) : ; In the formula: Vxel Represents the volume of a single physical voxel; Burial depth: ; In the formula: Dep top , Dep cen , Dep bot Representing candidate targets O n Top burial depth, centroid burial depth, bottom burial depth.

[0024] Step S3, Digging and Attitude Control: Based on the target recognition results, the industrial computer sets the desired digging depth. Dep target This avoids directly damaging the Panax notoginseng with the tip of the excavator; since the Panax notoginseng root has a distinct taproot structure, its lower end is usually most prone to breakage during excavation, therefore, the desired excavation depth is... Dep target The setup needs to completely cover the bottom of the roots to prevent root breakage and avoid excessive digging, reducing energy consumption and soil disturbance. It also needs to consider measurement uncertainties caused by EIT inversion errors and soil heterogeneity. Therefore, the initial setup is based on the bottom burial depth. Dep bot With the root center of mass z c,n Obtain the basic mining depth: ; In the formula: The morphological weighting coefficient (which belongs to [0, 1]) is 1 when the root of Panax notoginseng is slender and long, and 0.7 to 0.8 when the root is short and thick. Secondly, considering the limited spatial resolution of EIT, the variation in electrode contact impedance, and the errors caused by the non-uniformity of soil conductivity, a safety redundancy depth is introduced. : ; In the formula: The vertical resolution of the EIT system is determined by the system's maximum detection depth and the number of inversion grids. The difference in average electrical conductivity between the Panax notoginseng root region and the background soil region is used to characterize imaging contrast and positioning uncertainty. k d1 , k d2 This represents the correction factor (obtained empirically). The desired mining depth is ultimately achieved as follows: ; Simultaneously, the pitch angle of the excavating shovel is adjusted. The rationality of the excavating shovel's pitch angle directly determines the magnitude of the shear force and bending stress on the root during excavation, and is a key factor in avoiding root breakage and damage. By comprehensively utilizing the root's centroid location, root volume, and burial depth distribution characteristics, the direction of the excavating shovel's movement is made as consistent as possible with the main axis of the Panax notoginseng root. Specifically: First, within the root region, for the three-dimensional point set ( x,y,z )∈ O n Principal component analysis is performed to obtain the principal axis direction vectors of the spatial distribution of each root body: ; In the formula: n r This indicates the main growth direction of the Panax notoginseng root in three-dimensional space; The vertical height of the root is: ; The depth concentration index is: This indicator reflects the length and thinness of the root body; the larger the indicator, the thinner and longer the root body. The projection angle of the root body in the vertical plane is obtained from the direction vector of the principal axis and used as the basic pitch angle: ; Considering the depth of the root centroid burial z c,n The impact on digging posture stability, and the greater the root volume, the greater the resistance generated during digging. To avoid excessive impact from the shovel body on the root, and based on the root depth concentration index... A correction term needs to be introduced. Considering all the above factors, the desired pitch angle is: ; In the formula: z max Indicates the maximum allowable detection depth of the system; Vol max This indicates the maximum identifiable root volume preset by the system; Indicates standard morphological parameters set empirically; These represent the corresponding correction coefficients (obtained from experimental data); Step S31: Based on the position and size of the Panax notoginseng tuber in the image, the system plans the optimal avoidance excavation trajectory: Tuber voxel set obtained from EIT image segmentation O n The centroid coordinates, volume, and bottom burial depth were calculated, and the principal axis direction vector of the tuber was obtained through principal component analysis (PCA). n rAs a reference direction for the excavation trajectory; The equivalent envelope radius is defined based on the tuber volume: ; And obtain the avoidance safety radius: In the formula: The vertical resolution of the T system is determined by the system's maximum detection depth and the number of inversion grids. The starting point for excavation is: ; The starting point is located below and to the side of the tuber, allowing the digging shovel to "lift" the tuber along the main axis. The corresponding initial pitch angle is: ; In the formula: ( n x0 ,n y0 ,n z0 ) represents the principal axis direction vector of the corresponding point; To automatically reduce the rate of change of pitch angle when the excavator approaches the tuber, thus avoiding impact, a smooth parametric curve is used as the excavation trajectory. : ; ; In the formula: Dep 0 indicates the starting depth of the trajectory; z(t) Indicates the current depth of the excavator shovel; Indicates the adjustment coefficient; express t The pitch angle at any given moment; Step S32: After the target recognition and positioning module successfully obtains the position and size of the Panax notoginseng tuber and successfully plans the optimal avoidance excavation trajectory, it immediately enters the excavation operation execution stage. The core task of this stage is to convert the theoretical expected trajectory generated by the trajectory planning module into a precise and smooth physical motion of the excavator shovel actuator through high-precision PID closed-loop servo control. A dual-channel independent PID control strategy is adopted to control the digging depth of the excavator shovel separately. Depth ) and pitch angle ( Angle These two key degrees of freedom are controlled in parallel and in real-time closed-loop control; Depth control loop: ; ; In the formula: K p,d , Ki,d , K d,d These represent proportional gain, integral gain, and derivative gain, respectively. D step1 (t) Indicates the desired digging depth. D act1 (t) Indicates the actual depth of the excavator shovel; Angle control loop: ; ; In the formula: K p,a , K i,a , K d,a These represent proportional gain, integral gain, and derivative gain, respectively. D step2 (t) Indicates the desired pitch angle. D act2 (t) Indicates the actual pitch angle; Step S33: In the industrial computer (IPC), the control program operates at fixed time intervals. t (i.e., the sampling period) is executed cyclically; therefore, the above continuous-time PID formula is transformed into a discrete-time difference equation: ; In the formula: U [ T [] indicates the control output at the T-th sampling time; K p , K i , K d These represent proportional gain, integral gain, and derivative gain, respectively. e [ T ] represents the error at the Tth sampling time.

[0025] Step S4, Excavation Completion and System Reset: After the excavation operation is completed, the entire system resets; the electrical impedance imaging component and the controllable excavation component are raised. Based on the study of the adhesion characteristics of the excavation shovel and electrode rod to the soil, the vibration system set in the controllable excavation component causes the soil to fall off the surface of the excavation shovel. At the same time, high-pressure gas is sprayed through the jet cleaning head to remove the soil adhering to the surface of the electrode plate and electrode rod, realizing the self-cleaning of the equipment; specifically: Step S41: After the excavator and imaging system complete their work and lifting, the system enters the automated self-cleaning stage; the controller drives the vibration motor to reach the vibration frequency according to the preset cleaning mode in the program. f and vibration amplitude A Vibration frequency f and vibration amplitude A The calculation method is as follows: The physical condition for soil to detach from the surface of the excavator shovel is: the maximum inertial force generated by vibration. F ine-max Greater than the adhesion between the soil and the shovel surface F adhe (The maximum adhesion between the soil and the shovel surface is obtained through pre-calibration): ; Maximum inertial force F ine-max Determined by vibration parameters: ; In the formula: m soil Indicates soil quality; therefore: ; With vibration frequency f and vibration amplitude A Vibration state and vibration time t vib Through high-frequency vibration, large clumps of soil (especially clay soil) adhering to the inner arc surface of the excavator are detached due to inertial force, achieving macroscopic and large-area coarse cleaning. Step S42: After the sensing system completes its sensing work, to ensure the measurement accuracy and service life of the equipment, the system presets a cleaning mode for the electrical impedance imaging component in the controller program. Since vibration cannot effectively remove the thin layer of fine, moist soil particles adhering to the edges, gaps, and surfaces of the electrode plates, these residues will seriously affect the accuracy of EIT measurements. Therefore, it is necessary to perform non-contact, fine cleaning of the electrode array using the impact and purging action of high-speed airflow. Specifically: The physical conditions for fine soil particles to detach from the electrode surface are: the impact force generated by the high-pressure airflow. F jet It must be greater than the microscopic adhesion force between the fine particles and the electrode surface. F adhf : ; The impact force of high-pressure airflow on the surface is obtained through fluid dynamics: ; In the formula: H d Indicates the thrust coefficient; P p Indicates the nozzle outlet pressure; A noz This indicates the nozzle exit area.

Claims

1. A digging control system for a 37mm digging shovel based on electrical impedance imaging, characterized in that: The system includes an electrical impedance imaging component, a controllable excavation component, an industrial computer and PID controller, and a high-pressure air pump. The electrical impedance imaging component is located in front of the controllable excavation component and includes an electrode plate, an electrical impedance landing gear, electrode rods, electrode sheets, and jet cleaning heads. Two electrical impedance landing gears are symmetrically arranged on the end face of the electrode plate, with one end of the landing gear rotatably connected to the end face of the electrode plate and the other end connected to the excavator chassis. Two sets of jet cleaning heads are symmetrically arranged on the bottom face of the electrode plate, with electrode rods evenly distributed around the outer ring of the two sets of jet cleaning heads. Electrode sheets are arranged on the inner side of the electrode rods and along their vertical direction. The controllable excavation component includes an excavation support, an excavation shovel, an angle control hydraulic rod, and a depth control landing gear. The longitudinal section of the excavation support is a U-shaped structure, with multiple excavation shovels evenly arranged on its grooved end face. An angle control hydraulic rod and a depth control landing gear are rotatably arranged on both sides of the excavation support, with the depth control landing gear located below the angle control hydraulic rod. The end of the angle control hydraulic rod and the depth control landing gear away from the excavation support is connected to the excavator chassis. The industrial computer and PID controller, and the high-pressure air pump are located behind the controllable excavation component.

2. The three-seven digging shovel digging control system based on electrical impedance imaging according to claim 1, characterized in that: The specific steps for the excavation control system to perform excavation include: Step S1, System Initialization and Soil Background Imaging: Before the excavation operation begins, the electrical impedance imaging component is first calibrated; then the electrical impedance imaging component is placed into the soil of the area to be excavated, and a three-dimensional electrical conductivity background image of the area is generated through adaptive learning of soil background characteristics. Step S2, Target Recognition and Localization: The excavator slowly moves forward, the controllable excavation component controls the excavation shovel to enter the soil shallowly, and the electrical impedance imaging component continuously collects data and images; during real-time acquisition, the difference operation is performed between the real-time generated electrical impedance image and the background image; subsequently, image segmentation and morphological analysis algorithms are used to accurately locate the target area representing Panax notoginseng tubers in the image, and calculate its centroid coordinates, contour size and burial depth. Step S3, Excavation and Attitude Control: The industrial computer sets the desired excavation depth based on the target recognition results, and adjusts the pitch angle of the excavator shovel at the same time. Based on the position and size of the Panax notoginseng tuber in the image, the system plans the optimal excavation trajectory to avoid it; the PID controller adjusts the angle control hydraulic rod according to the deviation between the planned trajectory and the real-time angle of the excavating shovel, ensuring that the excavating shovel bypasses the Panax notoginseng tuber and excavates from its side and below; Step S4, Excavation Completed and System Reset: After the excavation operation is completed, the entire system is reset; the electrical impedance imaging component and the controllable excavation component are raised. Based on the study of the adhesion characteristics of the excavation shovel and electrode rod to the soil, the vibration system set by the controllable excavation component is used to make the soil fall off the surface of the excavation shovel. At the same time, the high-pressure gas sprayed by the jet cleaning head removes the soil attached to the surface of the electrode plate and electrode rod.

3. The three-seven digging shovel digging control system based on electrical impedance imaging according to claim 2, characterized in that: The specific steps in step S1, "calibrating the electrical impedance imaging component," are as follows: Step S11, System Calibration: Before the excavation operation begins, the electrical impedance imaging component is calibrated using a hemispherical test tank filled with a salt solution of known conductivity. The measurement module of the electrical impedance imaging component is immersed in the hemispherical test tank, the calibration program is run, the system automatically collects data, and compares the collected data with the theoretical voltage value generated based on the finite element model to generate a correction matrix, specifically: ； In the formula: V cor This represents the corrected voltage vector; V mea This represents the raw, unprocessed sampled voltage vector; G Represents the gain calibration matrix; O Indicates the bias calibration vector; For the i One test channel ( i =1,2,…, M Its component form is: ； In the formula: G i Indicates the first i The gain correction factor for each test channel compensates for the inconsistency in amplification between different measurement channels; O i Indicates the first i The DC bias error of each test channel compensates for the systematic offset caused by circuit common-mode voltage, contact potential difference, etc. For the gain calibration matrix G With bias calibration vector O The calibration is achieved through the acquisition of theoretical data from multiple channels. V ide Compared with actual measurement data V real Relationship acquisition, for the first i One channel: ； In the formula: Mean() Represents the mean function; Var() Represents the variance function; Cov() This represents the covariance function.

4. A three-seven digging shovel digging control system based on electrical impedance imaging according to claim 2 or 3, characterized in that: The step S1, "generating a three-dimensional electrical conductivity background image of the region through adaptive learning of soil background characteristics," specifically involves: Step S12, Adaptive learning of soil background characteristics: Step S121: By moving the excavator to the starting position of the plot to be worked, the electrical impedance imaging component is placed inside the soil so that its electrode plate is completely inserted into the surface soil, and the soil area to be imaged is discretized into a finite element mesh according to the geometry of the excavator electrode. Step S122: Using the "adjacent excitation-opposite edge measurement" mode, quickly complete one full electrode pair scan and acquire a set of data containing... Background data for each independent voltage measurement value, i.e. V bg N represents the number of electrodes; and the conductivity distribution is initialized. ; Then, regarding the current conductivity Estimate the partial differential equations and solve them iteratively using the finite element method: ； In the formula: Represents the nabla operator; Indicates the first k Conductivity distribution in the next iteration Indicates the first k The potential distribution corresponding to the next iteration; The simulated voltage vector is obtained by solving: ,in, F() Represents the forward operand operator; Step S123: Calculate the sensitivity matrix: Through the sensitivity matrix, i.e., the Jacobian matrix J k Describe the extent to which changes in conductivity affect voltage: ； In the formula: Indicates the first i Voltage measurement value of each grid cell. Indicates the first j The conductivity of each grid cell; Step S124: Since the inverse conductivity problem is ill-posed, a Tikhonov regularization constraint is introduced, and the objective function is: ； In the formula: L Represents the regularization operator matrix; Represents the regularization parameter; right Finding the derivative is the core of determining the optimization direction: ； In the formula: This represents the transpose of the Jacobian matrix. This represents the transpose of the regularization operator; Next, Gaussian-Newton linearization is performed, and the objective function is linearized in... Taylor's expansion: ； In the formula: Indicates the direction of conductivity update; Using an iterative solver: ； To avoid iterative divergence, a damping step size is introduced. To ensure the objective function decreases: ； In the formula: Indicates the relaxation factor; Output conductivity update vector with optimal step size ; Step S125, Conductivity Update: ； Convergence condition: Convergence condition for threshold of change in conductivity: ; Convergence condition for threshold of change in objective function: ; Voltage residual threshold convergence condition: ; In the formula: These represent the threshold for change in conductivity, the threshold for change in objective function, and the threshold for voltage residual, respectively. If any of the above convergence conditions are met, then stop the iteration; otherwise, let Return to step S122 and iterate until the requirements are met; Step S126: After convergence Perform 3D Gaussian smoothing to eliminate local fluctuations: ； In the formula: Indicates the smoothing coefficient; This represents the continuous three-dimensional space of the soil region to be imaged. r Represents the node coordinates of the finite element mesh. The position vector representing the integral variable; The final output is a three-dimensional conductivity background image: .

5. A three-seven digging shovel digging control system based on electrical impedance imaging according to claim 3 or 4, characterized in that: Step S2 specifically involves: Step S21: By continuously collecting data, obtain the boundary voltage dataset of the soil volume in front of the excavator shovel in real time. V rt ; For each frame of data, the image reconstruction module uses... Using prior information and initial values ​​for iteration, real-time conductivity images can be quickly reconstructed. ; Obtain the normalized difference image between the real-time conductivity image and the background image: ； Panax notoginseng tubers, due to their high water content, typically exhibit the following characteristics: Highlighted positive abnormal areas; Step S22, for Image data is preprocessed for each voxel in the image. v xel Define a v xel 3D neighborhood window centered W v Obtain the new voxel values ​​after filtering: ； In the formula: Median{} This indicates that the median of all values ​​in the set is taken. Represents a three-dimensional neighborhood window centered on the voxel (x,y,z); Step S23: Perform iterative self-organizing data analysis by analyzing the filtered image. Divided into three main categories, including background category C back Suspected Panax notoginseng C pn Similar to suspected rocks / cavities C stone , C back : Voxels with values ​​close to 0 C pn : Voxels with significantly positive values C stone : Voxels with significantly negative values; The clustering process involves iterative "split" and "merge" operations until the clustering results converge; during the t-th iteration, the... j Categories C j Cluster center u j The update is as follows: ； In the formula: Indicates the first j The new center values ​​for each category in the next iteration; Indicates the first j The total number of voxels included in each category in the current iteration; Step S24: From the segmented suspected Panax notoginseng species C pn In this process, the centroid coordinates, contour size, and burial depth of the corresponding target are obtained; centroid coordinates ( x c,n ,y c,n ,z c,n ): ； In the formula: Indicate candidate target O n The number of voxels contained therein; Outline size, i.e. volume size Vol(O n ) : ； In the formula: V xel Represents the volume of a single physical voxel; Burial depth: ； In the formula: Dep top , Dep cen , Dep bot Representing candidate targets O n Top burial depth, centroid burial depth, bottom burial depth.

6. The three-seven digging shovel digging control system based on electrical impedance imaging according to claim 5, characterized in that: In step S3, the steps of planning the excavation trajectory and completing the excavation are as follows: Step S31: Based on the position and size of the Panax notoginseng tuber in the image, the system plans the optimal avoidance excavation trajectory: Tuber voxel set obtained from EIT image segmentation O n The centroid coordinates, volume, and bottom burial depth are calculated, and the principal component analysis is used to obtain the tuber principal axis direction vector. n r As a reference direction for the excavation trajectory; The equivalent envelope radius is defined based on the tuber volume: ； And obtain the avoidance safety radius: In the formula: The vertical resolution of the T system is determined by the system's maximum detection depth and the number of inversion grids. The starting point for excavation is: ； The starting point is located below and to the side of the tuber, and the corresponding initial pitch angle is: ； In the formula: ( n x0 ,n y0 ,n z0 ) represents the principal axis direction vector of the corresponding point; Using smooth parametric curves as excavation trajectories : ； ； In the formula: Dep 0 indicates the starting depth of the trajectory; z(t) Indicates the current depth of the excavator shovel; Indicates the adjustment coefficient; Step S32: After the target recognition and positioning module successfully obtains the location and size of the Panax notoginseng tuber and successfully plans the optimal avoidance excavation trajectory, the excavation operation execution phase begins immediately. A dual-channel independent PID control strategy is adopted to perform parallel and real-time closed-loop control on the two key degrees of freedom of the digging shovel: digging depth and pitch angle. Depth control loop: ； ； In the formula: K p,d , K i,d , K d,d These represent proportional gain, integral gain, and derivative gain, respectively. D step1 (t) Indicates the desired digging depth. D act1 (t) Indicates the actual depth of the excavator shovel; Angle control loop: ； ； In the formula: K p,a , K i,a , K d,a These represent proportional gain, integral gain, and derivative gain, respectively. D step2 (t) Indicates the desired pitch angle. D act2 (t) Indicates the actual pitch angle; Step S33: In an industrial computer, the control program operates at fixed time intervals. t The process is cyclical; the above continuous-time PID formula is transformed into a discrete-time difference equation: ； In the formula: U [ T [] indicates the control output at the T-th sampling time; K p , K i , K d These represent proportional gain, integral gain, and derivative gain, respectively. e [ T ] represents the error at the Tth sampling time.

7. The three-seven digging shovel digging control system based on electrical impedance imaging according to claim 6, characterized in that: Step S4 specifically involves: Step S41: After the excavator and imaging system complete their work and lifting, the system enters the automated self-cleaning stage; the controller drives the vibration motor to reach the vibration frequency according to the preset cleaning mode in the program. f and vibration amplitude A Vibration frequency f and vibration amplitude A The calculation method is as follows: The physical condition for soil to detach from the surface of the excavator shovel is: the maximum inertial force generated by vibration. F ine-max Greater than the adhesion between the soil and the shovel surface F adhe : ； Maximum inertial force F ine-max Determined by vibration parameters: ； In the formula: m soil Indicates soil quality; therefore: ； With vibration frequency f and vibration amplitude A Vibration state and vibration time t vib Through high-frequency vibration, large clumps of soil adhering to the inner arc surface of the excavator are dislodged due to inertial force, achieving macroscopic and large-area coarse cleaning. Step S42: The electrode array is cleaned non-contactly through the impact and purging action of high-speed airflow. Specifically: The physical conditions under which fine soil particles are detached from the electrode surface are: the impact force generated by the high-pressure airflow. F jet It must be greater than the microscopic adhesion force between the fine particles and the electrode surface. F adhf : ； The impact force of high-pressure airflow on the surface is obtained through fluid dynamics: ； In the formula: H d Indicates the thrust coefficient; P p Indicates the nozzle outlet pressure; A noz This indicates the nozzle exit area.

Images