A slope portable base station multi-receiver hole site batch positioning method

Abstract

The application discloses a slope portable base station multi-receiver hole position batch positioning method, belongs to the technical field of engineering surveying and slope hole position positioning, and aims at solving the problems of batch coordinate fault, array relationship distortion when a plane hole position is mapped to an irregular slope, and receiver and hole number mismatch caused by portable base station relocation. The application establishes a design hole position constraint graph, uses a control anchor point to obtain a batch reference transformation with a version, constructs a joint constraint graph containing a slope sheet state and an obstacle infeasible region, integrally solves three-dimensional hole positions and node uncertainty, forms a receiver and hole number identity edge, and executes reference recalculation, slope reconstruction or identity reallocation according to point, edge, closed loop and cross-batch joint seam residual error orientation, so that the technical effect of keeping cross-batch coordinate continuity, slope array relationship and receiver identity consistency is realized.

Description

Technical Field

[0001] This invention relates to the field of engineering surveying and slope hole location technology, and in particular to a method for batch location of multiple receiver holes on a slope using a portable base station. Background Technology

[0002] Engineering surveying operations such as slope drilling, anchoring holes, and borehole layout typically require converting planar hole positions from design drawings into three-dimensional locations on-site. Portable base stations are easy to set up quickly in areas with insufficient fixed facilities, such as mountainous terrain and slopes, and can be used with multiple receivers to simultaneously stake out or remeasure multiple hole positions. Current operations usually involve first setting up a single base station to establish a local coordinate reference, and then projecting or transforming the design coordinates onto the slope surface one by one.

[0003] However, the temporary installation location and orientation of base stations can change during relocation, making it difficult to directly maintain continuous local coordinates across different batches. Furthermore, slope, elevation differences, local unevenness, and obstacle areas can alter the distance and orientation of planar boreholes on the actual slope. If only the coordinate deviation of a single borehole is examined, overall translation, overall rotation, borehole spacing distortion, and cross-batch seams may still be overlooked.

[0004] When multiple receivers operate simultaneously, mismatches may occur between equipment identification, calibration results, current location, and borehole number. Existing methods lack a mechanism to jointly solve for batch baselines, slope conditions, borehole topology, and receiver identity within the same constraint structure and to perform error-oriented backoff.

[0005] Therefore, there is a need for a method for batch positioning of multiple receiver apertures on a portable base station on a slope that can overcome the shortcomings of the existing technology. Summary of the Invention

[0006] One objective of this invention is to propose a batch positioning method for multiple receiver apertures on a slope using a portable base station. Addressing the problems in existing technologies such as batch coordinate discontinuity caused by multiple base station installations, distortion of the overall relationship of designed apertures on undulating slopes, and mismatch between receiver identity and aperture number, this invention proposes a technical solution that uses a designed aperture constraint diagram, a batch reference transformation with versioning, and a slope representation including slope patch states and infeasible regions to form a joint constraint diagram. This solution comprehensively solves for the uncertainty of the target three-dimensional aperture positions and nodes, forming a one-to-one identity edge between the receiver and the aperture number. Furthermore, it utilizes four levels of residuals—point, edge, closed loop, and cross-batch seam—to identify the source of reference, slope, or identity error, updating only the routed object. This invention achieves the technical effect of maintaining cross-batch coordinate continuity, slope array geometric relationships, and receiver identity consistency.

[0007] This invention provides a method for batch positioning of multiple receiver apertures on a portable base station on a slope, comprising:

[0008] S1. Based on the design hole position plane coordinates, hole position number and design adjacency relationship, obtain the design hole position constraint diagram;

[0009] S2. After the base station is set up or relocated, the batch reference transformation with version identifier is obtained according to the corresponding coordinates of the control anchor point in the project coordinate system and the batch local coordinate system, and the batch data is unified to the project coordinate system.

[0010] S3. Based on the slope sampling data in the project coordinate system, construct a slope representation including slope patch state and obstacle infeasible region. Combine the design hole location constraint diagram, batch datum transformation and the slope representation into a joint constraint diagram, and solve the overall solution to obtain the target three-dimensional hole location set and related node uncertainties.

[0011] S4. Based on the current location of the receiver, the topology of the designed hole position constraint diagram, and the version identifier, determine a one-to-one identity mapping between each receiver and the hole position number in the target three-dimensional hole position set, form an identity edge representing the identity mapping in the joint constraint diagram, and obtain the measured three-dimensional coordinates of each hole position.

[0012] S5. Based on the measured 3D coordinates and joint constraint diagram, obtain the hole location residuals, adjacent edge residuals, closed-loop residuals, and cross-batch joint residuals. Recalculate the common residuals of the entire array as the reference, route the residuals within the same slope patch as the local reconstruction of the slope, and route the residuals reduced by identity edge replacement as identity redistribution. Only update the routed nodes or identity edges, and output the qualified 3D hole location set with the aforementioned identity mapping when the uncertainty of each residual and node meets the acceptance conditions.

[0013] Optionally, S1 includes:

[0014] Establish hole position nodes according to the designed hole position plane coordinates;

[0015] Connect the hole nodes with design adjacency relationships as adjacency edges, and record the design distance and design orientation between the corresponding holes in the adjacency edges;

[0016] Select a closed loop from the adjacent edges that can cover each hole node, record the node order and design closure relationship of the closed loop, and form the design hole constraint diagram.

[0017] Optionally, S2 includes:

[0018] Select no fewer than three non-collinear control anchor points, and estimate the batch baseline transformation based on the corresponding coordinates and measurement covariance of each control anchor point. When the number of control anchor points is insufficient or they are collinear, reselect control anchor points.

[0019] Calculate the posterior covariance of the batch benchmark transformation based on the coordinate residuals of the control anchor points after the transformation;

[0020] When the installation location, attitude, or control anchor point combination of the portable base station changes, a new version identifier is generated, and the batch reference transformation, the posterior covariance, the control anchor point identifier used, and the version identifier are stored together.

[0021] Optionally, S3 includes:

[0022] The slope sampling data is obtained by using at least one of the following: slope point cloud, discrete elevation points, or digital elevation model. The slope sampling data is divided into slope patches, and the normal, fitting variance, roughness, and local metric parameters are calculated based on the sampling points in each slope patch.

[0023] The infeasible region of the obstacle is formed based on the obstacle boundary and the construction clearance.

[0024] Within the joint constraint diagram, batch transformation common nodes, slope patch state nodes, and borehole nodes are set up. Anchor point transformation factor, slope fit factor, local metric factor, array orientation factor, minimum closed loop factor, and obstacle net distance factor are constructed. The borehole nodes and their posterior covariance are jointly solved using the weighted residuals of each factor to obtain the target three-dimensional borehole set.

[0025] Optionally, S4 includes:

[0026] Using the unique device identifier of each receiver as a stable identity key, the device calibration information is associated with the corresponding unique device identifier, and the current position of the receiver is corrected according to the device calibration information;

[0027] The receiver adjacency relationship is formed based on the corrected position or the ranging relationship between the receivers;

[0028] The version identifier of the batch reference transformation used when the target three-dimensional hole position record was generated is used to associate the current position of the receiver and the corresponding measurement record with the version identifier of the batch reference transformation used when the data was generated, and the version identifier is used as the version identifier of the receiver.

[0029] Calculate the positional cost between the corrected position and each target 3D hole location, the topological cost between the receiver adjacency relationship and the design hole location constraint diagram, and the version cost between the version identifier used by the receiver and the version identifier recorded by the target 3D hole location;

[0030] In the joint constraint graph, identity edges are set between receivers and hole numbers. According to the one-to-one constraint that each receiver corresponds to a hole number and each hole number corresponds to a receiver, the identity mapping that minimizes the sum of the location cost, the topology cost, and the version cost is obtained.

[0031] Optionally, S5 includes:

[0032] The residual of the borehole position is obtained by the difference between the measured three-dimensional coordinates of each borehole position and the coordinates of the corresponding target three-dimensional borehole position;

[0033] In the project coordinate system, a measured adjacent edge is formed based on the measured three-dimensional coordinates of two holes with a design adjacency relationship, and a target adjacent edge is formed based on the corresponding two target three-dimensional holes. The residual of the adjacent edge is obtained from the length difference and direction difference between the measured adjacent edge and the target adjacent edge.

[0034] According to the node order of the closed loop, the absolute values ​​of the length difference and direction difference between each measured adjacent edge and the corresponding target adjacent edge are weighted and summed to obtain the closed loop residual.

[0035] The cross-batch joint residual is obtained from the coordinate difference between adjacent measurement batches at a common hole position or a common control anchor point;

[0036] The above four levels of residuals are compared with the corresponding acceptance thresholds pre-set according to the requirements of hole position positioning accuracy, hole relationship accuracy, array closure accuracy and cross-batch consistency to obtain the four levels of acceptance results.

[0037] Furthermore, the sensitivity of the translational component and rotational component of the hole location residuals and adjacent edge residuals of multiple hole locations to the nodes corresponding to the batch datum transformation in the joint constraint diagram are calculated, and the consistency of residual direction and the difference between the weighted sum of residuals before and after fitting the rigid body rotation are calculated. The difference is used as the rotation fitting benefit.

[0038] When the translation component sensitivity reaches a preset translation sensitivity threshold and the residual direction consistency reaches a preset correlation threshold, or when the rotation component sensitivity reaches a preset rotation sensitivity threshold and the rotation fitting gain reaches a preset rotation gain threshold, the corresponding residual is routed to the node corresponding to the batch baseline transformation, and the batch baseline transformation is re-estimated using the corresponding coordinates of the control anchor point.

[0039] Retain qualified nodes and identity edges that were not selected in this residual route and whose update impact assessment is not affected by the batch baseline transformation update;

[0040] Furthermore, for boreholes belonging to the same slope patch, the normal deviation between the model normal of the slope patch and the measured normal obtained by fitting the neighborhood of the measured three-dimensional coordinates of the corresponding borehole is calculated. The correlation between the normal component of the borehole site residual and the normal deviation and the fitting variance is calculated, and the sensitivity of the borehole site residual to the state node of the corresponding slope patch is calculated.

[0041] When the correlation reaches a preset slope correlation threshold and the sensitivity reaches a preset slope sensitivity threshold, the corresponding residual is routed to the slope patch state node and the hole node affected by the slope patch. The slope sampling data within the range of the slope patch is reacquired, and the normal, fitting variance, roughness and local metric parameters of the slope patch are updated. The affected target three-dimensional hole is locally resolved and the corresponding residual is recalculated.

[0042] Retain qualified nodes and identity edges that were not selected in this residual route and whose update impact assessment is not affected by the slope patch update;

[0043] Furthermore, candidate permutations that satisfy the one-to-one constraint are generated between the identity edges corresponding to adjacent receivers, and the weighted sum of the adjacent edge residuals and closed-loop residuals before and after the permutation is calculated respectively.

[0044] The difference between the weighted sums before and after the replacement is taken as the identity replacement benefit. When the identity replacement benefit reaches a preset benefit threshold and the cost of the replaced version meets the preset version consistency condition, the corresponding identity edge is updated, and the identity edges that did not participate in the candidate replacement are retained.

[0045] Furthermore, the posterior covariance of the nodes corresponding to the batch baseline transformation in the joint constraint diagram is used as the baseline uncertainty, and the posterior covariance of the nodes corresponding to the slope patch state and the hole position is used as the slope patch uncertainty and the hole position uncertainty, respectively.

[0046] The acceptance criteria include that the baseline uncertainty, the slope patch uncertainty, and the hole location uncertainty are all not greater than their respective preset covariance thresholds, the hole location residual, the adjacent edge residual, the closed loop residual, and the cross-batch joint residual are all not greater than their respective acceptance thresholds, and the target three-dimensional hole location is located outside the obstacle infeasibility region and the distance from the obstacle boundary is not less than the preset obstacle net distance.

[0047] When the baseline uncertainty exceeds the corresponding threshold, it is routed to the node corresponding to the batch baseline transformation to perform baseline recalculation.

[0048] When the uncertainty of the slope patch or the uncertainty of the hole location exceeds the corresponding threshold, it is routed to the corresponding slope patch status node or hole location node to perform local slope reconstruction;

[0049] When the target 3D hole does not meet the obstacle infeasibility region or the preset obstacle clearance condition, it is routed to the corresponding slope patch state node and hole node, and the local re-solution is performed based on the obstacle clearance constraint.

[0050] The above-mentioned targeted update is treated as an abnormal rollback. Only the routed nodes or identity edges are updated. After each update, the affected objects are unlocked and the corresponding level 4 residuals and node uncertainties are recalculated. After the verification is passed, the affected objects are relocked.

[0051] The qualified three-dimensional hole set is output after all four levels of acceptance results, node uncertainty and obstacle clearance meet the acceptance conditions.

[0052] The beneficial effects of this invention are:

[0053] 1. By controlling the anchor points to obtain the batch reference transformation with version identifier and posterior covariance, and unifying the data of different batches to the project coordinate system, the hole position data before and after the base station relocation can maintain a traceable coordinate reference, reducing the impact of cross-batch coordinate discontinuity on array positioning.

[0054] 2. By incorporating the design hole location constraint diagram, slope patch state, obstacle infeasible region, and batch datum transformation into the overall solution of the joint constraint diagram, hole location fit, hole-to-hole relationship, array orientation, closed loop, and obstacle clearance can be constrained simultaneously, avoiding the situation where individual hole positioning is qualified but the overall array relationship is distorted.

[0055] 3. By establishing a one-to-one identity edge between the receiver and the hole number, and based on the fourth-level residual and node uncertainty, the anomaly is routed to the baseline recalculation, local slope reconstruction or identity redistribution. Only the affected objects are updated, other qualified results can be retained and a verifiable local recovery closed loop can be formed. Attached Figure Description

[0056] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0057] Figure 1 This is a flowchart of the method for batch positioning of multiple receiver apertures on a slope portable base station according to the present invention.

[0058] Figure 2 This is a flowchart of the error diagnosis and targeted update process in S5 of the present invention. Detailed Implementation

[0059] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0060] refer to Figures 1-2 A method for batch positioning of multiple receiver apertures on a portable base station on a slope, comprising:

[0061] S1. Based on the design hole position plane coordinates, hole position number and design adjacency relationship, obtain the design hole position constraint diagram;

[0062] S2. After the base station is set up or relocated, the batch reference transformation with version identifier is obtained according to the corresponding coordinates of the control anchor point in the project coordinate system and the batch local coordinate system, and the batch data is unified to the project coordinate system.

[0063] S3. Based on the slope sampling data in the project coordinate system, construct a slope representation including slope patch state and obstacle infeasible region. Combine the design hole location constraint diagram, batch datum transformation and the slope representation into a joint constraint diagram, and solve the overall solution to obtain the target three-dimensional hole location set and related node uncertainties.

[0064] S4. Based on the current location of the receiver, the topology of the designed hole position constraint diagram, and the version identifier, determine a one-to-one identity mapping between each receiver and the hole position number in the target three-dimensional hole position set, form an identity edge representing the identity mapping in the joint constraint diagram, and obtain the measured three-dimensional coordinates of each hole position.

[0065] S5. Based on the measured 3D coordinates and joint constraint diagram, obtain the hole location residuals, adjacent edge residuals, closed-loop residuals, and cross-batch joint residuals. Recalculate the common residuals of the entire array as the reference, route the residuals within the same slope patch as the local reconstruction of the slope, and route the residuals reduced by identity edge replacement as identity redistribution. Only update the routed nodes or identity edges, and output the qualified 3D hole location set with the aforementioned identity mapping when the uncertainty of each residual and node meets the acceptance conditions.

[0066] In this specific embodiment, S1 includes:

[0067] This specific implementation method takes twenty-four designed borehole locations in a slope anchoring operation area as the object. The borehole locations are arranged in six columns and four rows. Six receivers complete the layout and re-measurement in four rounds. The data processing terminal reads the borehole number, the horizontal and vertical coordinates in the project coordinate system, the design row and column number, the borehole diameter category, and the design adjacent number from the design result file. The coordinates are uniformly expressed in meters, and the borehole number is used as the primary key that remains unchanged across steps. The terminal performs uniqueness checks on the borehole records, coordinate validity checks, and adjacent reference integrity checks. Duplicate numbers are rejected for import according to the issued version of the design result file. Records with missing coordinates or non-existent adjacent numbers are written to an exception list and the construction of that batch of design drawings is stopped, thus forming a complete set of design borehole records.

[0068] The data processing terminal establishes a hole location node for each valid design hole location record. The node fields include at least the hole location number, design plane coordinates, row and column numbers, hole diameter category, node status, and design version. In this specific implementation, adjacent rows, adjacent columns, and oblique reinforcement relationships explicitly given in the construction drawings are used as design adjacency relationships. Hole location pairs not included in the design adjacency list are not generated with adjacent edges. In the case where the same undirected hole location pair is recorded repeatedly, the terminal forms an edge key by lexicographically ordering the two endpoint numbers and retains only one adjacent edge, so that subsequent closed-loop and topology costs all reference the same edge object.

[0069] For the hole node and hole position node The adjacent edges formed are based on the terminal. and Calculate the design distance and design orientation, where the symbol... Indicates the index of the first hole node of the adjacent edge, symbol This represents the index of the second hole node adjacent to the edge, both of which are taken from the set of valid hole nodes in the current design hole position constraint diagram. These symbols are only used for numbering discrete hole positions; units or dimensions are not applicable. This represents the design x-coordinate of the first hole location node in the project coordinate system, in meters, with the symbol... This represents the design ordinate of the first hole location node in the project coordinate system, in meters, with the symbol... This represents the design x-coordinate of the second hole location node in the project coordinate system, in meters, with the symbol... This represents the design ordinate of the second hole location node in the project coordinate system, in meters, with the symbol... Indicates the design distance between two holes, in meters, symbol This function represents the design orientation from the first hole to the second hole, expressed in radians. The arctangent function is used to determine the quadrant based on the difference between the vertical and horizontal coordinates. The calculation results, along with the hole positions at both ends, the edge key, the design version, and the valid mark, are written into the adjacent edge record. If the distance is less than the zero distance protection value of 0.01 meters set in this specific implementation, the corresponding edge is marked as invalid and returned to the designer for review.

[0070] In this specific embodiment, the data processing terminal performs a depth-first traversal of the adjacency graph starting from the node with the smallest hole number. First, it obtains a spanning tree that covers all hole nodes. Then, it forms a basic closed loop with each non-tree adjacency edge and the unique path connecting its two endpoints in the spanning tree. All basic closed loops constitute a candidate closed loop set. The terminal sorts the closed loops in ascending order using the number of edges contained in the closed loop as the first sorting key and the hole number sequence as the second sorting key. It retains the candidate closed loops that can increase hole coverage or adjacency edge coverage and writes them into the closed loop selection result until all twenty-four hole nodes belong to at least one closed loop and each adjacency edge used for acceptance belongs to at least one closed loop. If the graph is not connected or there are still holes that cannot be covered by a closed loop, it outputs a design adjacency relationship incomplete flag and stops entering S2.

[0071] For each retained closed loop, the terminal writes the node number sequence arranged in a uniform clockwise direction, the corresponding adjacent edge key sequence, the design distance of each edge, the design orientation, and the consistent flag of the start and end nodes of the closed loop into the closed loop record, and sets the design plane displacement vector along the directed edge of the closed loop as the design closure relationship; in order to avoid the same geometric closed loop being repeated due to different starting points, the terminal rotates the smallest hole position number to the beginning of the sequence, and retains the smallest lexicographical value in the clockwise and counterclockwise sequences as the closed loop key, and deletes the other records with the same closed loop key;

[0072] After completing the above processing, the terminal combines the hole location node table, adjacent edge table, and closed-loop table into a design hole location constraint diagram, generates a diagram version identifier determined by the project number, design result issuance number, and diagram construction time, and calculates the content summary of the three tables as an integrity verification value. The design hole location constraint diagram is written to the project database as a read-only object. Subsequently, S3 reads its nodes, adjacent edges, and closed-loop constraints, S4 reads its topology and hole location number, and S5 reads its design distance, design orientation, and closure relationship. This ensures that the same design relationship remains consistent in joint solution, identity assignment, and level four residual acceptance.

[0073] In this specific embodiment, S2 includes:

[0074] After each deployment or relocation of the portable base station, the data processing terminal establishes a measurement batch record. In this specific embodiment, the first deployment is used for the first twelve boreholes, and the second deployment is used for the last twelve boreholes, including a re-measurement of the four common boreholes at the joint. The batch record includes the batch number, base station equipment number, deployment time, antenna height, batch local coordinate system definition, receiver original coordinate record, and control anchor point observation record. The control anchor point observation record uses the anchor point number as the key to simultaneously store the project coordinates, batch local coordinates, timestamps of both types of coordinates, and three-dimensional measurement covariance. The coordinate unit is meters, and the unit of each element of the three-dimensional measurement covariance is square meters, which is converted to square millimeters in the audit field. Records with timestamps differing by more than one second will not be included in the same baseline solution.

[0075] The terminal selects no fewer than four control anchor points from the visible control anchor points in this batch to ensure that the estimation condition of no fewer than three non-collinear anchor points as described in the claims is met and that pre-release leave-one verification can be performed. In this specific implementation, four permanent control points are preferred, and the non-collinearity criterion is that the area of ​​the triangle formed by any three points on the horizontal plane is not less than 0.5 square meters and the minimum side length is not less than 5 meters. When there are fewer than four valid anchor points, all three candidate points do not meet the area criterion, or the three-dimensional standard deviation of any anchor point is greater than 0.03 meters, the terminal does not release a batch change version and does not use the previous batch change. Instead, it outputs a command to supplement the measurement of the fourth anchor point or to re-observe the control anchor points until a combination of anchor points that meets the conditions is obtained.

[0076] For the The first of the measurement batches One control anchor point, the terminal adopts Establish a rigid body transformation from batch local coordinates to project coordinates, where the symbol... This represents the batch index in the current project's valid measurement batch set, symbol... Indicates the first Anchor indexes in the effective control anchor set used for batch estimation; both are used only for discrete record numbering, and units or dimensions are not applicable; symbols. This represents the three-dimensional coordinate vector of the control anchor point in the project coordinate system, with each component in meters. (symbol) Indicates the same control anchor point at the 1st Three-dimensional coordinate vectors in the local coordinate system of the batch, with each component in meters, and symbol... Indicates the first The three-dimensional rotation matrix from the batch local coordinate system to the project coordinate system, and which is a dimensionless orthogonal matrix, with the symbol... Indicates the first The local coordinate system origin of the batch is translated into the project coordinate system by the translation vector, and each component is in meters. The terminal first eliminates the translation by weighting the centroid, then performs singular value decomposition on the weighted cross-covariance matrix to obtain the rotation matrix, and finally substitutes back to obtain the translation vector.

[0077] The terminal minimizes Estimate the batch reference transformation, wherein the symbol Using the valid measurement batch index immediately preceding the aforementioned rigid body transformation formula, the symbol... Indicates the first The batch participates in the estimation of the effective control anchor index and takes values ​​from 1 to 1. Both are discrete numbered and their units or dimensions are not applicable; symbols Indicates the first The dimensionless weighted sum of residuals for each batch, with the following sign. Indicates the first The batch is used to estimate the number of control anchors and is a dimensionless positive integer count, with the sign... Indicates the first Batch No. The difference between the observed coordinates of each anchor point and the transformed coordinates, with each component in meters, superscript. Represents the transpose of a vector or matrix, symbol This represents the inverse of the anchor point coordinate residual covariance matrix, with units per square meter. The residual covariance is obtained by adding the project coordinate measurement covariance to the local coordinate measurement covariance after rotation propagation. The terminal uses three iterations for reweighting. Anchor points with standardized residuals exceeding 3 are reduced in weight in this iteration. If there are fewer than four effective anchor points after weighting or if a collinear combination is formed again, the system returns to measure the fourth anchor point or reselects the branch.

[0078] After the reference transformation converges, the terminal follows... Calculate the posterior covariance, where the sign is... The valid measurement batch index in the aforementioned rigid body transformation formula is used, but the unit or dimension is not applicable; the symbol is... Indicates the first The six-dimensional posterior covariance matrix of batch rotation and translation components, where the variance of the rotation component is in radians squared and the variance of the translation component is in square meters, with the cross terms using corresponding product units. (Symbols omitted). The Jacobian matrix represents the increment of the six-dimensional transformation of the metric residuals of all effective anchor points. Its three columns corresponding to the rotation increment are dimensionless (meters per radian), and its three columns corresponding to the translation increment are dimensionless. The symbols are... Indicates that the aforementioned The effective block weights retained after the last iteration of reweighting are arranged in a block diagonal matrix according to the observation order, and the unit of each coordinate residual block is per square meter. Indicates matrix transpose, superscript This indicates matrix inversion; before calculating the degradation condition number, the terminal constructs a diagonal scale matrix with 0.01 radians as the reference scale for three rotation increments and 0.01 meters as the reference scale for three translation increments, and performs dimensionless scale normalization on the normal matrix; if the normalized normal matrix condition number is greater than 10 to the power of 8 set in this specific implementation, the terminal determines that the anchor point geometry is degraded and reselects the anchor point combination, and does not output the batch transformation;

[0079] To verify the batch baseline transformation, the terminal performs leave-one-out cross-validation on each of the four valid control anchor points. Each time, the transformation is recalculated using the remaining three anchor points, and the project coordinates of the anchor points to be left out are predicted. The residual modulus of the left-out anchor points must not exceed 0.05 meters, and the root mean square of the residuals after four rounds must not exceed 0.03 meters. If any test fails, the terminal removes an abnormal anchor point from the standardized residuals in descending order and adds a new anchor point that meets the non-collinearity criterion from the candidate control points. The estimation and leave-one-out cross-validation are then re-executed. If no new anchor point can be added after two consecutive rounds, the batch status is set to baseline invalid, and on-site remeasurement is required.

[0080] The terminal performs back-calculation on the coordinates of all valid anchor points involved in the estimation using the current transformation. It obtains the mean residual vector by the arithmetic mean of the three-dimensional residual vectors of each anchor point and checks that the Euclidean modulus of this mean vector is no greater than 0.005 meters, the median of the standardized residuals is no greater than 1, and the maximum standardized residual is no greater than 3. The 0.005-meter threshold is derived from the upper limit of the benchmark system deviation specified in the project control measurement scheme and is written into the benchmark verification parameter record with version. If any check exceeds the limit, it returns to the reselection branch. If there are no available replacement anchor points for two consecutive rounds, it requires re-measurement and sets the batch status to benchmark invalid. The terminal writes the test sample, residual statistics, pass flag, and test time into the benchmark verification record. This record, together with the posterior covariance, serves as the conditions for the release of the transformation version. Transformations that fail verification cannot be used for unified receiver data and slope sampling data, thus providing clear calibration verification results and infeasible branches.

[0081] After S5 returns the hole location residuals and adjacent edge residuals for multiple hole locations, the terminal first divides the hole location and adjacent edge length residuals by 0.05 meters and 0.06 meters respectively, the direction residual by 2 degrees, the closed-loop residual by the dimensionless reference 1, and the cross-batch seam residual by 0.04 meters to form dimensionless whitening residuals. The terminal then applies fixed perturbations to the three translation components and three rotation components of the transformation node of this batch one by one. The translation perturbation is divided by the 0.01-meter reference scale, and the rotation perturbation is divided by the 0.01-radian reference scale to form dimensionless increments. The change in whitening residuals is divided by the corresponding dimensionless increments to obtain the dimensionless translation component sensitivity and the dimensionless rotation component sensitivity. The terminal uses the mean of the dot product of each residual unit direction and the weighted average residual direction as the residual direction consistency, and records the weighted sum of the residuals before and after the rigid body rotation fitting. All statistics, reference scales, and residual record numbers are stored together.

[0082] Terminal according to Calculate the rotation fitting gain, where the symbol The aforementioned valid measurement batch index is used only for discrete record numbering; units or dimensions are not applicable; symbols Indicates the first The dimensionless rotational fitting gain of the batch, with a larger value indicating that the rotational correction can eliminate more residuals. (Superscript) Indicates the application of rotational fitting, superscript Indicates the state before fitting, superscript Indicates the fitted state; the three superscripts are only used to distinguish the purpose or state, and are not applicable to units or dimensions; symbols Indicates the first The dimensionless weighted sum of residuals before rigid body rotation fitting in batches, with sign... Indicates the first The dimensionless weighted sum of the residuals after fitting the rigid body rotation in the batch; in this specific implementation, the translation sensitivity threshold and rotation sensitivity threshold are both set to 0.60, the direction consistency threshold is set to 0.80, and the rotation gain threshold is set to 15% of the weighted sum of the residuals before fitting. The thresholds are determined by twenty sets of calibration batches of known reference offsets and rotations.

[0083] When the translation component sensitivity reaches the translation sensitivity threshold and the residual direction consistency reaches the correlation threshold, or when the rotation component sensitivity reaches the rotation sensitivity threshold and the rotation fitting gain reaches the rotation gain threshold, the terminal will route the corresponding residual to the transformation node of this batch and re-obtain the batch baseline transformation using the verified control anchor points. The terminal first calculates the scope of the transformation update's dependence on the joint constraint graph nodes, and only unlocks the hole position nodes and identity edges that reference the version of this batch and are affected by the update. Qualified nodes and identity edges that are not selected by this residual routing and whose impact assessment result is zero continue to reference the original valid version and remain locked.

[0084] The terminal transforms the local coordinates of each receiver and the local coordinates of the slope sampling within this batch to the project coordinate system using the rotation matrix and translation vector. At the same time, it propagates the original measurement covariance and the batch baseline posterior covariance to the transformed record using first-order error propagation. Each transformed record retains the original record number, batch number, coordinates before transformation, coordinates after transformation, coordinate covariance, control anchor point combination summary, and transformation solution time, so that the slope sampling point of S3, the current location of the receiver in S4, and the cross-batch common point in S5 can all be traced back to the same batch transformation.

[0085] Before batch data was unified, the terminal used the original record number and measurement time to form a deduplication key. When multiple records appeared with the same deduplication key, the coordinate covariance trace was uniformly converted to square millimeters. The record with the smallest conversion value is retained. Then, the batch number, version identifier, and coordinate unit are checked for the receiver coordinates, slope sampling coordinates, and common point coordinates respectively. If the batch number of the record is inconsistent with the transformation version, any coordinate component exceeds the project boundary, or the covariance trace exceeds 0.0075 square meters, the record is written to the isolation queue and no coordinate transformation is performed. The remaining records are written to the project database as three output objects: receiver dataset, slope sampling dataset, and common point dataset.

[0086] When the base station's installation location changes by more than 0.10 meters, its attitude changes by more than 0.5 degrees, or the control anchor point combination changes, the terminal generates a new version identifier. The version identifier consists of the project number, batch number, and sequentially increasing transformation version number. The version record is associated with and stores the rotation matrix, translation vector, posterior covariance, effective anchor point number, standardized residuals of each anchor point, degradation check results, and status summary. The old version remains read-only. The data record after the conversion of this batch is written to the version identifier used and sent to S3 and S4. When the baseline recalculation occurs in S5, only the new version of this batch is generated, and the affected nodes are determined through the reference relationship.

[0087] In this specific embodiment, S3 includes:

[0088] The data processing terminal reads the slope sampling data unified to the project coordinate system by S2. In this specific implementation, ground laser scanning point cloud is used, and discrete elevation points are used to supplement the obscured areas. The point cloud record includes point number, three-dimensional coordinates, collection batch, timestamp, echo intensity, and coordinate covariance. The unit of three-dimensional coordinates is meters, and the unit of each element of the coordinate covariance is square meters. For quality auditing, it is converted to square millimeters. The terminal first retains the point closest to the centroid within each voxel with a voxel side length of 0.05 meters, and then deletes isolated points using three times the median distance of the neighborhood as the outlier threshold. Sampling points with inconsistent batch versions or a maximum eigenvalue of coordinate covariance greater than 0.0025 square meters are not included in the slope patch fitting and are added to the list to be retested.

[0089] The terminal performs region growth under the following conditions: the included normal angle is no greater than 8 degrees, the distance between adjacent sampling points is no greater than 0.30 meters, and the height is continuous. The height continuity is determined by the absolute value of the projection of the three-dimensional displacement of the candidate point and the current region boundary point onto the unit normal of the current region. If the absolute value is no greater than 0.08 meters, it is considered continuous. The 0.08-meter threshold is jointly determined by the sampling interval of the project point cloud and the allowable slope roughness and is written into the slope patch parameter record. If any condition is not met, the growth in that boundary direction is stopped, and a new slope patch seed is created with the unassigned candidate point. The terminal then divides the filtered sampling data into slope patches, each containing at least thirty sampling points. For areas with insufficient sampling points, the terminal merges them with adjacent slope patches that have the minimum normal difference and whose boundaries are connected. Areas that still have less than thirty points after merging are marked as low-confidence areas and do not carry target holes. The slope patch record saves the patch number, member point index, boundary polygon, batch version, and update time.

[0090] For slope area Terminal Calculate the covariance of the sampling points, where the sign is... Indicates the slope patch index in the current set of valid slope patches, symbol Indicates slope area The number of valid sampling points within the range is a dimensionless positive integer, with the sign... Indicates the index of the sampling point within the slope area and takes values ​​from 1 to... All three are only used for numbering or counting discrete objects; units or dimensions are not applicable. Indicates the first The three-dimensional coordinate vectors of each sampling point in the project coordinate system, with each component in meters, and the symbol... Indicates slope area The three-dimensional centroid of the internal sampling points and the unit of each component is meters, with the symbol... Indicates slope area The three-dimensional coordinate covariance matrix with units of square meters, superscript The vector transpose is represented; the unit eigenvector corresponding to the minimum eigenvalue of the terminal is taken as the slope patch normal and uniformly adjusted to face outwards from the slope. The minimum eigenvalue is taken as the fitting variance, and the root mean square distance from the sampling point to the fitting plane is taken as the roughness. The two are stored in square meters and meters respectively.

[0091] The terminal establishes a local orthogonal coordinate system at the centroid of each slope patch, consisting of two tangential basis vectors and the normal vector. After projecting the design adjacent directions onto the tangential plane, local measurement parameters are obtained. The local measurement record includes the two principal tangential dimensions, allowable normal offset, fitting variance, roughness, and applicable aperture category. The project measurement plan pre-issues control baseline records. Each control baseline record includes point pair identifiers, design project coordinates of the two endpoints, known baseline length, principal tangential direction, design version, and issuing source. The terminal projects the difference between the design coordinates of the two endpoints onto the principal tangential direction and takes the projection modulus as the design projection distance. The on-site data acquisition terminal re-measures at least two independent control baseline point pairs along each principal tangential direction on each slope patch, and collects at least five normal re-measurement neighborhoods on both sides of each boundary shared with adjacent slope patches. The re-measurement object number, point pair identifier, principal tangential direction, data acquisition subject, project coordinates, coordinate covariance, and batch version are recorded. The timestamp and corresponding slope patch number are used to ensure that the projection span of the retest point along any principal tangent is not less than 0.50 meters. All records must reference the current valid batch version consistent with the fitted sampling points. In this specific implementation, the initial value of the principal tangent scale is set to 1, and the allowable normal offset is set to the maximum value between 0.05 meters and three times the roughness. The terminal calculates the ratio of the measured tangent distance to the design projection distance for each valid point pair. The weight is obtained by weighted averaging after inverse normalization of the variance of the measured point pair distance. When the calibrated parameters exceed the dimensionless range of 0.5 to 2, the slope patch is marked as a geometric anomaly and re-segmented. When there are fewer than two valid independent point pairs in any principal tangent direction, the point pair identifier cannot be associated with the control baseline, the design projection distance is less than 0.10 meters, the spatial span is insufficient, the batch or design version is inconsistent, or the maximum eigenvalue of the coordinate covariance is greater than 0.0025 square meters, the slope patch is marked as to be resampled and the target hole position is prohibited.

[0092] The terminal reads the retaining wall boundaries, exposed pipelines, loose rock areas, and restricted zones from the construction drawings, unifies the boundaries of each obstacle to the project coordinate system, and performs buffer expansion within the slope tangent plane at a construction clearance of 0.40 meters, then expands it 0.20 meters along the normal direction to form a three-dimensional obstacle infeasibility region. The obstacle record includes obstacle number, boundary grid, clearance, source map version, and valid time. If a candidate point for a target borehole is located within the infeasibility region or its shortest distance to the obstacle boundary is less than the construction clearance, the candidate is not directly deleted. Instead, an obstacle clearance factor is added to the subsequent joint solver, and a local resampling flag for the borehole is output when the constraints cannot be met.

[0093] When the terminal constructs the joint constraint diagram, it sets the transformation common node for each batch, the state node for each slope patch, and the 3D hole node corresponding to each hole number in S1. The batch transformation common node is initialized with the transformation and posterior covariance of S2. The slope patch state node is initialized with centroid, normal, fitting variance, roughness, and local metric parameters. The hole node first projects the design plane coordinates onto the slope patch containing the plane position and uses the intersection as the 3D initial value. For holes falling on both sides of the slope patch boundary, the terminal simultaneously establishes two candidate fitting factors and uses the candidate with the smallest whitening residual as the effective association. For holes that do not support the slope patch, it outputs a supplementary sampling request.

[0094] The joint constraint diagram establishes anchor point transformation factor, slope fit factor, local metric factor, array orientation factor, minimum closed-loop factor, and obstacle clearance factor. The anchor point transformation factor constrains batch transformation to be consistent with S2 anchor point observation. The slope fit factor constrains borehole positions to be located in the corresponding slope patch. The local metric factor maintains the design distance between adjacent borehole positions under the local slope metric. The array orientation factor maintains the design orientation projection relationship. The minimum closed-loop factor constrains the cumulative displacement closure of the S1 closed loop. The obstacle clearance factor applies a one-way penalty to candidates that intrude into the infeasible region or have insufficient clearance. The original residual of each factor is first divided by its measurement standard deviation or calibration standard deviation to form a dimensionless whitened residual. The factor weights are jointly determined by the S2 covariance, slope fit statistics, and design acceptance accuracy and are recorded with each version.

[0095] The terminal uniformly extracts 20% of the sampling points from each slope patch according to a spatial grid as verification points. The remaining sampling points are used to fit the slope state. The root mean square distance from the verification point to the fitting plane must not exceed 0.04 meters. For boundaries with adjacent slope patches, the terminal records and pairs the normal remeasurement neighborhoods of the same batch version on both sides of the boundary, and fits the remeasurement normals respectively, requiring that the angle between the model normal and the corresponding remeasurement normal not exceed 3 degrees. For outer boundaries without adjacent slope patches, the terminal uses at least three uniquely distributed points evenly distributed along the boundary. The normal direction fitted by the remeasured neighborhood is used as the verification benchmark. If the number of remeasured neighborhoods is less than the specified number, the batch versions are inconsistent, or the number of valid points in any neighborhood is less than ten, the slope patch is marked as to be resampled and the bearing target hole position is prohibited. If the verification fails, the terminal refilters and fits the data at three voxel scales of 0.03 meters, 0.05 meters, and 0.08 meters respectively. The result with the smallest root mean square distance and the normal threshold is used as the slope patch state. If all three scales fail, the slope patch is marked as to be resampled and no bearing hole position is generated.

[0096] After S5 returns the borehole location residuals for at least five borehole locations within the same slope patch, the terminal refits the measured normal direction based on the 0.30-meter neighborhood of the corresponding borehole location's measured 3D coordinates. It calculates the angle between the model normal and the measured normal direction, and calculates the Spearman rank correlation coefficient between the absolute value of the borehole location residual normal component and the normal deviation, and the Spearman rank correlation coefficient between the absolute value of the borehole location residual normal component and the fitting variance. The minimum of the two correlation coefficients is taken as the slope correlation degree. If the sample size is less than five or the rank variance of any variable is zero, the correlation check is marked as insufficient data and a supplementary measurement is requested. The terminal applies a 0.01-meter perturbation to each component of the centroid of the slope patch state nodes. A small perturbation of 0.01 radians is applied to the normal direction, and a dimensionless perturbation of 0.01 is applied to the local measurement scale. The state increment is normalized using 0.01 meters, 0.01 radians, and 0.01 as reference scales, respectively. At the same time, the change in the residual of the borehole location is divided by 0.05 meters to form the dimensionless residual change. For each state component, the root mean square of the dimensionless residual change of the affected borehole location is calculated as the ratio to the dimensionless state increment. The maximum value of all state component results is taken as the dimensionless slope sensitivity. When the slope correlation is not less than 0.70 and the slope sensitivity is not less than 0.60, the residual is routed to the state node of the slope patch and the borehole node affected by the slope patch.

[0097] During local slope reconstruction, the terminal reacquires slope sampling data within a 0.50-meter range extending outward from the boundary of the routed slope patch. It then sequentially performs coordinate version checks, outlier deletion, slope patch division, normal and fitting variance calculation, and roughness and local metric updates. Only the fitting factor and local metric factor referencing the slope patch are reconstructed. The terminal determines the set of affected boreholes based on the joint Jacobian non-zero blocks, locally resolves the target 3D boreholes, and recalculates the corresponding residuals. Qualified nodes and identity edges that were not selected by the current residual routing and whose update impact assessment is zero are kept locked and their original versions remain unchanged.

[0098] Terminal according to Solve the joint constraint graph globally, where the symbols are... This represents the estimated mixed state vector composed of batch transformation common nodes, slope patch state nodes, and borehole location nodes. Its coordinate and translation components are in meters, its rotation component is in radians, its slope normal and local metric scales are dimensionless, and its fitting variance component is in square meters. The symbol is... The symbol represents the optimal state vector that has the same identity components and unit structure and minimizes the objective function. Represents the set of all effective factors within the joint constraint diagram, denoted by [symbol]. express The effective factor index in the text is only used for discrete sets or object numbering; it is not applicable to units or dimensions. (Symbols) Representation factor The dimensionless whitened residual vector, sign Representation factor Dimensionless information matrix, superscript Vector transpose operator This represents the state at which the objective function is minimized; the terminal employs damped Gauss-Newton iteration, constructing a diagonal scale matrix using coordinates and translation components (0.01 meters), rotation components (0.01 radians), slope normal and local metric scale components (0.01), and fitting variance components (0.0001 square meters) as class-specific reference scales. Each time, the state increment is divided by the corresponding reference scale to form a dimensionless normalized increment; the terminal uses... Calculate the first The relative change of the dimensionless objective function, where the sign... Indicates the first The dimensionless objective function changes relatively and is a dimensionless non-negative number, with the sign... This represents the index of the current iteration round and is a dimensionless positive integer. (symbol) Indicates the first The dimensionless objective function value, sign Indicates the first The round refers to the dimensionless objective function value of the previous round, with the subscript... Indicates the previous iteration round, operator To represent taking the absolute value, a function This indicates taking the maximum value of the two parameters within the parentheses; it is a dimensionless positive number. Derived from the numerical solution parameter records of the project and used as a lower limit for double-precision calculations to prevent division by zero; the Euclidean norm of the normalized increment is less than 10 to the power of -6 or Stop when the value is less than 10 to the power of negative 5, and the sign used thereafter. Following the complete explanation in this section: If the solution fails to converge after the first twenty iterations, the nodes unaffected by the anomaly are frozen and the high residual slope patches are re-segmented. Only one more twenty iterations of resolving are allowed, and the number of factors, the condition number of the normalization matrix, the objective function, and the reason for failure are recorded for each iteration. If the solution fails to converge again, there is no effective sampling after re-segmentation, there are insufficient effective factors, or the normal matrix is ​​not invertible, the terminal marks the relevant slope patches and boreholes as unsolved and to be sampled, outputs the slope patch number, borehole number, and batch identifier, does not write the corresponding records into the qualified target three-dimensional borehole set, and prohibits them from entering S4.

[0099] After the solution converges, the terminal uses... Obtain the posterior covariance of the relevant nodes, where the sign... Using the mixed state vector and its complete unit structure immediately adjacent to the aforementioned objective function, the symbol... This represents the posterior covariance matrix composed of all nodes and intersections at the optimal state. The variance of the coordinate components is in square meters, the variance of the rotation components is in radians squared, and the remaining terms use their corresponding product units. The symbol is... This represents the joint Jacobian matrix of all effective factor residuals with respect to the state vector. The derivatives of the coordinate and translation state columns have dimensions per meter, the derivatives of the rotation state columns have dimensions per radian, the derivatives of the dimensionless state columns have dimensions of the dimensionless reciprocal, and the derivatives of the fitted variance state columns have dimensions per square meter. The symbol is... This represents the information matrix of each of the aforementioned dimensionless factors. A dimensionless block diagonal matrix arranged in factorial order, with superscript... Indicates matrix transpose, superscript This indicates matrix inversion; the terminal only sends the optimal three-dimensional coordinates, hole position covariance, bearing slope plate number, batch version, and obstacle clearance of the hole position node that has converged and whose posterior covariance is invertible to form the target three-dimensional hole position set, and sends the batch transformation posterior covariance, slope plate state posterior covariance, and hole position posterior covariance to S4 and S5.

[0100] In this specific embodiment, S4 includes:

[0101] The data processing terminal uses the unique device identifiers written at the factory for each of the six receivers as stable identity keys to establish device calibration records. The record fields include the unique device identifier, antenna phase center offset, installation pole height correction, ranging zero offset, calibration date, validity period, and calibration covariance. The device calibration information is then associated with the corresponding unique device identifier. In this specific implementation, zero offset and pole height are checked at known control points before operation. Devices with calibration errors exceeding 0.02 meters are removed from identity mapping. Calibration offsets within the validity period are converted to the project coordinate system using the rotation matrix of the corresponding version of S2 and added to the current position of the receiver. The resulting corrected position, correction covariance, measurement time, and version identifier are written into the receiver status record.

[0102] The terminal forms receiver adjacency relationships based on the corrected position. In this specific implementation, the horizontal projection of the receivers at the same observation time is first triangulated by Delaunay, and then receiver edges with a three-dimensional distance not greater than 1.5 times the maximum adjacency distance of the S1 design hole position constraint map are retained. When a receiver does not have a retained edge, the terminal adds the receiver edge with the smallest three-dimensional distance to it and marks it as a weak adjacency. The receiver edge records and saves the unique identifier of the two end devices, the measured distance, the measured direction, the distance covariance, and the observation time, so that the receiver topology can be compared with the adjacent edges of the design hole position constraint map item by item.

[0103] Before the operation, the terminal reads four set-out round task sheets issued by the project scheduling terminal, each with a version number. Each task sheet uniquely lists six target holes according to the S1 hole position number, and records the round number, six hole position numbers, applicable batch transformation version, design drawing version, issuance time, and content summary. The union of the hole position numbers of the four task sheets must equal twenty-four design holes, and the intersection must be empty. In each round, six candidate target holes and their induced sub-graphs in the S1 design hole position constraint diagram are generated according to the task sheet. The hole position records use the batch reference transformation version used when generating coordinates in S3. Each receiver position and corresponding measurement record are associated with the batch reference transformation version identifier used when generating the data, and this version identifier is used as the receiver. Version identifier used; the terminal then groups the observations of six receivers within the same observation round whose first and last measurement time difference does not exceed two seconds and whose version records are traceable into a receiver set. Candidate identity edges are generated only between this receiver set and the task order hole position set for this round. The two-second condition is only used for receiver observation synchronization and not compared with static target hole positions. If the task order hole position is missing, duplicated, the version is untraceable, or the receiver observation does not meet the synchronization condition, the status of this round is set to unassigned and a request is made to reissue the task order or retest. If the version recorded by the receiver has been marked as invalid by S5, the receiver position is recalculated using the new version it references instead of directly matching it with the old version target hole position. Candidates whose versions are still inconsistent after recalculation are retained but the version cost is increased.

[0104] For the receiver With the target hole position Terminal Calculate the location cost, where the symbol Indicates the receiver index in the set of valid receivers in this round, symbol This represents the hole index in the candidate target hole location set for this round of tasks. Both are used only for discrete object numbering and are not applicable to units or dimensions; symbol Indicates receiver With hole position Dimensionless positional cost, symbol Indicates receiver The project coordinates are a 3D vector after calibration and correction, with each component in meters, and the symbol is... Indicates the target hole position The project coordinates are a three-dimensional vector, with each component in meters, and the symbol... The three-dimensional covariance matrix representing the receiver correction position, with units of square meters, is denoted by [symbol]. The three-dimensional posterior covariance matrix representing the target borehole coordinates, in square meters, is indicated by the superscript. Indicates vector transpose, superscript This represents matrix inversion; in this specific implementation, the dimensionless threshold of 11.345 corresponding to the 99% confidence level of the three-degree-of-freedom chi-square distribution is used. This confidence level, threshold value, and three-dimensional degrees of freedom are written into the identity assignment parameter record with project version and task sheet version. The smaller the position cost, the higher the spatial consistency. Candidate identity labels with a position cost greater than 11.345 are considered infeasible.

[0105] For each candidate identity edge, the terminal computes a fixed unary topological cost and a version cost independent of the identities of unknown neighbors; (This section contains symbols) Following the complete interpretation of the receiver index in the aforementioned location cost formula, the symbol... Using the complete interpretation of the candidate aperture index in the aforementioned location cost formula; for the receiver The terminal forms a receiver local topology signature by prioritizing strong adjacencies, ascending order of azimuth relative to the project's northward direction, ascending order of three-dimensional distance, and lexicographical order of the unique identifier of the other end device; for candidate aperture locations... The terminal extracts all design adjacent edges from the single-hole location induced subgraph of this round of tasks, and forms a local topological signature for the hole location according to the following order: strong adjacency priority, design orientation ascending order, design distance ascending order, and the lexicographical order of the hole location number at the other end; symbols Indicates receiver The local topological signature edge item count is a dimensionless non-negative integer count, with the sign... Indicates candidate hole position The local topological signature edge item count is a dimensionless non-negative integer; the terminal uses Determine the number of common pairing bits, where the sign The function represents the number of common paired bits in two local topological signatures, and is a dimensionless, non-negative integer count. This indicates taking the minimum value of the two parameters within the parentheses; the terminal uses... Determine the number of unpaired edge terms, where the symbol is... The operator represents the number of unpaired edge items between two local topological signatures, and is a dimensionless non-negative integer count. This indicates taking the absolute value; the terminal first determines the absolute value before calculating any mean or ratio. When using the first complete explanation of this paragraph At that time, skip all calculations of the mean of distance standardized deviation, the mean of direction standardized deviation, and the proportion of inconsistencies in adjacency existence; do not generate candidate identity edges with valid topological costs; directly output a flag indicating insufficient topological data and request supplementary testing; thus, in Division is not performed at this time; it is only when the first complete definition in this paragraph is used. Only then will the terminal pair the previous ones according to their sorting order. Each edge item, for this Calculate the arithmetic mean of the absolute value of the distance difference divided by 0.10 meters and the arithmetic mean of the absolute value of the circumferential azimuth difference divided by 3 degrees for each common position, and then... Calculate the adjacency existence inconsistency ratio, where the symbol Indicates receiver With candidate pore positions The dimensionless nonnegative adjacency existence inconsistency ratio, sign ,symbol and symbols All definitions used are the first complete explanations in this section, functions. This indicates taking the maximum of the two parameters within the parentheses; the terminal then sums the mean of the distance standardized deviation, the mean of the direction standardized deviation, and... The fixed topology cost is obtained by weighting the values ​​by 0.45, 0.35, and 0.20. When only weak adjacency edges exist in branches where the number of common pairing bits is greater than zero, a weak topology flag is output and the topology cost calculated according to the aforementioned rules is retained. The three weights are determined with the goal of minimizing the mismatch rate from sixty sets of historical observations correctly labeled before deployment, and their sum is 1. The version cost is 0 when the version identifiers are the same or the same valid version chain exists, 0.5 when it can be converted by a valid transformation chain, and an infeasible value when there is no valid transformation chain. Both the topology cost and the version cost are dimensionless and the smaller the value, the more consistent they are.

[0106] Terminal with Constructing the total cost of identity assignment, where symbol ,symbol and symbols Using the receiver index, aperture index, and dimensionless location cost, along with their complete definitions, from the aforementioned location cost formula, the symbols are as follows: This represents the dimensionless total cost of the one-to-one allocation of receivers and aperture positions in this round, denoted by [symbol]. This represents the number of receivers participating in the allocation in this round, equal to the number of candidate apertures, and is a dimensionless positive integer. (Symbol: ) Indicates receiver Is it assigned to a hole position? A dimensionless binary variable that takes the value 0 or 1, with the sign... This represents the dimensionless location cost weight, which is set to 0.50 in this specific implementation. This represents the dimensionless topological cost weight, which is set to 0.35 in this specific implementation. The symbol represents the dimensionless version cost weight, which is 0.15 in this specific implementation. Indicates receiver With hole position Dimensionless topological cost, symbol Indicates receiver With hole position The dimensionless version cost; the three total cost weights come from the same historical annotation set grid search and sum to 1, the smaller the cost, the better the identity matching;

[0107] The terminal imposes a constraint that each receiver corresponds to a hole position number and each candidate hole position number corresponds to a receiver, meaning the sum of the binary variables in each receiver row is 1 and the sum of the binary variables in each hole position column is 1. It also removes candidate edges that are positionally infeasible or versionally infeasible from the feasible candidate bipartite graph. Before solving, the terminal performs maximum matching on the bipartite graph containing only feasible candidate edges and checks if the number of matching edges is equal to six. If there is no fully feasible perfect matching covering all receivers and candidate holes, the current state is set to unassigned, and the minimum impulse is output. The system retrieves the receiver set, hole location set, and reasons for missing feasible edges and requests a retest, without generating actual hole location test records for S5 to read. Only when the feasibility check passes is the Hungarian algorithm used to find the identity mapping with the minimum total cost. After solving, each selected identity edge is checked to ensure it is not marked as infeasible in location or version. If the total cost difference between the optimal solution and the second-best solution is less than 0.10, the terminal uses the solution with more strongly adjacent consistent edges as the winning result. If they are still the same, the unique result is determined by the lexicographical order of the sequence composed of the receiver device unique identifier and the hole location number.

[0108] After solving the identity mapping, the terminal generates candidate permutations that satisfy the one-to-one constraint, with the receiver edges adjacent to each identity edge as the range. The candidate set includes the exchange of aperture numbers between two adjacent receivers and the cyclic permutation of three receivers within a closed loop. The permutation must not result in one aperture number corresponding to multiple receivers. The terminal recalculates the affected adjacent edge residuals, closed-loop residuals, and version costs for each candidate. The candidates are sorted in ascending order by the weighted sum of the residuals after the permutation. When the residuals are the same, the version cost and ascending order are used as the second sorting key. The candidate ranked first and its affected identity edges are written into the permutation evaluation record.

[0109] The terminal uses the dimensionless weighted sum of the residuals of adjacent edges before replacement and the residuals of closed loops before replacement, minus the corresponding weighted sum after replacement, as the identity replacement benefit. In this specific implementation, the identity replacement benefit is not less than 0.20 and the version cost of each identity edge after replacement is not greater than 0.5 as the update condition. When the update condition is met, the terminal only unlocks and updates the identity edges that participate in the candidate replacement. Identity edges that do not participate in the candidate replacement continue to maintain their original hole position number and version. When the update condition is not met, the effective flag of the candidate record is deleted and the current identity mapping is retained.

[0110] After the identity edge is updated, the terminal checks again that each receiver corresponds to a hole number and each hole number corresponds to a receiver, and recalculates the measured 3D coordinate index, location cost, topology cost, version cost, adjacent edge residual, and closed-loop residual of the affected hole. If any one of the one-to-one constraint, version consistency condition, or residual improvement condition is not met, the terminal rolls back to the identity edge snapshot before the update. If all are met, an incremental identity mapping version is generated and the version is passed to S5 to ensure that identity redistribution has a clear verification and rollback path.

[0111] The terminal uses the device calibration record, receiver status record, receiver adjacent edge, target hole position version record, candidate identity edge, and replacement evaluation record as a joint index composed of device unique identifier, hole position number, and measurement round. After each round of allocation, it checks that the number of records of six receivers, the number of six candidate holes, the number of identity edges, and the number of unique keys are all equal. If any of the numbers are inconsistent, the status of this round is set to unallocated and the missing object identifier is output. Only when the numbers are consistent can the hole position measurement record for S5 to read be generated.

[0112] For each set of receiver and hole position numbers obtained, the terminal establishes an identity edge in the joint constraint diagram. The identity edge stores the unique identifier of the device, hole position number, location cost, topology cost, version cost, total cost, version used by the receiver, version recorded by the hole position, and effective time. The terminal reads the calibrated and unified three-dimensional position of the receiver to the project coordinate system as the index using the identity edge as the actual measured three-dimensional coordinates of the corresponding hole position. The actual measured three-dimensional coordinates, measurement covariance, and identity edge identifier are written into the hole position measurement record and sent to S5. Other holes positions not assigned in this round and their existing identity edges remain read-only.

[0113] In this specific embodiment, S5 includes:

[0114] The data processing terminal reads the measured hole position records of S4, the target three-dimensional hole position set and joint constraint diagram of S3, first aligns the measured values ​​with the target values ​​according to the hole position number and identity edge identifier, and then... Calculate the residual at the pore site, where the symbol is... This represents the hole index in the valid set of holes that has been successfully aligned by hole number and identity edge identifier. It is used only for discrete object numbering and is not applicable to units or dimensions. Indicates hole position The three-dimensional pore site residual vector, with each component in meters, and the symbol... Indicates hole position The measured three-dimensional coordinate vector, with each component in meters, and the symbol... Indicates hole position The target three-dimensional coordinate vector and each component is in meters; the terminal simultaneously saves the residual modulus, residual direction, measurement batch, bearing slope piece and identity edge mark. The hole position mark missing any alignment item is not accepted and does not participate in the qualified output;

[0115] For S1, the hole position Pointing hole position The adjacent edges are formed by the difference between two measured 3D coordinates to form a measured adjacent vector, and by the difference between two target 3D coordinates to form a target adjacent vector, and so on. Obtain the length difference, where the symbol Using the first well location index from the residual formula immediately preceding the well location, the symbol... and symbols The metric measured three-dimensional coordinate vector and the target three-dimensional coordinate vector are used in this formula; symbols Indicates the second hole position index of the same valid adjacent edge. and All are taken from the current set of valid hole positions and This is only used for discrete numbering; units or dimensions are not applicable. (Symbol) Indicates hole position The measured three-dimensional coordinate vector, with each component in meters, and the symbol... Indicates hole position The target is a three-dimensional coordinate vector, and each component is in meters. The symbol is... Indicates adjacent edges The length difference in meters, double bar operator The Euclidean norm of a three-dimensional vector is represented; the direction difference is the angle between the measured adjacent unit vector and the target adjacent unit vector, and the unit is radians. The length difference and direction difference, together with the edge key, form the adjacent edge residual. If the length of any adjacent vector is less than 0.01 meters, the edge is marked as degenerate and a remeasurement is requested.

[0116] The terminal, following the node order of the S1 closed-loop record, performs a weighted summation of the absolute values ​​of the length differences of adjacent edges within the closed loop, divided by the length acceptance benchmark and the absolute values ​​of the direction differences, divided by the direction acceptance benchmark. The edge weight is the inverse of the joint variance of the corresponding measured edge and the target edge, normalized to a sum of 1. The resulting closed-loop residual is dimensionless, and the smaller the value, the higher the array closure consistency. For adjacent measurement batches, the terminal pairs common objects with the same name in the order of priority for common hole positions and supplementation for common control anchor points. First, the two batches of records are converted to their respective current valid versions. Then, the weighted root mean square of the coordinate difference of the common objects is calculated as the cross-batch seam residual, in meters. Each seam residual record simultaneously saves the previous batch number, the next batch number, the identifiers of the two batch transformation nodes, and the list of common objects. Fixed perturbations are applied to the transformation nodes of the previous and next batches respectively. The node with the largest reduction in seam residual value is taken as the primary node, and the other node is taken as the associated node. When the reduction values ​​are the same, the transformation node of the next batch is taken as the primary node. If there are fewer than three common objects, the seam acceptance is marked as insufficient data, and an additional common point is requested.

[0117] This specific implementation sets four levels of acceptance criteria: hole location residual modulus no greater than 0.05 meters, absolute value of adjacent edge length difference no greater than 0.06 meters and absolute value of direction difference no greater than 2 degrees, closed-loop dimensionless residual no greater than 1, and cross-batch joint residual no greater than 0.04 meters. The thresholds are derived from the construction hole location positioning accuracy, hole relationship accuracy, array closure accuracy, and cross-batch consistency requirements, and are written into the acceptance parameter record with project version. The terminal writes the pass, failure, or insufficient data status level by level. If any level is insufficient data, no qualified result is output. If any level fails, the error source routing is entered instead of directly modifying all nodes.

[0118] For hole location residuals, adjacent edge residuals, and existing cross-batch joint residuals with multiple hole locations sharing a common direction within the same batch or across batches, the terminal first divides the hole location and adjacent edge length residuals by 0.05 meters and 0.06 meters respectively, the direction residual by 2 degrees, the closed-loop residual by the dimensionless baseline 1, and the cross-batch joint residual by 0.04 meters according to the same rule as S2 to form dimensionless whitening residuals. Common residuals within the same slope area and the same batch are also included in the baseline recalculation candidate diagnosis. Cross-slope area or cross-batch distributions are only used as evidence to enhance the baseline anomaly confidence and are not a necessary prerequisite for generating the candidate. The terminal applies fixed perturbations to the three translation increments and three rotation increments of the corresponding batch transformation nodes, and divides the translation perturbation by 0.01 meters and the rotation perturbation by 0.01 radians to form dimensionless increments. The dimensionless translation component is obtained by dividing the change in whitening residual by the corresponding dimensionless increment. The sensitivity of the translation component and the sensitivity of the dimensionless rotation component are calculated. The sensitivity of the cross-batch seam residuals is calculated according to the aforementioned primary and associated nodes and incorporated into the benchmark diagnosis of the node. The terminal takes the mean of the dot product of the unit vector of each hole position residual and its weighted average direction as the residual direction consistency, and takes the difference of the weighted sum of the residuals before and after rigid body rotation fitting as the rotation fitting gain. In this specific embodiment, 0.60 is used as the sensitivity threshold of the dimensionless translation component and the sensitivity threshold of the dimensionless rotation component, respectively. When the translation sensitivity is not less than 0.60 and the direction consistency is not less than 0.80, or the rotation sensitivity is not less than 0.60 and the rotation fitting gain is not less than 15% of the weighted sum of the four levels of residuals, the corresponding residual is routed to the batch transformation node, the effective control anchor point of S2 is called to re-estimate the benchmark transformation of the batch and generate a new version. The obtained candidate and other effective candidates are entered into the unified arbitration described later.

[0119] For borehole location residuals concentrated in the same slope patch, the terminal uses the angle between the slope patch model normal and the measured normal obtained by fitting the measured 30cm neighborhood of the corresponding borehole location as the normal deviation, and the projection of the borehole location residual onto the model normal as the normal residual. When there are no fewer than five valid borehole locations in the slope patch and the normal residual, normal deviation, and fitting variance all have non-zero rank variances, the terminal calculates the Spearman rank correlation coefficients between the absolute value of the normal residual and the normal deviation, and between the absolute value of the normal residual and the fitting variance, and takes the minimum of the two correlation coefficients as the correlation coefficient of the slope patch. When there are fewer than five samples or the rank variance of any variable is zero, an insufficient slope correlation data flag is output, and no candidate for local slope reconstruction is generated based on correlation conditions. The terminal follows the sensitivity of S3. The aperture is calculated by applying a 0.01-meter perturbation to the centroid of the slope patch state node, a 0.01-radian perturbation to the normal, and a 0.01-dimensional perturbation to the local metric parameters. The change in the residual modulus of the corresponding borehole location is divided by 0.05 meters and then by the ratio of the corresponding perturbation to the aforementioned reference scale to obtain the dimensionless sensitivity of each state component. The maximum absolute value of the sensitivity of all state components is taken as the normalized sensitivity. In this specific implementation, when the correlation coefficient is not less than 0.70 and the normalized sensitivity is not less than 0.60, the residual is routed to the slope patch state node and the borehole node affected by it. The sampling data within the range of the slope patch is reacquired, and the normal, fitting variance, roughness, and local metric parameters are updated according to the S3 rule. The affected borehole locations are only locally resolved.

[0120] For cases where residuals are concentrated near adjacent receiver identity edges, regardless of whether the baseline or slope routing conditions are simultaneously met, the terminal generates pairwise candidate permutations between the corresponding identity edges of adjacent receivers, maintaining a one-to-one constraint. It then calculates the weighted sum of the adjacent edge residuals and the closed-loop residuals before and after the permutation, and applies them accordingly. Calculate the benefit of identity swapping, where the symbol This represents the dimensionless benefit of identity substitution, with a larger value indicating a more significant improvement. (Symbol: ...) The symbol represents the dimensionless weighted sum of the adjacent edge residuals and the closed-loop residuals before the permutation. This represents the dimensionless weighted sum of the adjacent edge residuals and the closed-loop residuals after the permutation, indicated by the subscript. Indicates the purpose of identity verification, subscript Indicates the state label before the replacement, subscript The three text subscripts are used only to distinguish the purpose or status and are not applicable to units or dimensions. When the identity replacement benefit is not less than 0.20 and the version cost of each identity edge after replacement is not greater than 0.5, the replacement is regarded as an identity update candidate that can participate in the unified arbitration, but it will not take effect directly in this section.

[0121] The terminal uses the posterior covariance of the batch transformation nodes in the joint constraint diagram as the baseline uncertainty, and the posterior covariance of the slope patch state nodes and hole location nodes as the slope patch uncertainty and hole location uncertainty, respectively. In this specific implementation, the standard deviation of translation is no greater than 0.03 meters, the standard deviation of rotation is no greater than 0.3 degrees, the standard deviation of slope patch normal is no greater than 1.5 degrees, and the standard deviation of hole location coordinates is no greater than 0.04 meters as the corresponding covariance thresholds. At the same time, it checks that each target hole location is outside the infeasible region of the obstacle and that the shortest distance to the obstacle boundary is no less than 0.40 meters. When the baseline uncertainty exceeds the limit, it routes to the batch transformation node to perform baseline recalculation. When the uncertainty of the slope patch or hole location exceeds the limit, it routes to the corresponding slope patch state node or hole location node to perform local reconstruction. When the obstacle conditions are not met, it adds net distance constraints to the corresponding slope patch and hole location node and resolves locally.

[0122] Before each targeted update, the terminal locks all qualified nodes and identity edges that are not selected by the route, and only unlocks the routed objects and directly affected objects determined by factor dependencies. After the update, the residuals of hole points, adjacent edges, closed loops, cross-batch seams, and node uncertainties involved in these objects are recalculated. If the update causes any acceptance quantity of a previously qualified object to exceed the threshold, the update is canceled and the dependency boundary is expanded once before resolving. After the re-verification is passed, the new version is written, the affected objects are relocked, and the original versions and acceptance status of unaffected nodes and identity edges are retained.

[0123] The terminal first generates and evaluates all candidates for benchmark recalculation, slope local reconstruction, and identity reassignment that meet their respective basic premises. Before testing any candidate, the terminal establishes a common evaluation object set based on the union of all valid candidate nodes and their directly related objects in this round. The set is locked by the valid record identifier before the candidate test, so that the same record set is used before and after the update of all candidates. For each record in the set, the terminal divides the residual modulus of the hole location by 0.05 meters, the absolute value of the difference in the length of adjacent sides by 0.06 meters, the absolute value of the difference in the circumferential direction by 2 degrees, the closed-loop residual by 1, and the cross-batch joint residual by 0.04 meters to obtain five types of non-negative dimensionless record values. Each category uses the arithmetic mean of all valid record values ​​as its unique category value, which is directly used for a single record. If a category has no valid record in the common evaluation object set, the category is marked as missing and not assigned a zero value. The original weight of the category is excluded from the current calculation, and the original weights of the remaining valid categories are normalized by dividing by the sum of their original weights. The original weights of the five categories are 0.25, 0.20, 0.15, 0.20, and 0.20 in the acceptance parameter records, respectively. The terminal multiplies the normalized valid category weights by the corresponding category values ​​and sums them to obtain the dimensionless unified residual total. No valid candidates are generated when all five categories are missing. After candidate trial calculation, any candidate in the set is locked. If an existing record cannot be recalculated, the candidate is marked as unavailable. Records newly appearing during the trial calculation but not belonging to the locked set are reserved for the next round of arbitration and not added to this round, ensuring that the objects and weights of each candidate are consistent. The benchmark and slope candidates are linearized using the current Jacobian and candidate increments. Identity candidates are precisely recalculated after replacement, with the expected reduction calculated as the total unified residual before the update minus the total unified residual after the trial or recalculation. If the identity replacement benefit is insufficient, the identity redistribution candidate is marked as unavailable, without affecting the generation and comparison of benchmark recalculation or slope local reconstruction candidates. As long as there is a candidate with an expected reduction greater than 0 among the remaining candidates, the terminal will still retain that candidate for entry. Unified arbitration; the terminal uses the one with the largest expected reduction as the sole route for this round. If the difference between two expected reductions is not greater than 0.01, they are considered to be tied and will be subject to a fixed priority of recalculating the baseline, reconstructing the slope locally, and reassigning the identity in sequence. Only one type of update is performed in each round, and other routing conditions are recalculated after verification. Automatic updates will stop when the affected object still cannot pass the acceptance after three consecutive rounds. When any level of the four-level acceptance fails and all applicable baseline, slope, identity, and obstacle clearance candidates are unavailable, or the expected reduction of all applicable candidates is not greater than 0, the terminal will not count the idle rounds but will directly generate an undiagnostic state and enter a safe rollback to avoid the object of acceptance failure being suspended.

[0124] When automatic updates stop, automatic diagnosis fails, common point data is insufficient, anchor point recalculation yields no valid solution, slope patch reconstruction lacks valid sampling, identity assignment does not have a fully feasible perfect match, or when all applicable candidates fail at Level 4 acceptance and none have a valid solution or the expected reduction is greater than 0, the terminal enters a safe rollback state. Specifically, if the replacement benefit of a single identity candidate is insufficient, a single obstacle net distance cannot be locally recalculated, or a single other candidate is unavailable, only the corresponding candidate is eliminated, and a global safe rollback is not triggered. After entering safe rollback, the terminal marks the current borehole set as unaccepted and prohibits the issuance of qualified borehole coordinates. It keeps the most recently accepted node and identity edge version read-only, outputs the borehole location residual, adjacent edge residual, closed-loop residual, or cross-batch joint residual that triggered the rollback to the field terminal, and simultaneously outputs the relevant batch, slope patch, borehole, control anchor point, or receiver identifier. Supplementary measurement data is checked through the corresponding input and a new recovery cycle is created, continuing calculation from the most recently accepted locked state, without using the results of failed cycles as construction coordinates.

[0125] When all 24 borehole locations pass the Level 4 acceptance test, and the uncertainty of all batches, slope patches, and borehole locations does not exceed the corresponding thresholds, and all target borehole locations meet the conditions of obstacle infeasibility region and construction clearance, the terminal outputs a set of qualified 3D borehole locations. The output record for each borehole includes the borehole number, target 3D coordinates, measured 3D coordinates, borehole covariance, unique equipment identifier, identity edge identifier, batch change version, slope patch number, Level 4 residual summary, and acceptance time. A file summary is also calculated for verification by the construction terminal, making the update results traceable to the updated and retained objects.

[0126] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

[0127] The aforementioned joint constraint processing links batch benchmarks, local slope geometry, borehole array topology, and receiver identity within the same solution and acceptance chain, enabling different error sources to be distinguished by their corresponding residual space patterns and node sensitivities, and allowing the correction results to re-enter point, edge, closed-loop, and cross-batch joint acceptance.

[0128] This invention introduces batch transformation nodes, slope patch state nodes, hole location nodes, and identity edges into the joint constraint diagram, and uses node uncertainty and obstacle clearance as parallel acceptance conditions in addition to residuals, so that the update range can be limited to the routed object, while avoiding local corrections from destroying the already qualified hole locations and identity relationships.

Claims

1. A method for batch positioning of multiple receiver apertures in a portable base station on a slope, characterized in that, include: S1. Based on the design hole position plane coordinates, hole position number and design adjacency relationship, obtain the design hole position constraint diagram; S2. After the base station is set up or relocated, the batch reference transformation with version identifier is obtained according to the corresponding coordinates of the control anchor points in the project coordinate system and the batch local coordinate system, and the batch data is unified to the project coordinate system. S3. Based on the slope sampling data in the project coordinate system, construct a slope representation including slope patch state and obstacle infeasible region. Combine the design hole location constraint diagram, batch datum transformation and slope representation into a joint constraint diagram, and solve the overall solution to obtain the target three-dimensional hole location set and related node uncertainties. S4. Based on the current location of the receiver, the topology of the hole position constraint diagram and the version identifier, determine the one-to-one identity mapping between each receiver and the hole position number in the target three-dimensional hole position set, form identity edges representing the identity mapping in the joint constraint diagram, and obtain the measured three-dimensional coordinates of each hole position. S5. Based on the measured 3D coordinates and joint constraint diagram, obtain the hole location residuals, adjacent edge residuals, closed-loop residuals, and cross-batch joint residuals. Recalculate the common residuals of the entire array as the baseline, route the residuals within the same slope patch as the local reconstruction of the slope, and route the residuals reduced by identity edge replacement as identity redistribution. Only update the routed nodes or identity edges, and output the qualified 3D hole location set with identity mapping when the uncertainty of each residual and node meets the acceptance conditions.

2. The method for batch positioning of multiple receiver apertures on a slope portable base station according to claim 1, characterized in that, S1 includes: Establish hole position nodes according to the planar coordinates of the designed hole positions; connect the hole position nodes with the designed adjacency relationship as adjacency edges, and record the design distance and design orientation between the corresponding hole positions in the adjacency edges; select a closed loop from the adjacency edges that can cover each hole position node, record the node sequence and design closure relationship of the closed loop, and form the designed hole position constraint diagram.

3. The method for batch positioning of multiple receiver apertures on a slope portable base station according to claim 1, characterized in that, S2 includes: selecting no fewer than three non-collinear control anchor points; estimating the batch reference transformation based on the corresponding coordinates and measurement covariance of each control anchor point; reselecting control anchor points when the number of control anchor points is insufficient or they are collinear; calculating the posterior covariance of the batch reference transformation based on the coordinate residuals of the control anchor points after the transformation; generating a new version identifier when the installation location, attitude, or combination of control anchor points of the portable base station changes; and storing the batch reference transformation, the posterior covariance, the control anchor point identifiers used, and the version identifier in association.

4. The method for batch positioning of multiple receiver apertures on a slope portable base station according to claim 1, characterized in that, S3 includes: using at least one of slope point cloud, discrete elevation points, or digital elevation model as the slope sampling data; dividing the slope sampling data into slope patches; calculating the normal, fitting variance, roughness, and local metric parameters based on the sampling points in each slope patch; forming the obstacle infeasibility region based on the obstacle boundary and construction clearance; setting batch transformation common nodes, slope patch state nodes, and borehole nodes in the joint constraint diagram; constructing anchor point transformation factors, slope fitting factors, local metric factors, array orientation factors, minimum closed-loop factors, and obstacle clearance factors; and jointly solving the borehole nodes and their posterior covariance using the weighted residuals of each factor to obtain the target three-dimensional borehole set.

5. The method for batch positioning of multiple receiver apertures on a slope portable base station according to claim 1, characterized in that, S4 includes: Using the unique device identifier of each receiver as a stable identity key, the device calibration information is associated with the corresponding unique device identifier, and the current position of the receiver is corrected according to the device calibration information; the receiver adjacency relationship is formed according to the corrected position or the ranging relationship between receivers. The version identifier of the batch reference transformation used when generating the target 3D hole position is recorded. The current position of the receiver and the corresponding measurement record are associated with the version identifier of the batch reference transformation used when generating the data, and this version identifier is used as the version identifier of the receiver. The position cost between the corrected position and each target 3D hole position, the topology cost between the receiver adjacency relationship and the design hole position constraint diagram, and the version cost between the version identifier of the receiver and the version identifier recorded in the target 3D hole position are calculated. Identity edges are set between the receiver and the hole position number in the joint constraint diagram. According to the one-to-one constraint that each receiver corresponds to one hole position number and each hole position number corresponds to one receiver, the identity mapping that minimizes the sum of the position cost, the topology cost, and the version cost is obtained.

6. The method for batch positioning of multiple receiver apertures on a slope portable base station according to claim 2, characterized in that, S5 includes: The residual of the borehole position is obtained by the difference between the measured three-dimensional coordinates of each borehole position and the coordinates of the corresponding target three-dimensional borehole position; In the project coordinate system, measured adjacent edges are formed based on the measured three-dimensional coordinates of two holes with a design adjacency relationship, and target adjacent edges are formed based on the corresponding two target three-dimensional holes. The adjacent edge residual is obtained from the length difference and direction difference between the measured adjacent edges and the target adjacent edges. According to the node order of the closed loop, the absolute values ​​of the length difference and direction difference between each measured adjacent edge and the corresponding target adjacent edge are weighted and summed to obtain the closed loop residual. The cross-batch joint residual is obtained from the coordinate difference between adjacent measurement batches at a common hole or common control anchor point. The above four levels of residuals are compared with the corresponding acceptance thresholds preset according to the requirements of hole positioning accuracy, hole relationship accuracy, array closure accuracy, and cross-batch consistency to obtain the four levels of acceptance results.

7. The method for batch positioning of multiple receiver apertures on a slope portable base station according to claim 3, characterized in that, Calculate the sensitivity of the translation component and rotation component of the hole location residual and adjacent edge residual of multiple hole locations to the nodes corresponding to the batch datum transformation in the joint constraint diagram, and calculate the consistency of residual direction and the difference between the weighted sum of residuals before and after fitting the rigid body rotation. Use the difference as the rotation fitting benefit. When the translation component sensitivity reaches a preset translation sensitivity threshold and the residual direction consistency reaches a preset correlation threshold, or when the rotation component sensitivity reaches a preset rotation sensitivity threshold and the rotation fitting gain reaches a preset rotation gain threshold, the corresponding residual is routed to the node corresponding to the batch baseline transformation, and the batch baseline transformation is re-estimated using the corresponding coordinates of the control anchor point. Retain qualified nodes and identity edges that were not selected in this residual route and whose update impact assessment is not affected by the batch baseline transformation update.

8. The method for batch positioning of multiple receiver apertures on a slope portable base station according to claim 4, characterized in that, For boreholes belonging to the same slope patch, the normal deviation between the model normal of the slope patch and the measured normal obtained by fitting the neighborhood of the measured 3D coordinates of the corresponding borehole is calculated. The correlation between the normal component of the borehole site residual and the normal deviation and the fitting variance is calculated, and the sensitivity of the borehole site residual to the corresponding slope patch state node is calculated. When the correlation reaches a preset slope correlation threshold and the sensitivity reaches a preset slope sensitivity threshold, the corresponding residual is routed to the slope patch state node and the borehole node affected by the slope patch. The slope sampling data within the range of the slope patch is reacquired, and the normal, fitting variance, roughness, and local metric parameters of the slope patch are updated. The affected target 3D boreholes are locally resolved and the corresponding residuals are recalculated. Qualified nodes and identity edges that were not selected in this residual routing and whose update impact assessment is not affected by the slope patch update are retained.

9. The method for batch positioning of multiple receiver apertures on a slope portable base station according to claim 5, characterized in that, Candidate permutations that satisfy the one-to-one constraint are generated between the identity edges corresponding to adjacent receivers. The weighted sum of the adjacent edge residuals and the closed-loop residuals before and after the permutation is calculated respectively. The difference between the weighted sums before and after the permutation is taken as the identity permutation benefit. When the identity permutation benefit reaches the preset benefit threshold and the version cost after the permutation satisfies the preset version consistency condition, the corresponding identity edge is updated, and the identity edges that did not participate in the candidate permutation are retained.

10. The method for batch positioning of multiple receiver apertures on a slope portable base station according to claim 6, characterized in that, The posterior covariance of the nodes corresponding to the batch baseline transformation in the joint constraint diagram is used as the baseline uncertainty, and the posterior covariance of the nodes corresponding to the slope patch state and hole position are used as the slope patch uncertainty and hole position uncertainty, respectively. The acceptance conditions include that the baseline uncertainty, the slope patch uncertainty, and the hole position uncertainty are all not greater than their respective preset covariance thresholds, the hole position residual, the adjacent edge residual, the closed-loop residual, and the cross-batch joint residual are all not greater than their respective acceptance thresholds, and the target three-dimensional hole position is located outside the obstacle infeasibility region and the distance from the obstacle boundary is not less than the preset obstacle net distance. When the baseline uncertainty exceeds the corresponding threshold, The target is routed to the node corresponding to the batch baseline transformation to perform baseline recalculation; when the uncertainty of the slope patch or the hole position exceeds the corresponding threshold, it is routed to the corresponding slope patch state node or hole position node to perform local slope reconstruction; when the target three-dimensional hole position does not meet the obstacle infeasibility region or the preset obstacle clearance condition, it is routed to the corresponding slope patch state node and hole position node, and local re-solution is performed according to the obstacle clearance constraint; the above directional update is treated as an abnormal rollback, and only the routed node or identity edge is updated. After each update, the affected object is unlocked and the corresponding fourth-level residual and node uncertainty are recalculated. After the verification is qualified, the affected object is relocked. The qualified three-dimensional hole set is output after all four levels of acceptance results, node uncertainty and obstacle clearance meet the acceptance conditions.

Images