Strip mine slope weak plane dynamic modeling system based on real-time working condition feedback

Abstract

The invention discloses a strip mine slope weak plane dynamic modeling system based on real-time working condition feedback, and belongs to the technical field of geographic model 3D dynamic modeling of computer graphics. Comprising a data acquisition module, a data preprocessing module, a modeling and updating module, a working condition feedback calculation module and a working condition feedback adjustment module which are connected in sequence, the method comprises the following steps: calculating an overlapping region and an adjacent overlapping region of a working condition operation influence region and a strip mine slope three-dimensional model, and intensity data of coordinate points of a three-dimensional weak surface under the influence of industrial and mining operation, and labeling and dynamically updating to generate inspection information of the overlapping region and the adjacent overlapping region; meanwhile, the three-dimensional model of the strip mine slope with more accurate dangerous area form is reconstructed and updated, the accuracy and timeliness of weak plane stability analysis and early warning are improved, and the problems that in the prior art, serious defects exist in the aspect of associating real-time mining working conditions, weak plane stability analysis lags behind, accuracy is low, and early warning accuracy and timeliness are low are solved.

Description

Technical Field

[0001] This invention relates to the field of 3D dynamic modeling technology of geographic models in computer graphics, and in particular to a dynamic modeling system for weak surfaces of open-pit mine slopes based on real-time working condition feedback. Background Technology

[0002] Early identification of slope weaknesses primarily relied on geological mapping and simple surveying methods. Technicians, armed with geological hammers and compasses, conducted on-site work, visually observing slope outcrops, measuring the attitude of rock strata, recording the distribution of joints and fissures, and judging the location and orientation of weaknesses based on their experience. This method is limited by subjective judgment, inefficient, and struggles to detect hidden weaknesses, significantly limiting its effectiveness in analyzing slopes under complex geological conditions.

[0003] With the advent of surveying technologies such as total stations and GPS, the accuracy of open-pit mine surveying has been improved. More detailed topographic data can now be obtained, and combined with geological data, simple two-dimensional slope models can be constructed to aid in the identification of weak surfaces. Compared to earlier purely manual methods, this approach can more intuitively display some weak surface information. However, because it is a two-dimensional model, it cannot comprehensively and intuitively present the true spatial form of weak surfaces, and it still has shortcomings in identifying weak surfaces with complex spatial distributions.

[0004] In the digital age, the rise of 3D laser scanning technology has brought about a significant revolution in slope weakness identification. This technology can rapidly acquire high-precision 3D point cloud data of the slope surface, thereby constructing a realistic 3D model. This model clearly reveals the slope's topography and structural features, making weakness identification more accurate and intuitive, and greatly improving the understanding of the overall slope structure. Simultaneously, geophysical detection technologies such as ground-penetrating radar and seismic tomography have begun to be applied. These technologies can penetrate rock and soil masses to detect internal structures and anomalies, effectively discovering hidden weaknesses, further expanding the scope of weakness identification and improving its accuracy.

[0005] Multi-source data fusion technology has emerged to address this need, integrating data from multiple sources such as geology, surveying, and geophysics. By preprocessing data from different sources, including standardizing data formats and coordinate systems, and then performing registration and fusion analysis, a more comprehensive and accurate slope geological model is constructed. Early multi-source data fusion primarily focused on static fusion. This involved collecting various types of data at specific time points, integrating them, and then analyzing them. For example, in the early stages of a mining phase, geological exploration data, topographic survey data, and geophysical exploration data were collected, and then these data were processed uniformly to construct a model for identifying weak points. This static fusion method integrates multi-source information, significantly improving the reliability and efficiency of weak point identification compared to identifying weak points from a single data source, providing stronger support for open-pit mine slope stability analysis.

[0006] However, this method of identifying weak surfaces by statically fusing multi-source data has gradually revealed significant drawbacks, particularly in its severe inadequacy in relating to real-time mining conditions. This results in lagging and inaccurate weak surface stability analysis, as well as low accuracy and timeliness of early warnings. Mining operations are a dynamic process. Factors such as the movement of excavators, the vibrations generated by digging force, and the vibrations caused by the powerful energy released during blasting operations all have a significant impact on the stability of weak surfaces in slopes. Traditional static fusion methods only focus on the initial geological conditions of the slope and do not take real-time mining conditions into account. When mining operations change the stress distribution and structural integrity of the slope's soil and rock mass, the weak surface model built based on static data cannot reflect these changes in a timely manner, leading to a disconnect between the weak surface stability analysis and the actual situation, and making it impossible to accurately assess the current safety status of the weak surface. Summary of the Invention

[0007] The purpose of this invention is to provide a dynamic modeling system for weak surfaces of open-pit mine slopes based on real-time working condition feedback. The system includes a data acquisition module, a data preprocessing module, a modeling and updating module, a working condition feedback calculation module, and a working condition feedback adjustment module, connected in sequence. It calculates the overlapping and adjacent overlapping areas between the affected area of ​​the working conditions and the three-dimensional model of the open-pit mine slope, as well as the intensity data of each coordinate point on the three-dimensional weak surface under the influence of mining operations. This data is then labeled and dynamically updated to generate inspection information for the overlapping and adjacent overlapping areas. Simultaneously, it reconstructs and updates the three-dimensional model of the open-pit mine slope with a more accurate morphology of the hazardous area, improving the accuracy and timeliness of weak surface stability analysis and early warning. This solves the problem of existing technologies having serious deficiencies in associating with real-time mining conditions, resulting in lagging and inaccurate weak surface stability analysis, and low accuracy and timeliness of early warnings.

[0008] The specific contents of this invention are as follows: The dynamic modeling system for weak surfaces of open-pit mine slopes based on real-time working condition feedback includes a data acquisition module, a data preprocessing module, a modeling and updating module, a working condition feedback calculation module, and a working condition feedback adjustment module. The data acquisition module is used to collect slope data, working condition data, and environmental data. The data preprocessing module is used to perform coordinate unification and time synchronization processing on the above data, and store the processed data in a distributed database. The modeling and updating module is used to create a three-dimensional model of the open-pit mine slope and update it dynamically. It marks the operation positions of excavators and blasting in the three-dimensional model of the open-pit mine slope and updates it dynamically. The working condition feedback calculation module is used to calculate the overlapping area and adjacent overlapping area between the working condition operation influence area and the three-dimensional model of the open-pit mine slope, and to mark and dynamically update the overlapping area and adjacent overlapping area. At the same time, it calculates the intensity data of each coordinate point of the three-dimensional weak surface in the three-dimensional model of the open-pit mine slope under the influence of the working and mining operations, and to mark, compare thresholds and dynamically update them. The working condition feedback adjustment module issues early warning information based on the results of the working condition feedback calculation, and at the same time generates inspection information of overlapping areas and adjacent overlapping areas, reconstructs and updates the three-dimensional model of the open-pit mine slope, so that the shape of overlapping areas and adjacent overlapping areas is more accurate.

[0009] Furthermore, the data acquisition module includes a UAV submodule, a borehole television submodule, a GNSS locator submodule, an excavator data acquisition submodule, and a blasting data acquisition submodule; The UAV submodule controls the UAV to fly periodically along a pre-set slope path, controls the lidar on the UAV to acquire three-dimensional coordinate data of the slope surface, controls the infrared thermal imager on the UAV to acquire thermal infrared data of the slope surface, and transmits the UAV parameters back. The borehole television submodule controls the borehole television to periodically acquire video data of the borehole wall on the slope and transmit the drilling parameters back. The GNSS locator submodule periodically acquires the location data of key points on the slope through the GNSS locator. These key points include the top of the slope, the toe of the slope, the stepped slope surface, the borehole opening, and the blasting center point. The excavator data acquisition submodule controls the GNSS locator installed on the excavator to obtain the excavator's position coordinates, controls the vibration sensor installed on the excavator to obtain the vibration intensity data during excavator operation, and controls the bucket weight sensor installed on the excavator to obtain the excavation volume of the excavator. The blasting data acquisition submodule controls the blasting vibration meter to acquire blasting vibration velocity data. The blasting vibration meter is installed at several point cloud coordinates on the weak surface of the slope, and blasting parameters are input through the human-machine interface unit.

[0010] Furthermore, the data preprocessing module includes a coordinate unification submodule and a time synchronization submodule; The coordinate unification submodule measures the national geodetic coordinates of a pre-selected benchmark point in the mining area, and uses this benchmark point as the origin. The X-axis is set to be parallel to the main mining direction of the mining area, the Y-axis is perpendicular to the X-axis, and the Z-axis corresponds to the altitude along the vertical direction. The independent coordinates of the mining area are established according to the national geodetic coordinates. The coordinate transformation software is used to convert the built-in coordinate systems of UAVs, borehole televisions, excavators, and GNSS locators into the independent coordinates of the mining area. The time synchronization submodule is equipped with a unified clock source based on an NTP server, configured with "year-month-day hour:minute:second:millisecond". It controls the drone, excavator, borehole television, GNSS locator and blasting vibration meter to synchronize with the time of the NTP server and periodically calibrate the time of each device to be consistent with the NTP server.

[0011] Furthermore, the modeling and updating module includes a 3D modeling submodule and a model updating submodule; The 3D modeling submodule includes a point cloud library construction unit and a point cloud library fusion modeling unit; The point cloud library construction unit retrieves the 3D coordinate data of the slope surface acquired by the aforementioned UAV and the key location data of the slope acquired by the GNSS locator. Using point cloud library construction software, it reads the data, merges discrete points and image points, and generates a preliminary slope surface point cloud library. Outliers are deleted, and the preliminary point cloud library is imported into point cloud density adjustment and optimization software. Key and non-key areas are divided, and point density is adjusted using encryption and thinning algorithms to verify that the density meets the standards. The unit also retrieves the borehole wall video data, borehole opening coordinate data, and borehole parameters from the aforementioned slope boreholes. The three-dimensional point cloud data of the weak surface within the slope is obtained through the above data calculation. The three-dimensional point cloud data of the weak surface within the slope is read using point cloud library construction software, and discrete points and image points are fused to generate a preliminary point cloud library one for the weak surface within the slope, and outliers are deleted. The thermal infrared data of the slope surface and the parameters of the UAV are retrieved from the infrared thermal imager. The point cloud data of the cracks on the slope surface is obtained through the above data calculation. The point cloud data of the cracks is read using point cloud library construction software, and discrete points and image points are fused to generate a preliminary crack point cloud library one, and outliers are deleted. The point cloud database fusion modeling unit fuses the point cloud databases of the slope surface, the weak surface, and the crack after verifying that the density meets the standard, forming a complete slope point cloud database. This database is then imported into the Surfer triangulation modeling software to construct a 3D model of the open-pit mine slope. This model includes 3D weak surface surfaces, 3D slope surface surfaces, and 3D cracks. The unit periodically retrieves the location coordinates of the excavator and imports them into the 3D model of the open-pit mine slope, generating an "excavator model" at the corresponding coordinate location. Periodically, the unit obtains the blasting center point coordinates via a GNSS locator and imports these coordinates into the 3D model of the open-pit mine slope, generating a bolded red blasting center warning dot at the corresponding coordinate location.

[0012] Furthermore, the model update submodule includes a point cloud library update unit and a model dynamic update unit; The point cloud library update unit periodically collects new three-dimensional coordinate data of the slope surface and key location data of the slope obtained by GNSS locators to generate a new slope surface point cloud library; periodically collects new borehole wall video data, borehole opening coordinate data and borehole parameters to generate a new weak surface point cloud library of the slope; and periodically collects new thermal infrared data of the slope surface and UAV parameters to generate a new crack point cloud library. The model dynamic update unit imports the new slope surface point cloud library 1, the slope internal weak surface point cloud library 1, and the crack point cloud library 1 into the Surfer triangulation modeling software, updates the old slope point cloud library 1, and forms a new slope point cloud library 1. The new slope point cloud library 1 is used to construct a three-dimensional model of the open-pit mine slope, realizing the dynamic update of the three-dimensional model of the open-pit mine slope. The unit periodically collects new excavator position coordinates, updates the old excavator position coordinates, imports the new excavator position coordinates into the three-dimensional model of the open-pit mine slope, and generates a new "excavator model" at the corresponding coordinate position. The unit periodically obtains new blasting center point coordinates through the GNSS locator, updates the old blasting center point coordinates, and the point cloud library 1 fusion modeling unit imports the new blasting center point coordinates into the three-dimensional model of the open-pit mine slope, generating a new bolded blasting center warning red dot at the corresponding coordinate position.

[0013] Furthermore, the working condition feedback calculation module includes an influence range definition submodule, an overlapping area calculation submodule, and a weakness strength calculation submodule; The influence range definition submodule predefines the influence range of excavation as "a sphere with a radius of 5m centered on the excavator's location", the influence range of blasting as "a sphere with a radius of 10m centered on the blasting center point", the influence range of surface cracks as "an area of ​​2m on both sides and below the crack, and 2m in front and behind along the crack's direction", and the influence range of internal interlayers as "an area of ​​3m above and below the interlayer, 10m in front and behind along the interlayer's direction, and 3m on both sides perpendicular to the interlayer's direction". The overlapping area calculation sub-module uses the grid sampling method to construct the working condition influence range and the weak plane influence range respectively. The working condition influence range and the weak plane influence range are divided according to the grid size of 1m×1m×1m. Taking the coordinates of the point cloud library 1 in each influence range area as the reference, calculate the three-dimensional coordinates of each grid vertex in the independent coordinate system of the mining area, obtain the three-dimensional coordinates of each grid vertex, and construct the three-dimensional coordinates of the above grid vertices in each influence range area and the point cloud library 1 of the corresponding influence range area into a new point cloud library 2; use the point cloud library construction software to traverse each point in the point cloud library 2 in the working condition influence range area, and judge whether each point belongs to the weak plane influence range area. If each point belongs to both ranges, the working condition influence range and the weak plane influence range completely overlap. If some points belong to both ranges, the working condition influence range and the weak plane influence range partially overlap. If no point belongs to both ranges, the working condition influence range and the weak plane influence range have no overlap. Finally, mark the overlapping points as flashing red dots, and at the same time generate a warning icon and the three-dimensional coordinate range of the overlapping area in the overlapping area. Clicking on the icon with the mouse can view the overlapping details; for the case where there is no overlap between the working condition influence range and the weak plane influence range, use the spatial two-point distance formula to calculate the spatial distance between any point in the working condition influence range and any point in the weak plane influence range, and take the minimum value, denoted as Y. If Y belongs to 0 < Y ≤ threshold 1, it is judged that the working condition influence range and the weak plane influence range are in near overlap. In this case, mark the point cloud library 2 of the working condition influence range and the point cloud library 2 of the weak plane influence range as solid red dots respectively, and at the same time generate a warning icon and the three-dimensional coordinate range of the solid red dot area in the solid red dot area. Clicking on the icon with the mouse can view the overlapping details; the above working condition influence range includes the influence range of excavator excavation and the influence range of blasting, and the above weak plane influence range includes the influence range of surface cracks and the influence range of internal interlayers.

[0014] Further, the weak point strength calculation sub-module includes a data acquisition and preprocessing unit, a blasting calculation unit, an excavation calculation unit, and a threshold comparison unit; The data acquisition and preprocessing unit acquires the blasting vibration velocity measured by the blasting vibration measuring instrument at several point cloud coordinates installed on the slope weak plane. According to the vibration velocity of each marked point, use the spatial interpolation algorithm to calculate the vibration velocity of the remaining coordinate points in the weak plane cloud library 2. The vibration velocity of each point is denoted as V 振 ; At the same time, collect blasting parameters, including the total charge of a single blast, denoted as Q; use the spatial two-point distance formula to calculate the straight-line distance between each coordinate point in the point cloud library 2 and the blasting center coordinate point, denoted as R; obtain the excavation volume per square meter unit area of the excavator through the bucket weight sensor, denoted as W; obtain the three-dimensional coordinates of the excavator in the independent coordinate system of the mining area through the GNSS locator on the excavator, and obtain the vertical coordinates of each coordinate point in the weak plane cloud library 2 directly below the excavator through the three-dimensional model. Calculate the vertical distance H between the excavation area and each coordinate point of the weak plane, H = Z 挖掘机 -Z弱面 By using borehole video data from a borehole television instrument, the number of different types of rock layers and the Z-axis thickness of each rock layer from the slope below the excavator to the corresponding weak surface are obtained. This is then calculated using the formula: P0 = Calculate the total load per unit area between the weak surface and the excavator above, where For the first The natural density of the ore-bearing rock layer. Let be the vertical thickness of the i-th layer of ore. Given the number of strata of ore and rock above the weak surface, calculate the residual load P above the weak surface, where P = P0 - W; The blasting calculation unit, based on the aforementioned vibration V, total charge Q for a single blast, and straight-line distance R, uses the cohesion reduction rate formula: k1 = 0.01V 振 Calculate the cohesion attenuation rate k1 at each coordinate point in the point cloud library 2 using +0.05×(Q / R), and the corrected cohesion C=C 原 ×(1-K1); Using the formula for the reduction in internal friction angle: ∆φ=0.03V 振 Calculate the reduction in internal friction angle at each coordinate point in the point cloud database using +0.1×(Q / R). The corrected actual internal friction angle φ = φ 原 -∆φ; and generate numbers at the corresponding coordinate points for the attenuated cohesion and the reduced internal friction angle; The excavation calculation unit calculates k1 based on the excavation volume W, the vertical distance H between the excavation area and each coordinate point of the weak surface, and the remaining load P above the weak surface, according to the cohesion reduction rate formula: k1=0.02+0.0005W. The corrected actual cohesion C=C 原 ×(1-K1); Using the formula for the reduction rate of internal friction angle: K2=0.001W+0.02×(H / P), K2 is calculated, and the corrected actual internal friction angle φ=φ 原 ×(1-K2); and generate numbers at the corresponding coordinate points for the attenuated cohesion and the reduced internal friction angle. The threshold comparison unit sets an internal friction angle threshold and a cohesion threshold. It then iterates through the internal friction angle values ​​of each coordinate point in the aforementioned weak slope point cloud database 2 and compares them with the preset internal friction angle threshold to find coordinate points whose internal friction angle values ​​are less than or equal to the internal friction angle threshold. Simultaneously, it iterates through the cohesion values ​​of each coordinate point in the aforementioned weak slope point cloud database 2 and compares them with the preset cohesion threshold to find coordinate points whose cohesion values ​​are less than or equal to the cohesion threshold.

[0015] Furthermore, it also includes a working condition feedback adjustment module, which includes an early warning submodule, a UAV cruise control submodule, a manual inspection submodule, a 3D model reconstruction submodule, and a reconstructed model update submodule; The early warning submodule triggers "emergency reinforcement early warning", "key early warning" and "routine early warning" for areas with intensity less than or equal to the threshold, overlapping areas and adjacent overlapping areas, respectively. The above early warning signals are sent to the system terminal and the inspection terminal. After receiving the early warning signal, relevant personnel arrange preventive measures in a timely manner within the early warning time. The UAV cruise control submodule receives the three-dimensional coordinate range of the flashing red dot region calculated by the overlapping region calculation submodule, the three-dimensional coordinate range of the solid red dot region, the coordinate point information less than or equal to the threshold sent by the threshold comparison unit, and the warning information. It generates a UAV patrol route from the coordinate information and sets the collection point cloud density information and patrol frequency information. It encapsulates the UAV patrol route, collection point cloud density information, patrol frequency information, and warning information into instructions and sends them to the UAV. The manual inspection submodule receives the three-dimensional coordinate range of the flashing red dot area calculated by the overlapping area calculation submodule, the three-dimensional coordinate range of the solid red dot area, the coordinate point information less than or equal to the threshold sent by the threshold comparison unit, and the early warning information. It generates manual inspection route and inspection frequency information, selects several marked coordinate points 1 for installing GNSS positioning instruments and sets the location data acquisition frequency, and selects several marked coordinate points 2 for installing borehole television instruments and sets the borehole wall video data acquisition frequency. It sends the manual inspection route and inspection frequency information, the GNSS positioning instrument installation coordinate points 1 and location data acquisition frequency information, and the borehole television instrument installation coordinate points 2 and borehole wall video data acquisition frequency information to the manual inspection terminal. The 3D model reconstruction submodule sets the weight of the overlapping region point cloud to 1.2, the weight of the adjacent overlapping region point cloud to 1.0, and the weight of the non-overlapping region point cloud to 0.8. Based on the incremental insertion algorithm of Delaunay triangulation, it first selects high-weight points of the overlapping region from the weak face point cloud and prioritizes inserting high-weight points into the empty set of the triangulation to be constructed. It then uses Delaunay triangulation to construct a high-weight core skeleton, ensuring that each high-weight point can be preferentially connected to the surrounding high-weight points. On the basis of the high-weight core skeleton, it first inserts the second-highest weight points of the adjacent overlapping region on the periphery. It then constructs the adjacent overlapping region skeleton of the model through Delaunay triangulation. Finally, it inserts low-weight points of the non-overlapping region and constructs the peripheral details of the model through Delaunay triangulation. The overlapping region point cloud is set as a flashing red dot, and the adjacent overlapping region point cloud is set as a solid red dot. At the same time, warning icons and 3D coordinate ranges are generated in the flashing red dot and solid red dot regions. Intensity information is marked at each coordinate point. Clicking the icon with the mouse can view the overlap details. The reconstruction model update submodule obtains the latest point cloud data of overlapping areas, adjacent overlapping areas, and non-overlapping areas from the working condition feedback calculation module. Based on the incremental insertion algorithm of Delaunay triangulation, it reconstructs the above-mentioned high-weight core skeleton, adjacent overlapping area skeleton, and peripheral details, and replaces the original 3D model.

[0016] Compared with the prior art, the present invention has at least one of the following technical effects: 1. This invention calculates the overlapping and adjacent overlapping areas between the operational impact area and the 3D model of the open-pit mine slope using a working condition feedback calculation module. It then marks and dynamically updates these overlapping and adjacent overlapping areas. Simultaneously, it calculates the strength data of each coordinate point on the 3D weak surface in the open-pit mine slope 3D model under the influence of mining operations, and marks and dynamically updates this data. The accuracy of identifying high-risk overlapping areas is improved by over 90%, eliminating missed detections caused by "static modeling without marking operational impact areas" (e.g., if the traditional model does not detect overlap between the blasting zone and the interlayer, this system can mark and warn 100% of these areas). Traditional static fusion methods only determine the weak surface strength (such as cohesion and internal friction angle) based on initial geological data, failing to respond to the dynamic impact of mining operations on strength. This application dynamically updates the weak surface strength parameters "according to the operation cycle (e.g., 10 minutes after blasting or per hour of excavation)," reducing the time difference with actual working conditions to ≤30 minutes, thus solving the core pain point of static models "lagging behind operational impacts." This solves the problem in the background technology that "when mining operations change the stress distribution and structural integrity of the slope rock and soil, the weak surface model built based on static data cannot reflect these changes in a timely manner, resulting in a disconnect between the weak surface stability analysis and the actual situation, and making it impossible to accurately assess the current safety status of the weak surface".

[0017] 2. This invention generates "encrypted inspection routes" for high-risk areas by setting up a drone cruise control submodule (e.g., overlapping area scanning frequency of 2 hours / time, point cloud density of 500 points / ㎡). The manual inspection submodule deploys GNSS positioning devices (data acquisition every 10 minutes / time) and borehole television devices (borehole wall monitoring once a day) at risk points. High-risk areas are precisely inspected according to the inspection route, achieving targeted inspection and avoiding the inefficiency of traditional "all-area indiscriminate inspection". The inspection coverage rate of high-risk areas is increased from the traditional "60%" to nearly "100%", making the inspection accurate and efficient.

[0018] 3. Traditional static modeling-based early warning systems rely solely on fixed thresholds (such as warnings for displacement exceeding 5cm) and cannot correlate with operational conditions. These systems struggle to capture sudden changes in the stability of weak surfaces caused by operational activities, leading to false or missed warnings. For instance, blasting operations may instantly increase the displacement of soil and rock near weak surfaces and alter their mechanical parameters. Static data models cannot detect this in real time. By the time displacement or other indicators exceed safety thresholds and trigger warnings, danger may already be imminent, missing the optimal window for prevention and significantly reducing the accuracy of warnings, thus failing to provide reliable assurance for safe mine production. This invention triggers "Emergency Reinforcement Warning (response within 10 minutes)," "Key Warning (response within 1 hour)," and "Regular Warning (response within 2 hours)" for areas with strength less than or equal to the threshold, overlapping areas, and adjacent overlapping areas, respectively. The warning response time is shortened from the traditional average of "2-6 hours" to "10 minutes to 2 hours." By sensing in real time the impact of working conditions on the weak surface's influence range or other strength indicators exceeding the safety threshold, the timeliness and accuracy of warnings are improved, and the false warning rate is reduced by 70% (due to the association with working condition parameters, avoiding misjudgment of non-working areas by static thresholds). This avoids missing the best prevention opportunity and solves the problem that the existing technology "statically integrates multi-source data to identify weak surfaces, which has serious deficiencies in associating with real-time mining conditions, resulting in lagging and low accuracy of weak surface stability analysis, and low warning accuracy and timeliness."

[0019] 4. Traditional static models can only be updated through "full reconstruction" after construction, which cannot adapt to changes in slope shape caused by working conditions (such as slope deformation caused by excavation, crack propagation caused by blasting); this system achieves layered dynamic updates through a modeling and updating module + a reconstruction model updating sub-module: Basic Update: The model update submodule collects new data from UAV lidar and infrared thermal imaging every 2 hours, generates a new point cloud library, and uses Surfer software to achieve "incremental update" of slope surface, weak surface, and crack models (only updates the changed areas, not the full reconstruction). Precise Reconstruction: The 3D model reconstruction submodule uses the Delaunay triangulation incremental insertion algorithm for overlapping areas (weight 1.2) and adjacent overlapping areas (weight 1.0), prioritizing the construction of high-weight core skeletons to ensure the model shape accuracy of the operation-affected area (e.g., triangle side length ≤ 0.5m). For non-overlapping areas (weight 0.8), a side length of 1 to 2m is used to balance efficiency.

[0020] The time required to update the 3D model has been reduced from the traditional "8-12 hours for full reconstruction" to "30 minutes for incremental update / 2 hours for reconstruction", and the deviation between the model shape in the overlapping area and the actual on-site measurement is ≤3cm, which is 80% more accurate than the traditional static model.

[0021] 5. Quantifying the impact of mining operations on weak surfaces to support "operating condition optimization and risk prediction." Traditional static fusion cannot quantify the specific impact of mining operations on weak surfaces (such as "the degree of damage to the strength of a weak surface caused by a certain blast"), and can only make qualitative judgments. This system achieves quantification of impact through a weak point strength calculation submodule. Quantification of blasting impact: The impact of different blasting parameters (e.g., Q=500kg vs Q=300kg) on ​​the weak surface is directly quantified by k1 (cohesion attenuation rate) and Δφ (reduction in internal friction angle), providing data support for "optimizing charge amount and adjusting blast source position" (e.g., calculation shows that "when the blast source distance increases from 7m to 10m, Δφ decreases from 7.43° to 4.2°"). Quantification of the impact of excavation: Quantifying the relationship between excavation volume W and the stability of weak surfaces using K2 (internal friction angle attenuation rate) (e.g., W from 30m...). 3 / ㎡ increased to 50m 3 / ㎡, K2 increased from 3.17% to 5.17%), guiding "regional excavation and control of excavation rate". It realizes the transformation of mining conditions from "experience-based adjustment" to "data-based optimization". After application in a certain open mine, the excessive attenuation rate of weak surface caused by blasting parameter optimization was reduced by 40%, and the slope deformation caused by excavation sequence optimization was reduced by 30%.

[0022] 6. Traditional static models can only reflect the slope condition at a certain stage and cannot adapt to changes in slope morphology and weak surface distribution caused by long-term mining (such as the exposure of new weak surfaces due to increased mining depth); this system achieves full lifecycle management through a closed-loop process of "continuous data acquisition - dynamic model update - working condition feedback adjustment". Long-term data accumulation: A distributed database stores point cloud data, strength parameters, and working condition records from previous years, forming a "slope evolution dataset" that can trace the complete change process of a weak surface from "initial state → impact of multiple blasting / excavation → current state"; Evolution trend prediction: Based on historical data, a weak surface strength decay curve is fitted (e.g., "each blast reduces cohesion by an average of 3.5%), predicting the stability of the weak surface under future mining conditions (e.g., "after 5 more blasts, the weak surface cohesion will drop below the safety threshold"); This achieves an upgrade from "short-term monitoring" to "long-term management" of weak slopes. After application in a certain open-pit mine, the incidence of slope instability accidents has significantly decreased, while extending the slope service life by 2-3 years (due to precise management avoiding overly conservative excavation schemes). Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a module architecture diagram of the dynamic modeling system for weak surfaces of open-pit mine slopes based on real-time working condition feedback, as described in this invention. Figure 2 This is a module architecture diagram of the data acquisition module of the present invention; Figure 3 This is a module architecture diagram of the data preprocessing module of the present invention; Figure 4 This is a module architecture diagram of the modeling and model updating module of this invention; Figure 5 This is a module architecture diagram of the working condition feedback calculation module of the present invention; Figure 6 This is a module architecture diagram of the working condition feedback adjustment module of the present invention. Detailed Implementation

[0025] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0026] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0027] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0028] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if [the described condition or event] is detected" may be interpreted, depending on the context, as "once determined," "in response to determination," "once [the described condition or event] is detected," or "in response to detection of [the described condition or event]."

[0029] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0030] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0031] The dynamic modeling system for weak surfaces of open-pit mine slopes based on real-time working condition feedback includes a data acquisition module, a data preprocessing module, a modeling and updating module, a working condition feedback calculation module, and a working condition feedback adjustment module. The data acquisition module is used to collect slope data, working condition data, and environmental data. The data preprocessing module is used to perform coordinate unification and time synchronization processing on the above data, and store the processed data in a distributed database. The modeling and updating module is used to create a three-dimensional model of the open-pit mine slope and update it dynamically. It marks the operation positions of excavators and blasting in the three-dimensional model of the open-pit mine slope and updates it dynamically. The working condition feedback calculation module is used to calculate the overlapping area and adjacent overlapping area between the working condition operation influence area and the three-dimensional model of the open-pit mine slope, and to mark and dynamically update the overlapping area and adjacent overlapping area. At the same time, it calculates the intensity data of each coordinate point of the three-dimensional weak surface in the three-dimensional model of the open-pit mine slope under the influence of the working condition operation, and marks and dynamically updates it. The working condition feedback adjustment module is used to generate inspection information for overlapping areas and adjacent overlapping areas, and at the same time reconstruct and update the three-dimensional model of the open-pit mine slope to make the shape of overlapping areas and adjacent overlapping areas more accurate.

[0032] The aforementioned dynamic modeling system for weak surfaces of open-pit mine slopes with real-time working condition feedback incorporates the impact of working conditions into the model construction, enabling precise inspection of areas affected by working conditions. Furthermore, through model reconstruction, it achieves a prominent and accurate display of the morphology of overlapping and adjacent overlapping areas. The model morphology of overlapping areas is more accurate than that of traditional static models, facilitating identification and early warning.

[0033] The data acquisition module includes a UAV submodule, a borehole television submodule, a GNSS locator submodule, an excavator data acquisition submodule, and a blasting data acquisition submodule; The UAV submodule controls the UAV to fly periodically along a pre-set slope path, controls the lidar on the UAV to acquire three-dimensional coordinate data of the slope surface, controls the infrared thermal imager on the UAV to acquire thermal infrared data of the slope surface, and transmits the UAV parameters back. In order to update the three-dimensional slope model in a timely manner so as to display the changes in the slope morphology in a timely manner, the UAV can be set to take pictures periodically at a frequency of, for example, 2 hours / time, or even less.

[0034] The borehole television submodule controls the borehole television to periodically acquire borehole wall video data of the slope borehole and transmit the borehole parameters back. The borehole television is set at equal intervals along the direction of the weak surface inside the slope, and is preferentially set evenly around the periphery and inside the weak surface inside the slope. This is to facilitate the subsequent use of interpolation to obtain the point cloud of the entire weak surface from the local data. The data is updated periodically, for example, every 2 hours or even less, according to the actual situation, so as to display the changes in the morphology of the internal weak surface in a timely manner and realize dynamic tracking.

[0035] The GNSS locator submodule periodically acquires the location data of key points on the slope through the GNSS locator. These key points include the top of the slope, the toe of the slope, the stepped slope surface, the borehole opening, and the blasting center point. The top of the slope, the toe of the slope, and the stepped slope surface are prone to deformation or movement. Timely updating of these coordinate data is beneficial for dynamically capturing changes in the three-dimensional model of the slope.

[0036] The excavator data acquisition submodule controls the GNSS locator installed on the excavator to obtain the excavator's position coordinates, controls the vibration sensor installed on the excavator to obtain the vibration intensity data during excavator operation, and controls the bucket weight sensor installed on the excavator to obtain the excavation volume of the excavator. The blasting data acquisition submodule controls a blasting vibration meter to acquire blasting vibration velocity data. This meter is installed at several point cloud coordinates on the weak surfaces of the slope, and blasting parameters are input through a human-machine interface unit. The blasting vibration meter is preferentially set at equal intervals along the direction of the weak surfaces within the slope, and is also preferentially and uniformly set around and within the weak surfaces. This facilitates the subsequent use of interpolation to infer the blasting vibration velocity data of the entire weak surface point cloud from the local blasting vibration data.

[0037] The data preprocessing module includes a coordinate unification submodule and a time synchronization submodule; The coordinate unification submodule measures the national geodetic coordinates of a pre-selected benchmark point in the mining area, and uses this benchmark point as the origin. The X-axis is set to be parallel to the main mining direction of the mining area, the Y-axis is perpendicular to the X-axis, and the Z-axis corresponds to the altitude along the vertical direction. The independent coordinates of the mining area are established according to the national geodetic coordinates. The coordinate transformation software is used to convert the built-in coordinate system of equipment such as drones and excavators into the independent coordinates of the mining area. Typically, a long-term fixed point is selected in the mining area as the reference point for the independent geodetic coordinate system. Coordinate transformation software includes Geodetic Coordinate Transformation V2.1 software, etc. For example, when a drone is flying, the lidar and infrared thermal imager will by default record "the drone's own GPS coordinates + the relative position of the equipment" (such as the distance and angle of a point relative to the drone measured by the lidar). The conversion requires two steps: ① First, convert the drone's GPS coordinates (usually in the WGS84 coordinate system, i.e., the global coordinates used for mobile phone navigation) into the "mine area-specific coordinates" of the drone's location using "National Geodetic Coordinate Conversion Parameters" (obtained from the local surveying and mapping department). (For example, the drone's position in the mine area-specific coordinates is X=200m, Y=150m, Z=550m). ② Then, calculate the mine area-specific coordinates of the slope surface points based on the "relative distance and angle" measured by the lidar. For example, if the lidar measures a crack point 100m in front of the drone, 5m to the right, and 30m below, the coordinates of the crack point can be calculated as X=200+100=300m, Y=150+5=155m, Z=550-30=520m. For example, the GNSS positioning module on the excavator also outputs WGS84 coordinates by default. The conversion logic is similar to that of the drone: the WGS84 coordinates collected by the module in real time are first converted into mine-specific coordinates using a preset conversion formula (such as "mine area X = national geodetic coordinate X - 3849900m, mine area Y = national geodetic coordinate Y - 11249500m", the specific values ​​are determined by the national geodetic coordinates of the origin). For example, if the module measures the national geodetic coordinates X = 3850120m and Y = 11250300m, after substituting them into the formula, we get the mine area coordinates X = 3850120 - 3849900 = 220m and Y = 11250300 - 11249500 = 800m, meaning the excavator is operating at "X = 220m, Y = 800m" in the mine area. Similarly, when installing GNSS locators on slopes, their initial coordinates in the mining area's independent coordinate system are directly measured using professional instruments (such as a total station). (For example, the coordinates of the slope top monitor are X=50m, Y=200m, Z=530m). During subsequent monitoring, only the coordinate changes need to be recorded (for example, at a certain moment, X=50.003m, Y=200.001m, indicating that the monitoring point has moved 3mm in the positive X direction), without any additional conversion. Drilling data is linked through "hole opening coordinates"—before drilling, the mining area's independent coordinates at the hole opening are measured using GNSS (for example, X=180m, Y=600m, Z=510m). Combined with the drilling depth (for example, 50m) and drilling angle (for example, vertically downward), the coordinates of the weak surface inside the hole can be calculated (for example, the coordinates of the weak surface 20m inside the hole are X=180m, Y=600m, Z=510-20=490m). Unless otherwise specified in this application, all spatial three-dimensional data are converted from coordinates to an independent coordinate system for the mining area. The conversion technology is existing in this field and will not be described in detail here.

[0038] The time synchronization submodule establishes a unified clock source based on an NTP server, configured with "year-month-day hour:minute:second:millisecond". It controls the drone, excavator, borehole television camera, GNSS locator, and installed blasting vibration meter to synchronize with the NTP server's time and periodically calibrates the time of each device to ensure consistency with the NTP server. The core significance of time synchronization and periodic calibration is to ensure that the timestamps of slope data and working condition data (such as blasting vibration and excavation volume) collected by all devices, including drones, excavators, and borehole television cameras, are completely consistent. This provides a precise time reference for subsequent multi-source data association and matching of working conditions with weak surface influences, avoiding data association failures or analysis biases caused by time differences.

[0039] For example, the time synchronization submodule controls the UAV to synchronize with the reference time via ground station software, recalibrating every 5 minutes during flight. It connects the excavator and borehole television to the NTP server via the vehicle-mounted controller or a wired network, calibrating every 1 minute and 30 seconds respectively. It also controls the on-site calibrator or network to periodically synchronize GNSS locators and other equipment with the NTP server, ensuring that the time deviation of all devices from the reference clock is ≤1 millisecond, and automatically appends a millisecond-level timestamp of the reference clock to the collected data. Unless otherwise specified in this application, all data acquisition times are synchronized with the aforementioned NTP server time.

[0040] The 3D modeling submodule includes a point cloud library construction unit and a point cloud library fusion modeling unit; The point cloud library construction unit retrieves the 3D coordinate data of the slope surface acquired by the aforementioned UAV and the key location data of the slope acquired by the GNSS locator. Using point cloud library construction software, it reads the data, merges discrete points and image points, and generates a preliminary slope surface point cloud library. Outliers are deleted, and the preliminary point cloud library is imported into point cloud density adjustment and optimization software. Key and non-key areas are divided, and point density is adjusted using encryption and thinning algorithms to verify that the density meets the standards. The unit also retrieves the borehole wall video data, borehole opening coordinate data, and borehole parameters from the aforementioned slope boreholes. The three-dimensional point cloud data of the weak surface within the slope is obtained through the above data calculation. The three-dimensional point cloud data of the weak surface within the slope is read using point cloud library construction software, and discrete points and image points are fused to generate a preliminary point cloud library one for the weak surface within the slope, and outliers are deleted. The thermal infrared data of the slope surface and the parameters of the UAV are retrieved from the infrared thermal imager. The point cloud data of the cracks on the slope surface is obtained through the above data calculation. The point cloud data of the cracks is read using point cloud library construction software, and discrete points and image points are fused to generate a preliminary crack point cloud library one, and outliers are deleted. The point cloud database fusion modeling unit fuses the point cloud databases of the slope surface, the internal weak surface, and the crack after verifying that the density meets the standards, forming a complete slope point cloud database. This database is then imported into the Surfer triangulation modeling software to construct a 3D model of the open-pit mine slope. This model includes 3D weak surface surfaces, 3D slope surface surfaces, and 3D cracks. The location coordinates of the excavator are periodically retrieved and imported into the 3D model of the open-pit mine slope, generating an "excavator model" at the corresponding coordinate location. The coordinates of the blasting center point are periodically obtained via a GNSS locator, and the point cloud database fusion modeling unit imports these coordinates into the 3D model of the open-pit mine slope, generating a bolded red blasting center warning dot at the corresponding coordinate location. The steps of the above data fusion and modeling process are as follows: 1) Data preprocessing and standardization: First, the data from the three types of point cloud databases mentioned above are converted into a unified independent coordinate system for the mining area through a preprocessing module to ensure the consistency of spatial location. All data are timestamped using an NTP server to ensure that the information was collected at the same or similar times. Data denoising (removing outliers such as floating points and outliers) and deduplication (such as removing data that was repeatedly collected from the same location by different devices) are also performed. 2) Multi-source point cloud registration: In different point cloud databases, feature points with consistent or similar spatial locations (such as the endpoints of the same crack in surface point clouds and crack point clouds) are searched. These feature points can be manually marked reference points or automatically matched by algorithms (such as ICP algorithm and SIFT feature matching). The ICP (Iterative Closest Point) algorithm or feature-based registration method is used to accurately locate the three types of point clouds in the same coordinate system so that they are aligned in space. For large-scale data, coarse registration (such as based on GNSS reference points) is used first and then fine registration is performed to reduce errors. 3) Weighting and fusion strategies: Surface point cloud: mainly used to construct the terrain surface framework, with moderate weight; Weak surface point cloud: directly related to stability analysis, with higher weight (e.g., 1.2); Crack point cloud: as a high-risk feature, with the highest weight, ensuring it is prominently displayed in the model; Fusion method: merge the three types of point clouds after registration into a complete point cloud set. According to different weights, high-weight data (weak surface, crack) is used first to construct fine structure during triangulation modeling, while low-weight data (non-critical surface areas) is used to fill and optimize the overall morphology; 4) Triangular network modeling: The fused point cloud data is imported into triangulation modeling software such as Surfer, and a 3D model is constructed using the Delaunay triangulation algorithm. This algorithm ensures that the model surface is smooth and can well reflect the geometric features of the original point cloud. High-weight regions (such as overlapping areas and crack areas) will generate denser triangular patches to ensure detail accuracy; while low-weight regions will reduce the patch density to improve efficiency. 5) Visualization and annotation: In the final model, key information such as weak surfaces, cracks, excavator location, and blasting center are marked, and attribute information such as coordinate range is added.

[0041] The aforementioned slope data includes 3D coordinate data of the slope surface acquired by UAV-controlled LiDAR and data on key slope locations, such as steps, periodically acquired by the GNSS locator submodule. Point cloud library construction software, such as CloudCompare and FaroScene, reads the aforementioned slope data (X / Y / Z coordinates in the mining area's independent coordinate system) and automatically merges the discrete coordinate points of the GNSS locator with the matching points of the UAV oblique photography imagery to generate a "Slope Surface Point Cloud Library 1" containing millions of points. For example, 5 million surface points are extracted from a 100-megapixel image of UAV oblique photography using a "feature point matching algorithm," and then merged with 100,000 key step points measured by the GNSS locator to form a point cloud library 1 covering the entire slope. The software can automatically delete "abnormal floating points" (such as false points caused by dust during UAV photography) and "duplicate points" (such as the same location being collected simultaneously by a total station and a UAV) from the point cloud, and save the point cloud library 1 in standardized formats such as LAS / PLY to ensure that subsequent modeling software can directly call them. GNSS discrete coordinate points can provide high-precision position anchoring for image matching points in UAV oblique photography, calibrate UAV data errors, and supplement the coordinates of key positions in the blind spots of the UAV's field of view, ensuring that the cloud database of slope surface points is accurate and complete.

[0042] The slope surface point cloud database includes coordinates of key locations such as the slope top, slope toe, and stepped slope surfaces (e.g., a point at the slope top has X=100m, Y=200m, Z=550m, and a point at the slope toe has X=150m, Y=250m, Z=480m). The point cloud density is adjusted according to the importance of the area. Point cloud density adjustment and optimization software, such as Trimble Real Works (adapted to mining scenarios) and Global Mapper (balancing ease of use and accuracy), supports manual or automatic division of key and non-key areas according to "slope area type." For example, "stepped slope surface and cracked area at the slope top" can be set as key areas, with a point density of ≥300 points / m² to ensure the restoration of the uneven shape of the steps and the winding shape of the cracks. "Gentle area at the slope toe and area without weak surfaces" can be set as non-key areas, with a point density of ≥100 points / m².

[0043] The detailed steps for calculating the 3D point cloud data of the weak surface within the slope using the borehole wall video data, borehole opening coordinate data, and borehole parameters are as follows: 1) Obtaining local point cloud of weak surface from single borehole coordinates 1.1) First step: Analyzing the core coordinates and parameters of the borehole (data foundation) Extract the necessary coordinates and parameters for generating weak surface point clouds from drilling records and exploration data to ensure data integrity. Drilling foundation coordinates: Orifice coordinates (X0, Y0, Z0): measured by GNSSRTK (accuracy ±3cm), such as X0=3200m, Y0=5600m, Z0=510m (mine area independent coordinate system). Drilling attitude parameters: the inclination angle α (angle between the borehole and the horizontal plane, such as 60°) and azimuth angle β (angle between the borehole axis projection and true north, such as 30°) per 1m depth are obtained by the borehole inclinometer (accuracy ±0.1°). Drilling diameter D: e.g., 150mm, used to calculate the radial distance of the hole wall (r=D / 2=75mm). Characteristic parameters of the weak surface in the borehole: Weak surface depth range: Determine the start and end depths of the weak surface within the borehole (e.g., h) through borehole wall video or borehole core logging. start =10m, h end =12m); Weak surface circumferential range: In the borehole wall video frame, identify the angular range of the weak surface in the borehole circumferential direction (e.g., θ=60°~120°, corresponding to the ring distribution of the borehole wall). 1.2) Second step: Weak surface features → Local coordinate conversion within the hole (intermediate transition) Establish a local coordinate system within the borehole (with the borehole opening as the origin, the borehole axis as the h-axis, the radial direction as the r-axis, and the circumferential direction as the θ-axis) to transform the weak surface features into local coordinates: Depth h: Sampling points are taken at 0.1m intervals within the weak surface depth range (10m~12m) (a total of 21 points, such as h=10.0m, 10.1m, ..., 12.0m) to ensure point cloud density; Circumferential angle θ: Within the circumferential range of the weak surface (60°~120°), angle values ​​are taken at 5° intervals (a total of 13 angles, such as θ=60°, 65°, ..., 120°), covering the entire circumference of the weak surface; Radial distance r: The weak surface is located on the hole wall, and the drilling radius r = 75 mm is taken (if the weak surface is concave, the measured value is supplemented by laser ranging through the hole wall, with an accuracy of ±2 mm). Local coordinate output: The local coordinates inside the hole at each sampling point are (r, θ, h), such as (75mm, 60°, 10.0m), (75mm, 65°, 10.0m), etc., forming a set of local coordinates of the weak surface; 1.3) Third step: Convert the local coordinates inside the borehole into three-dimensional coordinates (X, Y, Z) in the independent coordinate system of the mining area through "spatial vector transformation". 2) Obtain the point cloud of the weak surface from multiple borehole coordinates (core process) Multi-hole drilling requires four steps: "generating local point clouds from a single borehole → registering point clouds from multiple boreholes → global stitching → sparse completion." This process integrates discrete local point clouds into a complete weak-surface point cloud, solving the "blind spot" problem of single boreholes. 2.1) First step: Generate local point clouds of weak surfaces for each borehole (basic data preparation) For each borehole, repeat the "single borehole local point cloud generation" process to obtain multiple sets of local point clouds, for example: Borehole 1 (X0=3200m, Y0=5600m): Local point cloud A of weak surface (coverage depth 10m~12m, coordinate range 3207.5m~3209.0m, 5604.3m~5605.8m, 515.0m~516.0m). Borehole 2 (X0=3210m, Y0=5610m, spacing 14.14m): Weak surface local point cloud B (coverage depth 11m~13m, coordinate range 3217.5m~3219.0m, 5614.3m~5615.8m, 515.5m~516.5m). Key requirement: The local point cloud of each borehole must be unified to the "mine area independent coordinate system" to avoid coordinate reference deviation (e.g., if borehole 2 is in relative coordinates, it must first be converted to mine area coordinates through borehole coordinates). 2.2) Second step: Multi-hole point cloud registration (achieving local point cloud alignment) By using "identical point identification → ICP registration", the local point clouds of different boreholes are aligned to the same coordinate system. The core is to find the common features of weak surfaces across boreholes. Same-name point identification: From the local point clouds of each borehole, pairs of points with the same name that "correspond to the same weak surface feature" are selected. The identification methods include: ① Feature matching method: Compare the depth, dip angle, and orientation of the weak surfaces (e.g., the dip angle of the weak surface of borehole 1 is 55°, the dip angle of the weak surface of borehole 2 is 54.8°, and the deviation is ≤0.5°) and filter out points with similar features; ② Spatial distance method: Calculate the spatial distance between the local point cloud of borehole 1 and the local point cloud of borehole 2. Points with a distance ≤ 0.5m and consistent characteristics are identified as points with the same name. Example: Point P1 (3208.0m, 5605.0m, 515.5m) in borehole 1 and point P2 (3218.0m, 5615.0m, 516.0m) in borehole 2 have the same weak surface characteristics and their spatial distance conforms to the weak surface extension law, so they are determined to be a pair of points with the same name. ICP registration (using CloudCompare as an example): ① Import data: Set the local point cloud A of borehole 1 as the "target point cloud" and the local point cloud B of borehole 2 as the "source point cloud"; ② Initial alignment: Using the same point pair as a reference, manually move the source point cloud B to the vicinity of the target point cloud A (error ≤ 1m) to avoid ICP getting trapped in local optima; ③ Iterative optimization: Run the ICP algorithm, set the registration error threshold to ≤3cm, and the number of iterations to ≤50. The algorithm calculates the rotation matrix R and translation vector t by minimizing the sum of squared distances between corresponding points in the point cloud (e.g., R is a rotation of 1° around the Z-axis, and t is (-10.0m, -10.0m, -0.5m)). ④ Result verification: After registration, calculate the average distance of the corresponding point pairs (e.g., average distance = 2.5cm ≤ 3cm), and determine that the registration is qualified. At this time, the local point cloud B of borehole 2 and the local point cloud A of borehole 1 are accurately aligned. 2.3) Third step: Global stitching (forming a global point cloud framework for weak surfaces) For point clouds registered from three or more boreholes, a global registration error is minimized using a global optimization algorithm (such as Bundle Adjustment) to construct the overall framework of weak surfaces. Algorithm principle: Treat the local point clouds of all boreholes as a whole system and construct a global error function. This function comprehensively considers the relative registration error between the point clouds of each borehole (e.g., the error of borehole 1-2 is 2.5cm, and the error of borehole 2-3 is 2.8cm) and the matching error of corresponding points. Optimization solution: The Levenberg-Marquardt nonlinear optimization method is used to iteratively adjust the pose parameters (rotation, translation) of each borehole point cloud until the global error converges to the minimum value (e.g., global average error = 2.6cm). Overall framework output: After splicing, an overall point cloud framework of the weak surface is formed, for example, covering the range X=3207.5m~3219.0m, Y=5604.3m~5615.8m, Z=515.0m~516.5m, including the local point cloud features of each borehole, without obvious splicing faults; 2.4) Fourth step: Sparse region completion and optimization (achieving "full point cloud") For point cloud blank areas caused by large borehole spacing (e.g., >100m), the gaps are filled and the point cloud quality is improved by "Kriging interpolation completion → noise reduction and smoothing → integrity verification": Sparse region completion (Kriging interpolation): ① Construct the variability function: Analyze the spatial variability of the existing point cloud (e.g., the weak surface orientation is along the X-axis, the coefficient of variation of the point cloud in the X direction is 0.1, and in the Y direction it is 0.08), and fit a spherical variability function model; ② Interpolation calculation: For the blank area (e.g., X=3212.0m~3215.0m, Y=5609.0m~5612.0m), set the interpolation points at intervals of 0.2m×0.2m. Calculate the three-dimensional coordinates of the interpolation points (e.g., (3213.0m, 5610.0m, 516.0m)) based on the weights of the known points around the interpolation points (the closer the distance, the greater the weight). ③ Weight assignment: The weight of the virtual interpolation point is ω=0.9 (lower than 1.0 of the actual measured point), and the "virtual point" attribute is marked to distinguish between the actual measured data and the interpolated data; ④ Density guarantee: After completion, the overall point cloud density of the weak surface is ≥50 points / ㎡, and there is no obvious data loss in the blank area. Denoising and smoothing: ①Statistical filtering: Set the number of neighboring points to 15 and the standard deviation factor to 2 to remove isolated noise points (such as points that deviate from the average position of the neighboring area by more than 0.1m). ② Moving least squares method: Construct a quadratic polynomial fitting function in the local neighborhood (such as a 5m×5m range) to smooth the point cloud surface (preserving key features such as weak surface tilt angle and orientation, and eliminating small fluctuations). Integrity verification: ① Compare with the geological survey report: Compare the extended range of the weak surface point cloud after completion (e.g., X=3207.5m~3219.0m, Y=5604.3m~5615.8m) with the distribution range of the weak surface in the geological report (X=3207.0m~3219.5m, Y=5604.0m~5616.0m), and the deviation should be ≤10m; ② On-site verification: Measure the coordinates of 3 to 5 points in the weak surface outcrop area (such as the outcrop point of the interlayer on the slope surface). The deviation from the corresponding point in the point cloud should be ≤5cm to ensure the authenticity of the point cloud.

[0044] The steps for calculating point cloud data of cracks on the slope surface using thermal infrared data of the slope surface acquired by an infrared thermal imager and UAV parameters are as follows: 1) Preprocessing of thermal infrared data and UAV parameters (data foundation preparation) First, the raw data is cleaned and calibrated to eliminate noise and systematic errors, laying the foundation for subsequent crack identification and coordinate conversion. 1.1) Thermal infrared data preprocessing: Image denoising: UAV infrared thermal imaging data is susceptible to environmental interference (such as atmospheric radiation and wind noise). Median filtering (3×3 window) is used to remove isolated noise points, and Gaussian smoothing (σ=1.5) is applied to reduce image fluctuations while preserving the continuous characteristics of crack temperature anomalies. Temperature calibration: Using the infrared thermal imager's blackbody calibration data (e.g., photographing a blackbody target with a known temperature), a mapping relationship between pixel grayscale value and actual temperature is established (e.g., grayscale value 200 corresponds to a temperature of 30℃, grayscale value 180 corresponds to a temperature of 28℃), converting the thermal infrared image from a grayscale image to a temperature field distribution map. Image stitching: If a single thermal infrared image has insufficient coverage (e.g., only covering a 20m×20m area), a feature point matching algorithm (e.g., SIFT algorithm) is used to stitch multiple frames together to generate a complete panoramic thermal infrared temperature field map of the slope area (e.g., covering the overlapping area of ​​X=3195m~3210m and Y=5595m~5610m). 1.2) UAV parameter calibration: Extract key parameters from the UAV flight control system to ensure a consistent benchmark for coordinate transformation: Flight position parameters: The "three-dimensional coordinates (X, Y, F) of the fuselage" during thermal imaging are obtained through the UAV's GNSSRTK module. u Y u Z u (Accuracy ±3cm, e.g., X) u =3200m, Y u =5600m, Z u =570m, Z is the flight altitude); Flight attitude parameters: Obtain "heading angle ψ (angle between the fuselage and true north, such as 30°), pitch angle θ (angle between the fuselage and the horizontal plane, such as -5°, negative downwards), roll angle φ (angle of the fuselage roll, such as 0°)" (accuracy ±0.1°) from the UAV IMU (Inertial Measurement Unit); Camera intrinsic parameters: Obtain "focal length f (such as 8mm), principal point coordinates (u0, v0), image center pixel, such as (960, 540)", distortion coefficients (radial k1=0.01, tangential p1=0.002)" from the infrared thermal imager manual or previous calibration, for subsequent pixel coordinate → camera coordinate system conversion.

[0045] 2) Crack region identification in thermal infrared images (extraction of two-dimensional pixel range) Locating crack regions by temperature differences involves distinguishing between "crack temperature anomaly zones" and "normal rock mass areas," and extracting the pixel coordinate contours of the crack. 2.1) Temperature threshold segmentation: Analyze the thermal infrared temperature field distribution map to determine the temperature difference threshold between the crack and the rock mass (e.g., the measured crack temperature in the mining area is 2-3℃ lower than the surrounding rock mass), and use the "double threshold method" to segment the image: Low threshold Tlow The lowest temperature of the rock mass is -3℃ (e.g., if the average temperature of the rock mass is 30℃, T low =27℃); High threshold T high The lowest temperature of the rock mass is -1℃ (e.g., T). high =29℃); Segmentation results: Temperature in [T] low T high Pixels in the specified range are identified as "suspected crack areas," and are below T. low or higher than T high The pixels are classified as "background / noise" and a crack binary mask image is generated (white represents suspected cracks and black represents the background). 2.2) Crack morphology optimization: Morphological processing is performed on the binarized mask image to eliminate noise and optimize the crack profile: Expansion and corrosion: First, "3×3 structural element expansion" (fills small pores inside the crack) is used, followed by "3×3 structural element corrosion" (eliminates edge burrs) to obtain a continuous crack area; Skeleton extraction: The "Zhang-Suen skeleton extraction algorithm" is used to compress wide crack lines into a 1-pixel wide center line, accurately locating the geometric center of the crack (e.g., the pixel coordinate sequence of the crack center line: (u1, v1)=(100, 200), (u2, v2)=(200, 210), (u3, v3)=(300, 220)). Feature point extraction: Extract "endpoints (points with only 1 neighboring pixel), inflection points (points with a sudden change in direction > 30°), and equidistant points (sampling 1 point every 10 pixels)" from the skeleton line to form a list of discrete pixel coordinates of the crack (e.g., [(100, 200), (120, 205), ..., (300, 220)]). coordinate transformation The pixel coordinates (u, v) of the aforementioned crack are converted to three-dimensional coordinates (x, v) in the "camera coordinate system". c y c , z c Then, using the UAV's flight attitude parameters (heading angle, pitch angle, roll angle), the coordinates of the crack point in the camera coordinate system (x) are determined. c y c , z c The conversion of coordinates from X to Y to Z in the independent coordinate system of the mining area is a common knowledge in this field and will not be elaborated further. 4) Crack point cloud generation and density optimization (constructing 3D point cloud data) The three-dimensional coordinates of all crack feature points in the mining area are integrated to generate a preliminary crack point cloud. The density is then increased by interpolation to ensure the integrity of the point cloud. 4.1) Preliminary point cloud generation: For all the discrete pixel coordinates of the cracks extracted in the second step (e.g., 200 feature points), repeat the coordinate conversion in the third step to obtain 200 three-dimensional coordinate points of the mining area. Store them as XYZ format point cloud data (e.g., [(3190.5, 5595.2, 510.1), (3192.3, 5596.1, 510.2), ...]) according to the "crack direction order" to form a preliminary crack point cloud.

[0046] 4.2) Point cloud density completion: If the initial point cloud density is insufficient (e.g., <50 points / m, unable to reflect the fine morphology of the crack), "linear interpolation" is used to supplement the point cloud: between two adjacent feature points (e.g., P1=(3190.5, 5595.2, 510.1), P2=(3192.3, 5596.1, 510.2)), interpolation points are calculated at intervals of 0.03m (e.g., if the distance between P1 and P2 is ≈1.81m, 60 interpolation points are added); interpolation formula: X=X1+t×(X2-X1), Y=Y1+t×(Y2-Y1), Z=Z1+t×(Z2-Z1) (t is the interpolation ratio coefficient, from 0.016 to 0.96, with a step size of 0.016); after supplementation, the point cloud density is ≥100 points / m (e.g., a 10m long crack contains 1000 points), which meets the requirements for subsequent modeling.

[0047] 5) Crack point cloud verification and correction (ensuring data accuracy) Verify point cloud accuracy and correct outliers by comparing multi-source data and conducting on-site verification. 5.1) Internal verification: Comparison with LiDAR point cloud: Overlay the crack point cloud with the slope surface point cloud obtained by the UAV LiDAR in the previous stage (in CloudCompare) and check whether the crack point cloud is located on the slope surface (Z coordinate deviation ≤ 0.1m). If the Z deviation of a point is > 0.2m, it is judged as an abnormal point (possibly due to temperature segmentation error) and is removed. Geometric morphology verification: Calculate the "direction angle" (e.g., 30°) and "length" (e.g., 10m) of the crack point cloud, and compare them with the crack parameters measured in the infrared thermal image. The deviation should be ≤5% (e.g., if the measured length is 10.2m and the point cloud length is 10m, the deviation is 2%, which meets the requirements). 5.2) Field review: Select 3-5 crack outcrops (such as endpoints or inflection points) on the slope surface and use a GNSS locator to measure their coordinates in the mining area (accuracy ±3cm). Compare these coordinates with the corresponding points in the point cloud. The deviation should be ≤5cm (e.g., for the measured point (3190.5m, 5595.2m, 510.1m), the corresponding point in the point cloud (3190.52m, 5595.23m, 510.11m) has a deviation of ≈3.6cm, which meets the requirements). If the deviation exceeds the standard, recalibrate the camera intrinsic parameters or UAV attitude parameters and repeat the coordinate conversion steps.

[0048] The model update submodule includes a point cloud library update unit and a model dynamic update unit; The point cloud database update unit periodically collects new three-dimensional coordinate data of the slope surface and key location data of the slope obtained by GNSS locators to generate a new slope surface point cloud database; it periodically collects new borehole wall video data, borehole opening coordinate data and borehole parameters to generate a new weak surface point cloud database within the slope; it periodically collects new thermal infrared data of the slope surface and UAV parameters to generate a new crack point cloud database; the periodic update time can be determined according to the actual situation.

[0049] The model dynamic update unit imports the new slope surface point cloud library 1, the slope internal weak surface point cloud library 1, and the crack point cloud library 1 into the Surfer triangulation modeling software, updates the old slope point cloud library 1, and forms a new slope point cloud library 1. The new slope point cloud library 1 is used to construct a three-dimensional model of the open-pit mine slope, realizing the dynamic update of the three-dimensional model of the open-pit mine slope. The unit periodically collects new excavator position coordinates, updates the old excavator position coordinates, imports the new excavator position coordinates into the three-dimensional model of the open-pit mine slope, and generates a new "excavator model" at the corresponding coordinate position. The unit periodically obtains new blasting center point coordinates through the GNSS locator, updates the old blasting center point coordinates, and the point cloud library 1 fusion modeling unit imports the new blasting center point coordinates into the three-dimensional model of the open-pit mine slope, generating a new bolded blasting center warning red dot at the corresponding coordinate position.

[0050] The working condition feedback calculation module includes an influence range definition submodule, an overlapping area calculation submodule, and a weakness strength calculation submodule; The influence range definition submodule predefines the influence range of excavation as "a sphere with a radius of 5m centered on the excavator's location", the influence range of blasting as "a sphere with a radius of 10m centered on the blasting center point", the influence range of surface cracks as "an area of ​​2m on both sides and below the crack, and 2m in front and behind along the crack's direction", and the influence range of internal interlayers as "an area of ​​3m above and below the interlayer, 10m in front and behind along the interlayer's direction, and 3m on both sides perpendicular to the interlayer's direction". The overlapping area calculation sub-module uses the grid sampling method to construct the working condition influence range and the weak plane influence range respectively. The working condition influence range and the weak plane influence range are divided according to the grid size of 1m×1m×1m. Taking the coordinates of the point cloud library 1 in each influence range area as the reference, the three-dimensional coordinates of each grid vertex in the mine independent coordinate system are calculated to obtain the three-dimensional coordinates of each grid vertex. The three-dimensional coordinates of the above grid vertices in each influence range area and the point cloud library 1 of the corresponding influence range area are constructed into a new point cloud library 2. Use the point cloud library construction software to traverse each point in the point cloud library 2 in the working condition influence range area, and judge whether each point belongs to the weak plane influence range area. If all points belong to both ranges, the working condition influence range and the weak plane influence range completely overlap. If some points belong to both ranges, the working condition influence range and the weak plane influence range partially overlap. If no points belong to both ranges, the working condition influence range and the weak plane influence range have no overlap. Finally, mark the overlapping points as flashing red dots, and at the same time generate a warning icon and the three-dimensional coordinate range of the overlapping area in the overlapping area. Clicking on the icon with the mouse can view the overlapping details. For the case where the working condition influence range and the weak plane influence range have no overlap, use the spatial two-point distance formula to calculate the spatial distance between any point in the working condition influence range and any point in the weak plane influence range, and take the minimum value, denoted as Y. If Y belongs to 0 < Y ≤ threshold 1, it is judged that the working condition influence range and the weak plane influence range are in near overlap. In this case, mark the point cloud library 2 of the working condition influence range and the point cloud library 2 of the weak plane influence range as solid red dots respectively, and at the same time generate a warning icon and the three-dimensional coordinate range of the solid red dot area in the solid red dot area. Clicking on the icon with the mouse can view the overlapping details. The above working condition influence range includes the influence range of excavator excavation and the influence range of blasting. The above weak plane influence range includes the influence range of surface cracks and the influence range of internal interlayers.

[0051] The overlapping details are as follows (for example, "The influence range of excavation by excavator - 01 partially overlaps with the interlayer at point B, overlapping range: X = ********m~********m, Y = ********m~********m, Z = ********m~********m"). The detailed technical steps of traversing each point in the point cloud library 2 in the working condition influence range area by using the point cloud library construction software are as follows: 1). Load the target point cloud library and initialize the traversal parameters 1.1). Load data: Load the "point cloud library 2 of the working condition influence range" (denoted as Cloud W ) and the "point cloud library 2 of the weak plane influence range" (denoted as Cloud R ) through the point cloud library construction software (such as PCL, CloudCompare), and at the same time load the coordinate parameters (origin, axis direction, unit) of the mine independent coordinate system 1.2). Initialize parameters Define the traversal counter: count total (CloudW Total points), count overlap (Number of overlapping points), count non_overlap (Non-overlapping points); Define the spatial judgment accuracy threshold: ε=0.01m (due to measurement error, two points are allowed to be judged as "the same spatial location" when the coordinate difference is within 0.01m). Define the attribution tag field: for Cloud W Add is_in_R (a boolean value, True indicates that it belongs to the weak face range, False indicates that it does not belong) and overlap_type (a string, followed by the label "complete overlap / partial overlap / no overlap") to each point. 2) Traversing the point cloud library of the operating condition's influence range (Cloud) W ) By employing a "point-by-point traversal + index-based location" approach, brute-force matching of the entire point cloud is avoided, thus improving efficiency. Specifically: Index building: for Cloud R (Weak surface range point cloud) Constructing a spatial index (such as K-DTree, Octree) — by using Cloud R Points are divided into multiple grid cells according to X / Y / Z coordinates to form an index structure for fast retrieval (for example, Octree can divide the space into 2^n levels, with each leaf node corresponding to a small cube region that stores the points within that region). Point-by-point extraction: from the cloud W Extracting a single point P in sequence w (coordinates are (x) w y w , z w (), and record this point in Cloud. W index in w ; 3) In the weak area of ​​the cloud library (Cloud) R Search for "nearest neighbor" in ( ). Based on the spatial index constructed in step 2, P can be quickly located. w In the Cloud R "Potential matching points" in the cloud avoid traversing the cloud. R All points: Range search: using P w Centered on the cloud, with a radius of ε=0.01m, in the cloud... R The spatial index is used to retrieve "all points falling within the area of ​​the sphere", denoted as the candidate point set S. candidate ; If S candidate Empty: This indicates that P wNo clouds within 0.01m. R Point P is temporarily determined. w If S does not belong to the weak surface range, proceed to step 5; candidate Not empty: Proceed to step 4 to further determine P. w Whether it truly falls within the scope of the weak side; 4) Determine the attribution relationship based on both "spatial distance" and "range boundary". This is the core judgment step, which requires both "point coordinates coincide" and "the point is within the geometric boundary of the weak surface's influence range" to determine P. w Cloud R ): 4.1) Calculate the minimum spatial distance: Traverse the candidate point set S candidate Calculate P w With each candidate point P r (coordinates (x) r y r , z r The Euclidean distance d between the two sides is taken as the minimum value of d. min ; 4.2) Coordinate coincidence judgment: If d min ≤ε (i.e., d) min ≤0.01m), indicating P w With Cloud R If a point in the boundary conditions coincides with another point in the boundary conditions, the boundary condition is entered; if d min If >ε, then determine P w Not affiliated with Cloud R Proceed to step 5; 4.3) Weak surface range boundary verification: The geometric boundaries of the weak surface influence range (surface cracks / internal interlayers) have been predefined in the "Influence Range Definition Submodule" (e.g., the surface crack range is "2m on both sides and below the crack, and 2m along the front and back along the crack's direction"), and Cloud R All points have passed this boundary screening; Therefore, it is only necessary to verify P. w Do the coordinates satisfy the boundary formula for the weak surface region (taking surface cracks as an example)? Let the three-dimensional parametric equation of the crack centerline be L: (x-x0) / a=(y-y0) / b=(z-z0) / c (where a, b, and c are direction vectors, and (x0, y0, z0) are points on the line). Determine P w The vertical distance d to L ⊥ ≤2m (both sides / lower part of the crack), and P w The projected length in the direction of L is within the range of "2m before and after"; If the boundary conditions are met, then P is ultimately determined. w Scope of influence of the weak side (Cloud) R If the condition is not met, mark is_in_R as True; otherwise, mark it as False. 5) Record the results of single-point location determination. If P w is_in_R=True: P w index w Coordinates (x) w y w , z w Store in "overlapping point set S" overlap ", count overlap+ =1; If P w is_in_R=False: P w Store in "non-overlapping point set S" non_overlap ", count non_overlap+ =1; Repeat steps 2-5 until Cloud W All points have been traversed; 6) Statistical analysis and determination of the overall overlap type After the traversal is complete, based on count overlap count total Based on the proportional relationship and the "minimum spatial distance" of non-overlapping points, determine the overall overlap type between the working condition range and the weak surface range: Complete overlap determination: If count overlap / count total =100% (all P) w All belong to Cloud R If the condition is true, then it is determined that "the influence range of the working condition and the influence range of the weak surface completely overlap". Local overlap determination: If 0 <count overlap / count total <100% (partial P) w Belonging to Cloud R If the condition is met, then it is determined that there is a "local overlap"; No overlap / near overlap determination: If count overlap =0 (no P) w Belonging to Cloud R ): Calculate S non_overlap All points and Cloud R The minimum spatial distance Y of all points (by quickly retrieving the nearest weak face point of each non-overlapping point using spatial index, and taking the minimum of all distances); If 0 < Y ≤ Threshold 1 (for example, the system preset Threshold 1 is 5m), it is determined as "near overlap"; if Y > Threshold 1, it is determined as "no overlap". 7) Visualization Marking and Data Output Overlap Point Marking: Mark the points in S overlap as "flashing red dots" in the 3D model, and generate a warning icon (such as an exclamation mark), associating the coordinate range of the point and the working condition / weak surface type it belongs to (such as "excavator range - surface crack"); Near Overlap Point Marking: For the points in S non_overlap that meet the "near overlap", mark the points in the corresponding areas of Cloud W and Cloud R as "solid red dots", and also generate a warning icon and associate the coordinate range; Data Output: Store information such as the overlap type (complete / partial / near / no overlap), the number of overlap points, the 3D coordinate range of the overlap area (minimum / maximum X / Y / Z), the minimum distance Y, etc. in the distributed database, and simultaneously synchronize it to the modeling and update module to achieve dynamic annotation update of the model.

[0052] The weak point strength calculation sub-module includes a data acquisition and preprocessing unit, a blasting calculation unit, an excavation calculation unit, and a threshold comparison unit; Select several representative points in the point cloud library 1 of the weak surface, denoted as marked points. The marked points are evenly sampled on the boundary and inside of the weak surface. Install blasting vibration sensors at each marked point to measure the blasting vibration velocity. The data acquisition and preprocessing unit obtains the vibration velocity of each marked point, and uses the spatial interpolation algorithm to calculate the vibration velocity of the remaining coordinate points in the weak surface point cloud library 2. The vibration velocity of each point is denoted as V 振 ; At the same time, collect blasting parameters, including the total charge of a single blast, denoted as Q; use the spatial two-point distance formula to calculate the straight-line distance between each coordinate point in the point cloud library 2 and the blasting center coordinate point, denoted as R; obtain the excavation volume per square meter of the excavator through the bucket weight sensor, denoted as W; obtain the 3D coordinates of the excavator in the mine independent coordinate system through the GNSS locator on the excavator, and obtain the vertical coordinates of each coordinate point in the weak surface point cloud library 2 directly below the excavator through the 3D model, and calculate the vertical distance H between the excavation area and each coordinate point of the weak surface, H = Z 挖掘机 -Z 弱面 ; Obtain the number of layers of different types of ore rocks from the slope below the excavator to the corresponding weak surface and the Z-direction thickness of each layer of ore rock through the hole wall video data of the borehole TV instrument. Through the formula: P0 = Calculate the total load per unit area between the weak surface and the excavator above, where is the natural unit weight of the layer of ore rock, is the Vertical thickness of the ore-bearing rock layer. Given the number of strata of ore and rock above the weak surface, calculate the residual load P above the weak surface, where P = P0 - W; Blasting operations utilize "vibration wave impact on weak surfaces," causing the filling material within these weak surfaces to loosen and cracks to expand, thereby reducing its cohesion (the "adhesive force" resisting shear) and internal friction angle (the "friction angle" resisting sliding). The blasting calculation unit is based on the aforementioned V... 振 The total charge Q and straight-line distance R of a single blast are determined using the cohesion reduction rate formula: k1 = 0.01V. 振 Calculate the cohesion attenuation rate k1 at each coordinate point in the point cloud library 2 using +0.05×(Q / R), and the corrected cohesion C=C 原 ×(1-K1), where k1 represents the proportion of the attenuated cohesion to the original cohesion, and the blasting vibration velocity V 振 The larger the explosive charge Q, the smaller the distance from the blast source R (the stronger the blasting effect), and the larger k1 (the more severe the cohesion attenuation), 0.01 is an engineering empirical coefficient representing the weight of the blasting vibration velocity on the cohesion attenuation, i.e., when the blasting vibration velocity V at the weak surface... 振 For every 1 cm / s increase in vibration speed, the average decrease rate of cohesion attenuation k1 on the weak surface increases by 0.01 (i.e., 1%), reflecting the contribution of unit vibration intensity to cohesion attenuation. 0.05 is an empirical engineering coefficient representing the weight of the impact of blasting energy (measured by Q / R, where Q is the charge quantity and R is the distance from the blast source) on cohesion attenuation. Specifically, for every 1 increase in Q / R (unit: kg / m), the average decrease rate of cohesion attenuation k1 on the weak surface increases by 0.05 (i.e., 5%), reflecting the contribution of unit blasting energy density to cohesion attenuation. The formula for the reduction in internal friction angle is used: ∆φ = 0.03V. 振 Calculate the reduction in internal friction angle at each coordinate point in the point cloud database using +0.1×(Q / R). The corrected actual internal friction angle φ = φ 原 -∆φ; and generate numbers at the corresponding coordinate points for the attenuated cohesion and the reduced internal friction angle, where 0.03 is an engineering empirical coefficient representing the weight of the influence of blasting vibration velocity on the reduction of the internal friction angle, i.e., when the blasting vibration velocity V at the weak surface... 振 For every 1 cm / s increase, the absolute decrease in the internal friction angle of the weak surface, Δφ, increases by an average of 0.03, reflecting the contribution of unit vibration intensity to the decrease in the internal friction angle. 0.01 is an engineering empirical coefficient, representing the influence weight of blasting energy (measured by Q / R, where Q is the charge amount and R is the distance from the blast source) on the decrease in the internal friction angle. That is, when Q / R (unit: kg / m) increases by 1, the absolute decrease in the internal friction angle of the weak surface, Δφ, increases by an average of 0.1°, reflecting the contribution of unit blasting energy density to the decrease in the internal friction angle. Excavation by the excavator alters the stress state of the weak surface (e.g., reducing normal pressure) by "removing the rock load above the weak surface," indirectly reducing the internal friction angle and cohesion. The excavation calculation unit calculates k1 based on the excavation volume W, the vertical distance H between the excavation area and each coordinate point on the weak surface, and the remaining load P above the weak surface, using the cohesion reduction rate formula: k1 = 0.02 + 0.0005W. The corrected actual cohesion C = C 原 ×(1-K1), where k1 represents the cohesion attenuation rate (relative value) of the weak surface caused by excavation, representing the proportion of cohesion attenuated due to excavation disturbance to the original cohesion of the weak surface, 0.02 is the initial cohesion attenuation base (empirical coefficient) under excavation conditions, representing the cohesion attenuation rate of the weak surface caused by environmental disturbance, slight equipment vibration and other basic factors when there is no excavation (W=0, that is, the excavation operation has not affected the weak surface), the value is 0.02 (i.e. 2%), which is the minimum benchmark value of cohesion attenuation under excavation conditions, 0.0005 is the influence weight of excavation volume on cohesion attenuation (empirical coefficient), representing the unit excavation volume (for every 1m increase) 3 When the excavation volume per unit area (W) is / ㎡, the average increase in the weak surface cohesion attenuation rate K1 is 0.0005 (i.e., 0.05%), quantifying the engineering principle that "the larger the excavation volume, the more significant the cohesion attenuation." The coefficient needs to be calibrated on-site in conjunction with the lithology of the mining area (the coefficient can be increased for soft rock and decreased for hard rock). The internal friction angle reduction rate formula is used: K2=0.001W+0.02×(H / P) to calculate K2, and the corrected actual internal friction angle φ=φ 原 ×(1-K2), where K2 is the attenuation rate (relative value) of the internal friction angle of the weak surface caused by excavation, representing "the proportion of the internal friction angle attenuated by excavation disturbance to the original internal friction angle of the weak surface," dimensionless (usually presented as a percentage), directly reflecting the degree of damage to the internal friction angle of the weak surface caused by excavation (the larger K2 is, the more significant the attenuation of the internal friction angle, and the worse the anti-slip stability of the weak surface), and 0.001 is the weight of the influence of the unit area excavation volume on the attenuation of the intrinsic friction angle (engineering experience coefficient), indicating that when the unit area excavation volume W of the excavator increases by 1m... 3At a rate of / ㎡, the average increase in the attenuation rate K2 of the internal friction angle of the weak surface is 0.001 (i.e., 0.1%), quantifying the engineering principle that "the larger the excavation volume, the stronger the disturbance to the internal friction angle of the weak surface." The coefficient needs to be adjusted in conjunction with the lithology of the mining area (the coefficient can be increased for soft rock and decreased for hard rock). 0.02 is the weight of the influence of "vertical distance / remaining load" on the attenuation of the internal friction angle (engineering experience coefficient), indicating that when the ratio of "vertical distance H (m) between the excavation area and the weak surface ÷ remaining load P (kPa) above the weak surface" increases by 1 (m / kPa), the average increase in the attenuation rate K2 of the internal friction angle of the weak surface is 0.02 (i.e., 2%), reflecting the comprehensive influence of the excavation disturbance transmission depth and the stress stability of the weak surface on the attenuation (the larger H and the smaller P, the easier it is for the disturbance to be transmitted to the weak surface, and the more severe the attenuation). The excavation calculation unit generates numbers at the corresponding coordinate points for the attenuated cohesion and the reduced internal friction angle value, which will be displayed when the mouse touches the corresponding coordinate point.

[0053] The above formulas for cohesion reduction rate, internal friction angle reduction value, cohesion reduction rate, and internal friction angle reduction rate belong to the category of empirical formulas. They are derived from a large amount of field measurement and experimental data on open-pit mine slopes through data fitting and pattern summarization, including: 1) Long-term on-site monitoring data accumulation: A long-term monitoring network is deployed in multiple open-pit mines, using high-precision sensors to collect data in real time on the weak slopes of blasting and excavation areas. For example, during blasting, vibration velocity data under different blasting parameters is continuously collected for many years using vibration meters placed near the weak slopes; during excavation, weight sensors installed on the excavator bucket record the excavation volume data for each excavation operation, forming a massive raw dataset. 2) Supplementing Specialized Test Data: Conduct specialized tests to simulate the blasting and excavation conditions of open-pit mines in the laboratory; for example, prepare rock specimens containing simulated weak surfaces, conduct tests with different vibration intensities on a simulated blasting vibration table, and test the changes in cohesion and internal friction angle of the specimens after blasting; conduct disturbance tests of different degrees on specimens containing weak surfaces on simulated excavation equipment, determine the changes in the strength parameters of the weak surfaces after excavation, and supplement the lack of field data. 3) Multi-round data fitting and correction: Mathematical statistics methods are used to perform multi-round fitting on field monitoring and experimental data. For example, data on different blasting vibration velocities, explosive charges, blast source distances, and corresponding weak surface cohesion attenuation rates are input into professional statistical software. Through regression analysis and other methods, coefficients are repeatedly adjusted to fit the cohesion attenuation rate formula. For excavation conditions, data on excavation volume per unit area, vertical distance, remaining load, and internal friction angle attenuation rate are processed to obtain the internal friction angle reduction rate formula. 4) Industry Experience and Expert Evaluation: Senior experts in rock mechanics and open-pit mining were invited to evaluate the formulas. Based on their extensive experience and the specific conditions of different mining areas, the experts provided feedback on the rationality and applicability of the formulas. For example, in some soft rock mining areas, experts suggested adjusting the formula coefficients based on the susceptibility of the rock mass to disturbance, ensuring that the formulas conform to engineering realities. 5) Multi-mine verification and improvement: The initially constructed formula was applied to multiple open-pit mines with different geological conditions and mining processes for verification. The results of the weak surface strength attenuation calculated by the formula were compared with the actual monitored slope deformation and instability. If there were any discrepancies, the data collection and fitting stage was returned to further improve the formula. After years of repeated verification and optimization in multiple mines, the current formula form was formed.

[0054] The threshold comparison unit sets an internal friction angle threshold and a cohesion threshold. It then iterates through the internal friction angle values ​​of each coordinate point in the aforementioned weak slope point cloud database 2 and compares them with the preset internal friction angle threshold to find coordinate points whose internal friction angle values ​​are less than or equal to the internal friction angle threshold. Simultaneously, it iterates through the cohesion values ​​of each coordinate point in the aforementioned weak slope point cloud database 2 and compares them with the preset cohesion threshold to find coordinate points whose cohesion values ​​are less than or equal to the cohesion threshold.

[0055] For the first Natural unit weight of ore layer (unit: kN / m³) 3 For example, limestone γ=26kN / m 3 Mudstone γ = 22kN / m 3 (This is obtained by calling the "Open-pit Mine Rock Mechanics Parameter Library").

[0056] It also includes a working condition feedback adjustment module, which includes an early warning submodule, an UAV cruise control submodule, a manual inspection submodule, a 3D model reconstruction submodule, and a reconstructed model update submodule; The early warning submodule triggers alerts for overlapping areas (flashing red dots), adjacent overlapping areas (solid red dots), and areas with intensity below the threshold, respectively, such as "Emergency Reinforcement Warning (response within 10 minutes)," "Key Warning (response within 1 hour)," and "Regular Warning (response within 2 hours)." It also sends a "Reinforcement Warning" for areas with intensity less than or equal to the threshold. These warning signals are sent to the system and inspection terminals. Upon receiving the warning signals, relevant personnel, in conjunction with the inspection paths and frequencies generated below, promptly arrange inspections, work scheduling, risk assessments, and reinforcement measures within the warning time. The warning response time is shortened from the traditional average of "2-6 hours" to "10 minutes to 2 hours." By real-time sensing of the impact of work conditions on displacement or other intensity indicators exceeding safety thresholds, the timeliness and accuracy of warnings are improved, and the false alarm rate is reduced (due to the association with work condition parameters, avoiding misjudgments of non-work areas by static thresholds), thus preventing missed opportunities for optimal prevention.

[0057] The UAV cruise control submodule receives the three-dimensional coordinate range of the flashing red dot area calculated by the overlapping area calculation submodule, the three-dimensional coordinate range of the solid red dot area, the coordinate point information less than or equal to the threshold sent by the threshold comparison unit, and the warning time, etc. It generates the UAV patrol route from the above coordinate information and sets the collection point cloud density information and patrol frequency information. After encapsulating the above UAV patrol route, collection point cloud density information, patrol frequency information, warning time, etc. into instructions, it sends them to the UAV. The UAV starts the inspection in a timely manner within the warning time.

[0058] All key changes affecting the stability of weak surfaces (such as surface crack propagation, slope bulging, and rock loosening) may be directly reflected on the surface. The risk in three-dimensional overlapping areas or adjacent overlapping areas is essentially the "impact of working condition disturbances on weak surfaces," and the "external manifestation" of this impact is concentrated on the slope surface: for example, in the area where blasting overlaps with weak surfaces, if the weak surface shows a tendency to become unstable due to vibration, it will first cause the surface rock to loosen and new cracks to appear; in the area where excavation overlaps with interlayers, if the cohesion of the interlayers decreases, it will cause the upper slope to bulge or a small landslide. Surface cruise can accurately capture these millimeter-level surface changes (such as crack width increasing from 2mm to 5mm, and slope displacement increasing from 3mm to 8mm) by "increasing the scanning frequency (2 hours / time)" and "high point cloud density (500 points / ㎡)". These changes are the "direct signals" for judging the stability of the internal weak surfaces. Surface cruise can infer the stability state of the internal weak surfaces by observing surface changes (such as surface crack propagation likely corresponding to the decrease in mechanical parameters of the underlying weak surface).

[0059] The three-dimensional overlapping area (direct risk area) and the adjacent overlapping area (potential risk area, such as the area where the working condition influence range and the weak surface distance 0 < Y ≤ threshold one) together constitute the "risk gradient zone". Although the adjacent overlapping area is not directly disturbed by the working condition, it may gradually transform into the overlapping area due to the "secondary effects" of the working condition (such as the attenuation and transmission of blasting vibration, stress redistribution caused by excavation). If only cruising the three-dimensional overlapping area, it is easy to ignore the "gradual risk" in the adjacent area (such as slight cracks appearing on the slope surface of the adjacent area, which may expand to the weak surface in subsequent operations, forming a new overlapping risk). Through "gradient trajectory planning" in three-dimensional surface cruising (3m spacing flight lines in the overlapping area and 5m spacing flight lines in the adjacent area, both conforming to the three-dimensional terrain), the three-dimensional surfaces of both types of areas can be covered simultaneously, not only monitoring the immediate changes in the direct risk area but also capturing the early signals in the potential risk area, upgrading the risk monitoring from "single-point coverage" to "gradient coverage", eliminating the gap in traditional inspections where the "direct risk area is closely monitored and the potential risk area is ignored". The overall inspection coverage rate of both types of areas reaches over 98%.

[0060] The core value of the adjacent overlapping area is the "intermediate carrier of risk transmission". The impact of the working condition on the weak surface does not act on the overlapping area instantaneously but attenuates outward from the center of the working condition, first affecting the adjacent overlapping area and then gradually transmitting to the weak surface (for example, blasting vibration first loosens the slope surface of the adjacent area, and subsequent vibration superposition may cause the loosening range to expand to the weak surface, forming an overlap). Two-dimensional cruising is difficult to capture this "transmission risk" because it cannot accurately quantify three-dimensional changes. However, through obtaining the three-dimensional point cloud data of both types of areas in three-dimensional surface cruising, the "risk transmission trajectory" can be compared and analyzed. For example, 1 hour after blasting, a 2mm uplift (perpendicular to the slope surface) appears at a certain point in the adjacent overlapping area in the Z-axis direction. After 2 hours, the uplift expands to 3.5mm and extends 1m in the direction of the three-dimensional overlapping area. Combining the vibration attenuation law (vibration velocity in the adjacent area is 15cm / s and in the overlapping area is 25cm / s), it can be predicted that the uplift may transmit to the overlapping area within 4 hours, causing further attenuation of the weak surface. This "risk transmission prediction" based on three-dimensional changes enables the early warning to be advanced from "responding after obvious risks appear in the overlapping area" to "intervening when early signals appear in the adjacent area", extending the early warning time from 2 hours to more than 6 hours, leaving sufficient time for the mine to carry out reinforcement and adjustment operations.

[0061] The spatial boundary between the three-dimensional overlapping area and the adjacent overlapping area is essentially a "balance line between the impact of the working condition and the stability of the weak surface." By obtaining three-dimensional point clouds of the two types of areas through three-dimensional surface cruise, the "actual boundary of the working condition impact" can be accurately quantified (e.g., the original preset blasting impact range is a 10m sphere, but the three-dimensional cruise found that there is no obvious surface change at 5m in the adjacent overlapping area, and the actual impact boundary is only 8m). This allows for the optimization of working condition parameters (e.g., reducing the charge amount based on the actual boundary to avoid excessive disturbance). At the same time, by comparing the differences in surface changes between the two types of areas (e.g., the slope of the overlapping area bulges by 5cm, while that of the adjacent area bulges by 1.2cm), the "attenuation coefficient of the working condition impact" can be clarified (e.g., the surface change decreases by 0.8cm for every 1m away from the working condition center), providing data support for subsequent weak surface management (e.g., setting differentiated intensity thresholds for weak surfaces at different distances: the threshold for the weak surface corresponding to the adjacent overlapping area can be appropriately relaxed, while the threshold for the overlapping area needs to be strictly tightened). This "boundary optimization" based on three-dimensional quantitative data avoids the blindness of traditional experience-based setting of operating parameters, making operating condition adjustments and weak surface control more scientific. For example, after application in a certain mine, the excessive attenuation rate of weak surfaces caused by operating parameter optimization was reduced by 35%.

[0062] Mines need to allocate inspection resources according to risk levels (e.g., overlapping areas require a density of 500 points / m² and a frequency of 2 hours / time, while adjacent areas can have a density of 300 points / m² and a frequency of 4 hours / time). However, risk levels are not fixed (e.g., adjacent areas may be upgraded to high risk due to multiple overlapping operating conditions). 3D surface patrols periodically collect 3D data from two types of areas and can update risk levels based on 3D changes: for example, if the 3D displacement of a point in an adjacent area increases from 2mm to 5mm (exceeding the preset adjacent area threshold of 3mm), the system can automatically upgrade that area to a "quasi-overlapping area" and adjust patrol parameters (increasing point cloud density to 400 points / m² and reducing frequency to 3 hours / time); if the displacement of a point in an overlapping area decreases from 5mm to 1mm (stabilizing after reinforcement), it can be downgraded to a "quasi-adjacent area," appropriately reducing the inspection intensity. This "three-dimensional data-driven dynamic adjustment of risk levels" ensures that inspection resources are neither wasted in low-risk areas nor insufficient in potential risk areas that are being upgraded, increasing resource utilization by 40% while avoiding the problem of "insufficient resources in high-risk areas and excessive resources in low-risk areas" caused by fixed risk levels.

[0063] There is a causal relationship between the changes in the weak surface strength (such as cohesion decay) in the three-dimensional overlapping region and the surface changes (such as slope displacement) in the adjacent overlapping region. The three-dimensional changes in the adjacent region can serve as an indirect indicator of the changes in the weak surface strength (e.g., for every 1 mm increase in slope displacement in the adjacent region, the weak surface cohesion decreases by 0.5 kPa). The three-dimensional data of the two types of regions obtained by three-dimensional surface cruise can construct a linkage model of "surface change-strength decay": for example, the weak surface strength data of the overlapping region (provided by the working condition feedback calculation module, such as cohesion of 12 kPa) is correlated with the three-dimensional displacement data of the adjacent region (e.g., Z-axis bulge of 3 mm), and a regression equation is established. Subsequently, the weak surface strength can be quickly inferred from the three-dimensional changes in the adjacent region (without the need for frequent drilling tests). This linkage model not only makes up for the shortcomings of weak surface strength monitoring that "requires drilling and is costly", but also makes the stability analysis change from "relying only on the strength data of the overlapping region" to a dual verification of "strength data + adjacent region surface data" through the complementarity of data from the two types of regions. The reliability of the analysis results is increased from 80% to over 95%, avoiding misjudgments caused by a single data source.

[0064] In open-pit mines, 3D overlapping areas are often located in complex terrain characterized by steep slopes and numerous obstacles (e.g., slopes exceeding 50m in height, slope angles of 60°, and large deposits of ore). 3D surface cruise, based on a pre-constructed 3D slope model, can pre-plan obstacle avoidance trajectories. For example, in areas with ore deposits within the overlapping area, the flight altitude automatically adjusts from 5m to 8m to avoid obstacles. In steep slope areas, a spiral descent trajectory is used to conform to the slope, ensuring scanning accuracy while mitigating collision risks. Furthermore, the 3D surface cruise's dynamic altitude adjustment function adapts to different slope gradients in overlapping areas (e.g., a flight altitude of 5m in gentle slope areas and 3m in steep slope areas). This ensures adequate point cloud density (the point cloud density at 3m altitude in steep slope areas can reach 600 points / m², meeting the requirements for fine monitoring) while avoiding overly sparse or dense scanning in some areas due to a fixed altitude. This allows the cruise operation to balance safety and data validity in complex terrain, solving the problem of balancing safety and accuracy in complex terrain with 2D cruise.

[0065] Compared to traditional indiscriminate inspections that cover the entire area (using high-frequency scanning even in non-risk areas), 3D patrol concentrates limited UAV endurance and computing resources on areas where surface changes best reflect risk. This avoids over-inspection of flat areas without weak surfaces or operational conditions (such as areas without weak surfaces at the foot of slopes), ensuring resource consumption matches risk levels and significantly improving inspection efficiency. 3D patrol sets the route, altitude, and scanning angle for each patrol to be consistent with historical patrols of the area (e.g., setting three parallel routes along the boundary of overlapping areas, with a 2m spacing between each route and a height of 5m above the slope). This ensures that surface point cloud data acquired at different times have the same coordinate system and the same scanning perspective. Subsequent point cloud registration and difference analysis allow for precise quantification of changes (e.g., a crack extending 1.2m along its strike within 24 hours), and the change data can be stored and traced long-term, forming a "surface change ledger for overlapping areas."

[0066] By setting up a drone cruise control submodule to generate "encrypted inspection routes" for high-risk areas, such as scanning overlapping areas at a frequency of 2 times per hour (higher than the scanning frequency of non-overlapping areas) and a point cloud density of 500 points / m² (higher than the point cloud density of non-overlapping areas), high-risk areas are precisely inspected according to the inspection route, achieving targeted inspection and avoiding the inefficiency of traditional "all-area indiscriminate inspection". Traditional "all-area indiscriminate inspection" is not targeted and sometimes misses high-risk areas, while the inspection coverage of the high-risk areas mentioned in this application is significantly higher than that of "all-area indiscriminate inspection", making the inspection accurate and efficient.

[0067] The "3D model reconstruction submodule" of the working condition feedback adjustment module needs to optimize the model based on the latest point cloud data (e.g., point cloud weight of 1.2 in overlapping areas requires high-density point cloud support). The high-density 3D point cloud (above 500 points / ㎡) obtained by 3D surface cruise contains complete 3D morphological information of the slope surface and can be directly used as the basic data for model reconstruction. For example, in overlapping areas, the 3D point cloud can supplement the subtle undulations of the slope above the weak surface (e.g., bulges of 0.1-0.3m), so that the reconstructed 3D model can accurately reflect the spatial distribution of the internal weak surface and restore the three-dimensional changes of the surface, forming a closed loop of "3D cruise monitoring - point cloud data update - dynamic model reconstruction". This closed loop ensures that the 3D model always keeps consistent with the actual state of the slope, providing accurate model support for the strength parameter correction of the working condition feedback calculation module and the inspection route optimization of the working condition feedback adjustment module, ensuring the efficient operation of the entire dynamic modeling system.

[0068] The manual inspection submodule receives the three-dimensional coordinate range of the flashing red dot area calculated by the overlapping area calculation submodule, the three-dimensional coordinate range of the solid red dot area, the coordinate point information less than or equal to the threshold sent by the threshold comparison unit, and the warning time, etc., and generates manual inspection route and inspection frequency information. It selects several marked coordinate points one for installing GNSS positioning instruments and sets the location data acquisition frequency, and selects several marked coordinate points two for installing borehole television instruments and sets the borehole wall video data acquisition frequency. It sends the above-mentioned manual inspection route and inspection frequency information, GNSS positioning instrument installation coordinate points one and location data acquisition frequency information, borehole television instrument installation coordinate points two and borehole wall video data acquisition frequency information, warning time, etc. to the manual inspection terminal, and arranges manual inspection and preventive measures in a timely manner within the warning time. While 3D drone patrols and sensor monitoring can cover most areas, "physical blind spots" still exist due to technological limitations. For example, in areas with overlapping 3D regions, "deep concave cracks" (width < 5cm, depth > 1m) may be difficult for drone lidar to capture due to scanning angle obstruction. Similarly, in areas with loose rock accumulation near the overlapping region, sensors struggle to penetrate the rock to monitor the underlying weak surfaces. Manual inspections, however, can directly access high-risk areas through on-site surveys. For instance, in areas with flashing red dots (fully overlapping areas), manual inspections can use endoscopes to observe lithological changes on the crack walls (such as the presence of new shear surfaces). In areas with solid red dots (near the overlapping region), manual inspections can assess stability by tapping the rocks (a hollow sound indicates looseness), while simultaneously measuring real-time changes in crack width and length (with an accuracy of up to 1mm). This supplements the "microscopic morphological data" that technical equipment cannot obtain, upgrading high-risk area monitoring from "indirect inference from technical equipment" to "direct verification by manual on-site inspection," eliminating risk misjudgments caused by blind spots in technical monitoring.

[0069] One of the core functions of manual inspection is "equipment deployment"—the working condition feedback and adjustment module needs to obtain "fixed-point long-term data" through GNSS positioning instruments and borehole television cameras, and the accuracy of equipment deployment directly determines the validity of the data. Manual inspection can be based on the three-dimensional coordinates of flashing red dots and solid red dots, combined with the actual terrain (such as avoiding soft soil areas and choosing hard rock surfaces), to accurately install the GNSS positioning instrument at the "key monitoring point above the weak surface" (such as marked coordinate point one: X=102m, Y=205m, Z=548m, corresponding to directly above the weak surface), and set a high-frequency data acquisition frequency of 10 minutes / time to capture the three-dimensional displacement of the point in real time (especially the slight bulge in the Z-axis direction); at the same time, the operator can accurately install the borehole television camera at marked coordinate point two (such as borehole opening coordinates X=105m, Y=208m, Z=545m), ensuring that the lens is facing the weak surface, and observe the crack propagation inside the weak surface (such as whether it has increased from the original 2mm to 5mm) by acquiring borehole wall video once a day. This "precise manual deployment + long-term equipment monitoring" model constructs a three-dimensional network of "dynamic patrol (drone) + fixed-point monitoring (GNSS / drilling television)" to avoid invalid data caused by automatic deployment deviations (e.g., if the GNSS is installed in a loose soil area, the displacement data will be affected by surface subsidence).

[0070] When the threshold comparison unit sends "coordinate information of points less than or equal to the strength threshold" (e.g., weak surface cohesion ≤ 10 kPa, internal friction angle ≤ 10°), manual inspection can directly trigger "on-site emergency response." Unlike drones, which can only monitor and issue warnings, manual inspection can carry simple reinforcement equipment (e.g., anchor bolts, quick-setting concrete) to temporarily reinforce the risk area (e.g., for slopes in areas exceeding the threshold, manually drive in Φ20mm anchor bolts at 1m×1m intervals to suppress further slope deformation). Simultaneously, manual inspection can provide real-time feedback on the reinforcement effect through the inspection terminal (e.g., 2 hours after reinforcement, the GNSS positioning instrument shows the displacement decreased from 5mm / h to 0.5mm / h), providing a basis for decision-making regarding whether to initiate large-scale reinforcement. This instantaneous "warning-inspection-response-feedback" connection solves the shortcoming of technical equipment that "can only issue warnings but cannot intervene," shortening the risk response time from the traditional "4-6 hours of waiting for a professional team to arrive" to "within 1 hour," preventing further escalation of risks (e.g., the area exceeding the threshold developing from localized cracks into small landslides).

[0071] Data collected by drones and sensors may be deviated due to environmental interference (e.g., the Z-axis coordinate error of a drone's lidar may increase to 0.1m due to rain or fog; the vibration velocity data of a blasting vibration meter may be inflated due to interference from surrounding mechanical vibrations). Manual inspection can correct these deviations through on-site calibration. For example, for point cloud data in a three-dimensional overlapping area, a total station can be used to measure the precise coordinates at a preset reference point (e.g., X=100m, Y=200m, Z=550m in a mining area's independent coordinate system), and then compared with the point cloud coordinates acquired by the drone to calculate the error value (e.g., ...). The Z-axis deviation is 0.08m, and the calibration data is fed back to the data preprocessing module to correct subsequent point cloud data. For the strength data of the threshold comparison unit (such as the calculated cohesion value of 9.5kPa at a certain point), manual sampling can be carried out on-site (drilling rock cores near the marked coordinate points), and the actual cohesion can be tested in the laboratory (such as the measured 10.2kPa). The coefficients of the strength correction formula can be calibrated (such as adjusting 0.01 in the formula for the blasting cohesion attenuation rate to 0.009) to avoid model distortion caused by technical data deviation, so that the accuracy of dynamic modeling is always consistent with the actual on-site conditions.

[0072] Manual inspection is not an independent process, but rather forms a closed loop with the technical modules, characterized by "data exchange and functional complementarity": the field data obtained by manual inspection (such as crack changes and rock stability) can be supplemented to the modeling and updating module to optimize the 3D model; the long-term data obtained after equipment deployment (such as GNSS displacement and borehole television video) can be input into the working condition feedback calculation module to correct the weak surface strength parameters; the results of emergency response can be fed back to the working condition feedback adjustment module to optimize subsequent inspection routes (such as reducing the inspection frequency from once / day to once / 3 days in reinforced areas with excessive thresholds). This human-machine collaborative closed loop of "technical monitoring to identify risks - manual inspection to verify and handle issues - data feedback to optimize models" leverages the "wide-range, high-frequency" monitoring advantages of technical equipment while utilizing the "precise intervention and on-site verification" advantages of human intervention. This upgrades the management of weak slopes from "passive response" to "proactive prevention." Long-term application can reduce the incidence of open-pit mine slope instability accidents by more than 80%, while also reducing resource waste caused by over-reliance on technical monitoring (such as avoiding high-frequency drone patrols in areas that have already been confirmed as stable by humans).

[0073] The 3D model reconstruction submodule sets the weight of overlapping region point clouds to 1.2, the weight of adjacent overlapping region point clouds to 1.0, and the weight of non-overlapping region point clouds to 0.8. Based on the incremental insertion algorithm of Delaunay triangulation, it first selects high-weight points of overlapping regions from weak face point clouds and prioritizes inserting high-weight points into the empty set of the triangulation to be constructed. It uses Delaunay triangulation to construct a high-weight core skeleton, ensuring that each high-weight point can be preferentially connected to the surrounding high-weight points. On the basis of the high-weight core skeleton, the second-highest weight points of adjacent overlapping regions are inserted on the periphery. The adjacent overlapping region skeleton of the model is constructed through Delaunay triangulation. Finally, low-weight points of non-overlapping regions are inserted. The peripheral details of the model are constructed through Delaunay triangulation. The overlapping region point clouds are set as flashing red dots, and the adjacent overlapping region point clouds are set as solid red dots. At the same time, warning icons and 3D coordinate ranges are generated in the flashing red dot and solid red dot regions. Intensity information and thresholds are marked at each coordinate point. Clicking the warning icon with the mouse can view the overlap details, and touching the coordinate point with the mouse can display the intensity information and threshold of that point. The detailed steps for reconstructing the above model are as follows: 1) Point cloud data preprocessing and weight assignment The latest weak surface point cloud data (including overlapping areas, adjacent overlapping areas, and non-overlapping areas) is obtained from the working condition feedback calculation module, and all point clouds are imported into Delaunay triangulation software (such as Surfer or MeshLab). The three types of area point clouds are automatically identified and classified by point cloud attribute labels (marked by the overlapping area calculation submodule, such as "blinking red dot = overlapping area, solid red dot = adjacent overlapping area, no label = non-overlapping area"). Weights are assigned to three types of point clouds according to preset rules: overlapping point clouds are assigned a weight of 1.2 and labeled "high weight"; adjacent overlapping point clouds are assigned a weight of 1.0 and labeled "second highest weight"; and non-overlapping point clouds are assigned a weight of 0.8 and labeled "low weight". At the same time, the three-dimensional coordinate range of each type of point cloud is recorded (e.g., overlapping region X∈[95, 105]m, Y∈[195, 205]m, Z∈[540, 550]m) to define the boundaries for subsequent regional modeling. 2) Constructing a triangular mesh skeleton based on the incremental insertion algorithm with weighted distribution 2.1) Construct a "high-weight core framework" for overlapping regions. From the classified point clouds, the overlapping area point clouds with a weight of 1.2 are selected and sorted according to the principle of "spatial density uniformity" to avoid excessive local point cloud density leading to modeling redundancy: for example, in a 1m×1m×1m grid, if the number of point clouds exceeds 10, 5 key high-weight points are retained according to "from near to far from the grid center" to ensure that the point cloud distribution is both dense and non-redundant. Initialize the empty set of the triangulation network. First, insert the "reference high-weight points" of the overlapping regions (usually selected as the center coordinates of the region, such as X=100m, Y=200m, Z=545m) as the initial vertices of the triangulation network. Then, insert the remaining high-weight points into the empty set of the triangulation network one by one in the sorted order. For each insertion point, first determine its position in the existing triangulation network (whether it is inside, on, or outside the triangular facet). If it is outside, reconstruct the surrounding triangular facets using the "circumcircle criterion" (Delaunay triangulation core rule, ensuring that there are no other points inside the circumcircle of the triangular facet formed by the insertion point and the surrounding points), so that the newly inserted high-weight points are preferentially connected to the surrounding high-weight points. Repeat the insertion process until all high-weight points in the overlapping regions are inserted, forming a "high-weight core skeleton"—the side length of the triangular facets in this skeleton is strictly controlled between 0.3 and 0.5m (to ensure that the fine morphology such as slope bulging of 0.1 to 0.3m can be restored), and all triangular facets are composed of high-weight points to ensure the modeling accuracy of the overlapping regions. 2.2) Construct a "second-highest weight transition skeleton" in adjacent overlapping areas (connecting the core area and the outer area). Starting with the "boundary triangular facet" of the core skeleton of the overlapping region, the second-highest weight points are incrementally inserted towards the adjacent overlapping region: For each second-highest weight point, the spatial distance between it and the boundary point of the core skeleton is first calculated, and points with a distance ≤2m are inserted first. The second-highest weight points are connected to the boundary point of the core skeleton through Delaunay triangulation to form a "transition triangular facet" (side length controlled between 0.5 and 1m); the remaining second-highest weight points are gradually inserted outward until the entire adjacent overlapping region is covered to form a "second-highest weight transition skeleton" - this skeleton is seamlessly connected with the core skeleton, which maintains the modeling accuracy of the second-highest weight region and avoids morphological discontinuities between the two types of regions; Select non-overlapping point clouds with a weight of 0.8 and simplify them according to the principle of "reducing density and retaining key shapes": for example, within a 2m×2m×2m grid, retain only 3 to 5 low-weight points (prioritize retaining slope inflection points and edge points) to reduce the modeling workload of non-critical areas; starting from the boundary triangular facets of the transition skeleton of the adjacent overlapping area, insert low-weight points into the non-overlapping area: the side length of the triangular facets is controlled at 1 to 2m (no need for excessive precision, only to restore the overall slope shape), and the Delaunay circumcircle criterion is still followed during the insertion process, but the side length of local triangular facets is allowed to be slightly larger to balance modeling efficiency; after all low-weight points are inserted, "low-weight peripheral details" are formed, which are connected with the transition skeleton to finally form a complete weak-surface 3D triangular mesh model.

[0074] 3) Develop mouse interaction functionality: Clicking the warning icon will bring up the "overlap details window" (including region type, point cloud weight, intensity parameters, and historical change trends); clicking on a coordinate point will allow you to view the intensity data and corresponding thresholds for that point (such as raw parameters like blasting vibration velocity and excavation volume).

[0075] By setting the highest weight of 1.2 for overlapping areas, the core skeleton is constructed primarily using high-density, small-side-length triangular patches, which can accurately reproduce the subtle morphology of the area (such as 0.1m slope bulging and 0.05m crack width), with a modeling error of ≤0.03m. Non-overlapping areas (low-risk areas) are modeled with low weights and large side lengths, which reduces computational power consumption while ensuring the overall morphology. This "accuracy allocation on demand" design solves the problem of "insufficient accuracy in high-risk areas and excessive redundancy in low-risk areas" in the traditional "uniform accuracy modeling across the entire domain," enabling the model to support the intensity analysis of high-risk areas while also taking into account modeling efficiency. Traditional regional modeling often results in "boundary misalignment" due to the lack of correlation between point clouds in different regions (e.g., the model deviation between overlapping regions and adjacent regions exceeds 0.5m). However, this application adopts an incremental insertion logic of "core first, then transition, then periphery". When inserting the second highest weight points, it prioritizes connecting the boundary points of the core skeleton and connects the boundary points of the transition skeleton when inserting the low weight points. This allows the triangular meshes of the three types of regions to be seamlessly connected through "shared boundary triangular facets". The overall spatial continuity error of the model is ≤0.05m. This provides a complete model foundation for the subsequent working condition feedback calculation module to "accurately calculate the spatial relationship between overlapping regions and the range of working condition influence", avoiding misjudgment of overlapping regions due to model discontinuity. By using visual markers of "blinking red dots (high risk) and solid red dots (medium risk)," combined with warning icons and intensity labels, mine managers can intuitively distinguish risk levels through the model (without needing to analyze complex data). For example, seeing a flashing red dot area can directly identify it as an "overlapping risk area requiring urgent attention." Clicking the icon provides intensity parameters, allowing for a quick assessment of whether reinforcement needs to be initiated. This "visualization + interactivity" design reduces risk identification time from the traditional "1-2 hours of data analysis" to "1-2 minutes of model observation," significantly improving decision-making efficiency. When the subsequent model update submodule needs to update the model based on new point cloud data, it can only partially reconstruct the "point cloud within the weighted area" (e.g., after updating the point cloud in the overlapping area, only the high-weight core skeleton is rebuilt, without requiring a full model reconstruction). This reduces reconstruction time from 4-6 hours for a full model reconstruction to 1-2 hours for a partial update. This "modular modeling" design adapts to the system's "real-time dynamic update" requirements, avoiding resource waste caused by frequent full reconstructions and ensuring that the 3D model can quickly respond to changes in working conditions.By assigning differentiated weights to point clouds in different risk areas, the model's "accuracy priority" is perfectly matched with its "operating condition risk priority": overlapping areas, directly affected by operating conditions (such as blasting and excavation), require the highest accuracy modeling to reflect real-time changes; adjacent overlapping areas, as potential risk zones, require medium accuracy to monitor their trend of transformation into high-risk zones; non-overlapping areas are less affected by operating conditions, and low accuracy is sufficient to meet the requirements—this "risk-accuracy" correlation design makes the 3D model no longer a simple geometric shape reproduction, but a tool that can "quantitatively reflect the distribution of operating condition risks," providing core support for the closed loop of "operating condition feedback-model update-risk management" in the entire dynamic modeling system. The reconstruction model update submodule obtains the latest point cloud data of overlapping regions, adjacent overlapping regions, and non-overlapping regions from the working condition feedback calculation module. Based on the incremental insertion algorithm of Delaunay triangulation, it reconstructs one or all of the above-mentioned high-weight core skeleton, adjacent overlapping region skeleton, and peripheral details, and replaces the original 3D model.

[0076] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0077] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

Claims

1. An open-pit mine slope weak surface dynamic modeling system based on real-time working condition feedback, characterized in that, The system comprises a data acquisition module, a data preprocessing module, a modeling and updating module, a working condition feedback calculation module and a working condition feedback adjustment module. The data acquisition module is used for collecting slope data, working condition data and environmental data. The data preprocessing module is used for performing coordinate unification processing and time synchronization processing on the above data, and storing the processed data in a distributed database. The modeling and updating module is used for establishing a three-dimensional model of the open-pit mine slope and dynamically updating the three-dimensional model, marking the working positions of the excavator and the blasting in the three-dimensional model of the open-pit mine slope, and dynamically updating the three-dimensional model. The working condition feedback calculation module is used for calculating the overlapping area and the adjacent overlapping area of the working condition operation influence area and the three-dimensional model of the open-pit mine slope, marking and dynamically updating the overlapping area and the adjacent overlapping area, and calculating the strength data of each coordinate point of the three-dimensional weak surface in the three-dimensional model of the open-pit mine slope under the influence of the working condition operation, and marking, threshold comparison and dynamic updating. The working condition feedback adjustment module sends a warning message according to the calculation result of the working condition feedback, generates inspection information of the overlapping area and the adjacent overlapping area, reconstructs and updates the three-dimensional model of the open-pit mine slope, and makes the form of the overlapping area and the adjacent overlapping area more accurate.

2. The real-time working condition feedback based open-pit mine slope weak plane dynamic modeling system according to claim 1, characterized in that, The data acquisition module comprises a UAV sub-module, a borehole television instrument sub-module, a GNSS locator sub-module, an excavator data acquisition sub-module and a blasting data acquisition sub-module. The UAV sub-module controls the UAV to fly regularly along a pre-set slope path, controls the laser radar carried on the UAV to obtain three-dimensional coordinate data of the slope surface, controls the infrared thermal imager carried on the UAV to obtain thermal infrared data of the slope surface, and returns the UAV parameters. The borehole television instrument sub-module controls the borehole television instrument to obtain borehole wall video data of the slope regularly, and returns the borehole parameters. The GNSS locator sub-module obtains position data of key points of the slope regularly through the GNSS locator, the key points of the slope including the top of the slope, the foot of the slope, the step slope surface, the borehole opening and the blasting center point. The excavator data acquisition sub-module controls the GNSS locator installed on the excavator to obtain the position coordinates of the excavator, controls the vibration sensor installed on the excavator to obtain the vibration intensity data when the excavator is working, and controls the bucket weight sensor installed on the excavator to obtain the excavation amount of the excavator. The blasting data acquisition sub-module controls the blasting vibration meter to obtain blasting vibration velocity data, the blasting vibration meter being installed at a plurality of point clouds of the weak surface of the slope, and the blasting parameters being input through a man-machine interface unit.

3. The real-time working condition feedback based open-pit mine slope weak plane dynamic modeling system according to claim 2, characterized in that, The data preprocessing module comprises a coordinate unification sub-module and a time synchronization sub-module. The coordinate unification sub-module measures the national geodetic coordinates of the reference points pre-selected in the mining area, takes the reference point as the origin, sets the X axis parallel to the main mining direction of the mining area, the Y axis perpendicular to the X axis, and the Z axis along the vertical direction corresponding to the altitude, establishes the independent coordinates of the mining area according to the national geodetic coordinates, and converts the coordinate systems of the UAV, the borehole television instrument, the excavator and the GNSS locator into the independent coordinates of the mining area by using a coordinate conversion software. The time synchronization submodule builds a unified clock source taking the NTP server as a reference, configures "year-month-day time: minute: second: millisecond", controls the unmanned aerial vehicle, excavator, drilling television instrument, GNSS locator and blasting vibration instrument to synchronize the time of the above NTP server and periodically calibrates the time of each device to be consistent with the NTP server.

4. The real-time working condition feedback based open-pit mine slope weak plane dynamic modeling system according to claim 3, characterized in that, The modeling and updating module comprises a three-dimensional modeling submodule and a model updating submodule; The three-dimensional modeling submodule comprises a point cloud library construction unit and a point cloud library fusion modeling unit; The point cloud library construction unit calls the three-dimensional coordinate data of the slope surface obtained by the unmanned aerial vehicle and the key position data of the slope obtained by the GNSS locator, reads the data by using a point cloud library construction software, fuses discrete points and image points, generates a preliminary slope surface point cloud library one, deletes abnormal points, imports the preliminary point cloud library one into a point cloud density adjustment and optimization software, divides key and non-key areas, adjusts point density through encryption and thinning algorithms, and verifies the density to meet the standard; the hole wall video data of the slope drilling, the drilling orifice coordinate data and the drilling parameters are called, three-dimensional point cloud data of the internal weak surface of the slope are calculated through the above data, the three-dimensional point cloud data of the internal weak surface of the slope are read by using the point cloud library construction software, discrete points and image points are fused, a preliminary internal weak surface point cloud library one of the slope is generated, and abnormal points are deleted; the slope surface thermal infrared data obtained by the infrared thermal imager and the unmanned aerial vehicle parameters are called, the point cloud data of the cracks of the slope surface are calculated through the above data, the point cloud data of the cracks are read by using the point cloud library construction software, discrete points and image points are fused, a preliminary crack point cloud library one is generated, and abnormal points are deleted; The point cloud library fusion modeling unit fuses the slope surface point cloud library one, the internal weak surface point cloud library one and the crack point cloud library one data after the density verification to meet the standard, forms a complete slope point cloud library one, imports the slope point cloud library one into Surfer triangular net modeling software, and constructs an open-pit mine slope three-dimensional model, which comprises a three-dimensional weak surface curved surface, a three-dimensional slope surface curved surface and a three-dimensional crack; the position coordinates of the excavator are regularly called, the position coordinates are imported into the open-pit mine slope three-dimensional model, and a "excavator model" is generated at the corresponding coordinate position; the blasting center point coordinates are regularly obtained by the GNSS locator, the point cloud library fusion modeling unit imports the blasting center point coordinates into the open-pit mine slope three-dimensional model, and a thickened blasting center early warning red dot is generated at the corresponding coordinate position.

5. The real-time working condition feedback based open-pit mine slope weak plane dynamic modeling system according to claim 4, characterized in that, The model updating submodule comprises a point cloud library updating unit and a model dynamic updating unit; The point cloud library updating unit regularly collects new three-dimensional coordinate data of the slope surface and the key position data of the slope obtained by the GNSS locator, generates a new slope surface point cloud library one; regularly collects the hole wall video data of the new slope drilling, the drilling orifice coordinate data and the drilling parameters, generates a new internal weak surface point cloud library one of the slope; regularly collects new slope surface thermal infrared data and unmanned aerial vehicle parameters, and generates a new crack point cloud library one; The model dynamic updating unit respectively imports the above new slope surface point cloud library 1, the slope internal weak surface point cloud library 1 and the crack point cloud library 1 into Surfer triangular net modeling software, updates the old slope point cloud library 1, forms the new slope point cloud library 1, constructs the open-pit mine slope three-dimensional model by using the new slope point cloud library 1, and realizes dynamic updating of the open-pit mine slope three-dimensional model; new excavator position coordinates are collected regularly, the old excavator position coordinates are updated, the new excavator position coordinates are imported into the open-pit mine slope three-dimensional model, and a new "excavator model" is generated at the corresponding coordinate position; new blasting center point coordinates are obtained regularly by the GNSS locator, the old blasting center point coordinates are updated, the point cloud library 1 fusion modeling unit imports the new blasting center point coordinates into the open-pit mine slope three-dimensional model, and a new thickened blasting center warning red dot is generated at the corresponding coordinate position.

6. The real-time working condition feedback based open-pit mine slope weak plane dynamic modeling system according to claim 5, characterized in that, The working condition feedback calculation module comprises an influence range definition submodule, an overlapping area calculation submodule and a weak point strength calculation submodule; The influence range definition submodule defines the influence range of excavator excavation as "a sphere with a radius of 5 m with the excavator position as the center", the influence range of blasting as "a sphere with a radius of 10 m with the blasting center point as the center", and the influence range of surface cracks as "a region with a width of 2 m on both sides and below the cracks and a length of 2 m in front and behind along the crack direction"; the influence range of internal interlayer is "a region with a height of 3 m above and below the interlayer, a length of 10 m along the interlayer direction and a width of 3 m on both sides perpendicular to the interlayer direction"; The overlap area calculation sub-module respectively constructs the working condition influence range and the weak surface influence range by using a grid sampling method, divides the working condition influence range and the weak surface influence range according to a 1m*1m*1m grid size, takes the point cloud library-one of each influence range area as a reference, calculates the three-dimensional coordinates of each grid vertex in the independent coordinate system of the mining area, obtains the three-dimensional coordinates of each grid vertex, and constructs a new point cloud library-two of the three-dimensional coordinates of the grid vertex on each influence range area and the point cloud library-one of the corresponding influence range area; the point cloud library construction software traverses each point in the point cloud library-two of the working condition influence range area, judges whether each point belongs to the weak surface influence range area, if each point belongs to two ranges, the working condition influence range and the weak surface influence range completely overlap, if part of the points belong to two ranges, the working condition influence range and the weak surface influence range partially overlap, and if no point belongs to two ranges, the working condition influence range and the weak surface influence range do not overlap, finally, the overlapping points are marked as flashing red points, and a warning icon and a three-dimensional coordinate range of the overlapping area are generated in the overlapping area, and the icon can be clicked to view the overlap details; for the case that the working condition influence range and the weak surface influence range do not overlap, the spatial distance between any point in the working condition influence range and any point in the weak surface influence range is calculated by using a spatial two-point distance formula, the minimum value is taken as Y, if Y belongs to 0Y≤threshold-one, it is judged that the working condition influence range and the weak surface influence range belong to near overlap, in this case, the point cloud library-two of the working condition influence range and the point cloud library-two of the weak surface influence range are marked as solid red points respectively, and a warning icon and a three-dimensional coordinate range of the solid red point area are generated in the solid red point area, and the icon can be clicked to view the overlap details; the above-mentioned working condition influence range includes the influence range of the excavator excavation and the influence range of the blasting, and the above-mentioned weak surface influence range includes the influence range of the surface crack and the influence range of the internal interlayer.

7. The real-time working condition feedback based open-pit mine slope weak plane dynamic modeling system according to claim 6, characterized in that, The weak point strength calculation sub-module includes a data acquisition and preprocessing unit, a blasting calculation unit, a digging calculation unit and a threshold comparison unit; The data acquisition and preprocessing unit acquires the blasting vibration speed measured by the blasting vibration detector installed at the point cloud coordinates of the weak surface of the slope, calculates the vibration speed of each coordinate point in the point weak surface cloud library two by using a spatial interpolation algorithm according to the vibration speed of each marked point, and records each point vibration speed as V 振 ; simultaneously collects the blasting parameters, including the total charge of single blasting, recorded as Q; calculates the straight line distance between each coordinate point in the point cloud library two and the coordinate point of the blasting center by using a spatial two-point distance formula, recorded as R; acquires the excavation amount per square meter unit area of the excavator by the bucket weight sensor, recorded as W; acquires the three-dimensional coordinates of the excavator under the independent coordinate system of the mining area by the GNSS locator on the excavator, acquires the vertical coordinates of each coordinate point in the point weak surface cloud library two directly below the excavator by the three-dimensional model, calculates the vertical distance H between the excavation area and each coordinate point of the weak surface, H=Z 挖掘机 -Z 弱面 ; obtains the number of layers of different types of ore rock from the corresponding weak surface of the slope below the excavator and the Z-direction thickness of each layer of ore rock by the borehole television instrument, calculates the total load per unit area between the weak surface and the excavator above by the formula: P0= , wherein is the natural unit weight of the th layer of ore rock, is the vertical thickness of the th layer of ore rock, is the number of layers of ore rock above the weak surface, and the remaining load P above the weak surface is calculated, wherein P=P0-W; The blasting calculation unit calculates the cohesion attenuation rate k1 of each coordinate point in the point cloud library two by using the cohesion reduction rate formula: k1=0.01V+0.05×(Q / R) 振 , and corrects the cohesion C=C×(1-k1). 原 The blasting calculation unit calculates the internal friction angle reduction value of each coordinate point in the point cloud library two by using the internal friction angle reduction value formula: ∆φ=0.03V+0.1×(Q / R) 振 , and corrects the actual internal friction angle φ=φ-∆φ. 原 ​ And the reduced internal friction angle value is generated as a number at the corresponding coordinate point; The excavation computing unit calculates k1 according to the cohesion reduction rate formula: k1=0.02+0.0005W according to the above excavation amount W, the vertical distance H of each coordinate point of the excavation area and the weak surface, and the residual load P above the weak surface, calculates k1, and the actual cohesion C after correction is C=C 原 ×(1-K1); the internal friction angle reduction rate formula: K2=0.001W+0.02×(H / P) is used to calculate K2, the actual internal friction angle φ after correction is φ=φ 原 ×(1-K2); and the decayed cohesion and the reduced internal friction angle values are generated at the corresponding coordinate points. The threshold comparison unit sets the internal friction angle threshold and the cohesion threshold, simultaneously traverses the internal friction angle values of each coordinate point in the above-mentioned slope weak surface point cloud library-two, and compares them with the preset internal friction angle threshold to find the coordinate points with internal friction angle values less than or equal to the internal friction angle threshold; simultaneously traverses the cohesion values of each coordinate point in the above-mentioned slope weak surface point cloud library-two, and compares them with the preset cohesion threshold to find the coordinate points with cohesion values less than or equal to the cohesion threshold.

8. The real-time working condition feedback based open-pit mine slope weak plane dynamic modeling system according to claim 7, characterized in that, It also includes a working condition feedback adjustment module, which includes a pre-warning sub-module, an unmanned aerial vehicle cruise control sub-module, a manual inspection sub-module, a three-dimensional model reconstruction sub-module and a reconstructed model updating sub-module; The pre-warning sub-module triggers "emergency reinforcement warning", "key warning" and "routine warning" for the areas with strength less than or equal to the threshold, the overlapping area and the near overlapping area respectively, the above-mentioned warning signals are sent to the system end and the inspection terminal, and the relevant personnel receive the warning signals and arrange preventive measures in time within the warning time; The unmanned aerial vehicle cruise control submodule receives the flashing red dot area three-dimensional coordinate range calculated by the above-mentioned overlapping area calculation submodule, the above-mentioned solid red dot area three-dimensional coordinate range, the coordinate point information less than or equal to the threshold value sent by the threshold comparison unit, and the early warning information, generates an unmanned aerial vehicle patrol route based on the above-mentioned coordinate information, and sets the collection point cloud density information and the patrol frequency information, encapsulates the unmanned aerial vehicle patrol route, the collection point cloud density information, the cruise frequency information, and the early warning information as instructions, and then sends them to the unmanned aerial vehicle; The artificial patrol submodule receives the flashing red dot area three-dimensional coordinate range calculated by the above-mentioned overlapping area calculation submodule, the above-mentioned solid red dot area three-dimensional coordinate range, the coordinate point information less than or equal to the threshold value sent by the threshold comparison unit, and the early warning information, generates an artificial patrol route and patrol frequency information, and selects a plurality of marker coordinate points one for installing a GNSS positioning instrument and setting the position data acquisition frequency, and a plurality of marker coordinate points two for installing a borehole television instrument and setting the hole wall video data acquisition frequency, and sends the artificial patrol route and patrol frequency information, the GNSS positioning instrument installation coordinate points one and the position data acquisition frequency information, and the borehole television instrument installation coordinate points two and the hole wall video data acquisition frequency information to the artificial inspection terminal. The three-dimensional model reconstruction submodule sets the overlapping area point cloud weight to 1.2, the adjacent overlapping area point cloud weight to 1.0, and the non-overlapping area point cloud weight to 0.8, and based on the incremental insertion algorithm of Delaunay triangulation, first filters out the high-weight points in the overlapping area from the weak face point cloud, and preferentially inserts the high-weight points into the empty set of the triangular net to be constructed, constructs the high-weight core skeleton by using Delaunay triangulation, ensures that each high-weight point can be preferentially connected with the surrounding high-weight points, on the basis of the high-weight core skeleton, first inserts the next high-weight points in the adjacent overlapping area into the periphery, constructs the adjacent overlapping area skeleton of the model by using Delaunay triangulation, and finally inserts the low-weight points in the non-overlapping area, and constructs the periphery details of the model by using Delaunay triangulation. The overlapping area point cloud is set to a flashing red dot, the adjacent overlapping area point cloud is set to a solid red dot, the warning icons and three-dimensional coordinate ranges are generated in the flashing red dot and solid red dot areas, the intensity information is marked on each coordinate point, and the icon can be clicked by using a mouse to view the overlapping details. The reconstructed model updating submodule obtains the latest overlapping area point cloud, adjacent overlapping area point cloud, and non-overlapping area point cloud data information from the working condition feedback calculation module, reconstructs the above-mentioned high-weight core skeleton, adjacent overlapping area skeleton, and periphery details based on the incremental insertion algorithm of Delaunay triangulation, and replaces the original three-dimensional model.

Images