Point cloud fused blasting site digital management method and system

Abstract

The invention relates to the technical field of blasting engineering digital monitoring, and discloses a point cloud fused blasting site digital management method and system. The method comprises the steps of collecting multi-time-sequence point cloud data of a blasting site, extracting surface geometric information of an entity unit and generating a structure contour of the entity unit; performing space-time association on the entity units of adjacent time sequences, constructing a time dimension association network, and screening out stable units with constant geometric attributes and evolved dynamic units according to the time dimension association network; calculating a global statistical invariance interval for the stable unit, identifying a key time sequence node for the dynamic unit, and calculating an evolution stability boundary in stages; and when real-time point cloud data is received, qualified information is selected and combined to form a structured digital skeleton according to conformity judgment of a structure contour and a corresponding unit boundary. According to the method, automatic classification and differentiation stability modeling of entity behavior modes are realized, and the capability of continuously and finely evaluating the state of a blasting site is improved.

Description

Technical Field

[0001] This invention relates to the field of digital monitoring technology for blasting engineering, specifically to a digital management method and system for blasting sites that integrates point cloud data. Background Technology

[0002] Safety monitoring and effectiveness assessment at blasting sites typically employ 3D laser scanning to acquire point cloud data for digital management. Existing solutions primarily rely on independent analysis of multi-period discrete point cloud data. By registering and comparing point clouds from different time periods, geometric difference detection algorithms are used to identify areas where changes have occurred, and these areas are then modeled and their parameters extracted. This type of method enables macroscopic identification of scene changes.

[0003] Existing technologies have limitations. Essentially, they compare discrete-time "snapshots," only able to determine whether changes have occurred within a specific time interval, unable to construct a complete evolutionary trajectory of the same entity over a continuous timeline. Entities are isolated from each other in data from different time series, resulting in fragmented analysis results. Existing methods use a uniform analysis threshold and model to process all objects, failing to distinguish between geological structures that should remain static and dynamic targets such as explosive piles expected to move or break apart. This leads to the mixing of actual structural evolution with data noise and registration errors in the analysis results, reducing the accuracy and guiding value of state judgments.

[0004] A technology is needed to automatically identify and classify entities based on their behavioral patterns in time series. This technology needs to bridge the gap between discrete change detection and continuous behavior understanding, achieving effective separation between "stable" and "dynamic" entities. Furthermore, it requires the development of adaptive, quantitative state descriptions and stability assessment models for these two types of entities with different behavioral patterns, particularly methods capable of characterizing the stage-specific features of dynamic entity evolution, to support accurate real-time state perception and judgment. Summary of the Invention

[0005] The purpose of this invention is to provide a digital management method and system for blasting sites that integrates point clouds, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides a method for digital management of blasting sites based on point cloud data, the method comprising: Collect multi-temporal point cloud data of the blasting site, extract the surface geometric information of entity units with specific geological structures or artificial structures in the blasting scene; based on the surface geometric information of the entity units, generate structural contour information corresponding to each entity unit; Spatiotemporal associations are established between adjacent time-series entity units in the multi-time series point cloud data to create an association network of entity units along the time dimension. By traversing the association network, stable units whose geometric properties remain constant within the association network, and dynamic units whose geometric properties evolve within the association network are selected. For each stable unit, the statistical invariance interval of its structural profile information over the entire time dimension is calculated; for each dynamic unit, the key temporal nodes of its structural evolution in the interconnected network are identified, and the evolutionary stability boundary of the structural profile information in each time interval is calculated based on the time intervals divided by the key temporal nodes. Upon receiving real-time point cloud data, the system parses the real-time structural contour information and determines its conformity with the statistical invariance interval or evolutionary stability boundary of the corresponding unit. Based on the compliance judgment results, qualified structural contour information is selected from stable and dynamic units and combined to form a structured digital skeleton for describing the blasting site conditions.

[0007] Preferably, the extraction of surface geometric information of entity units with specific geological or artificial structures in the blasting scene includes, Geometric feature recognition is performed on the point cloud data of each time series to identify local high-density regions in the point cloud that exhibit planar, linear, and edge-like distributions. Boundary fitting is performed on each identified local high-density region to form a closed or semi-closed three-dimensional geometric surface; From the three-dimensional geometric surface, a set of geometric parameters that can characterize the spatial occupancy and surface morphology of the solid unit are extracted. The set of geometric parameters includes at least the center position, normal vector direction, area, perimeter, and three-dimensional coordinates of the boundary key points of the surface.

[0008] Preferably, the step of generating structural contour information corresponding to each solid unit based on the surface geometry information of the solid unit includes: Spatial clustering analysis is performed on the set of geometric parameters of multiple three-dimensional geometric surfaces belonging to the same entity unit, and geometric surfaces that are spatially adjacent and have the same normal vector direction are merged. For the merged geometric surface group, calculate its outer envelope three-dimensional cube. The size, spatial position, and orientation angle of the outer envelope three-dimensional cube constitute the structural contour information of the solid unit.

[0009] Preferably, the step of spatiotemporally associating adjacent temporal entity units in the multi-temporal point cloud data to establish an association network of entity units along the time dimension includes: Based on the spatial position of the outer envelope 3D cube in the structural contour information of the entity unit, unit matching is performed between data from different time series. If the spatial overlap of the outer envelope 3D cubes of two entity units in different time sequences exceeds a preset threshold, they are determined to be instances of the same entity unit in different time sequences. Directed connections are established between successfully matched entity unit instances. The starting point of the connection is the instance with an earlier time sequence, and the ending point is the instance with a later time sequence. All instances and connections together constitute an association network that describes the existence and evolution of entity units.

[0010] Preferably, the step of traversing the association network and filtering out stable units whose geometric properties remain constant within the association network, and dynamic units whose geometric properties evolve within the association network, includes: For each entity unit in the associated network, the size change sequence of the outer envelope 3D cube in the structural contour information of all its instances is statistically analyzed. Calculate the standard deviation of the size change sequence and compare it with a preset global size change threshold; If the standard deviation is less than the global size change threshold, the solid element is determined to be a stable element. If the standard deviation is greater than or equal to the global size change threshold, the entity unit is determined to be a dynamic unit.

[0011] Preferably, the step of calculating the statistical invariance interval of the structural profile information of each stable unit over the entire time dimension includes, Collect structural profile information of all instances of the stable unit in the associated network; For the structural contour information of all instances, the size, position and attitude angle parameters of the outer envelope three-dimensional cube are estimated by probability density, and the distribution curve of each parameter is obtained. For the distribution curve of each parameter, select the range of parameter values ​​corresponding to the region with the highest probability density. This range of parameter values ​​constitutes the statistical invariance interval of the parameter. By combining the statistical invariance intervals of size, position, and attitude angle, a multidimensional invariance interval describing the geometrical inherent properties of the stable unit is formed.

[0012] Preferably, the step of identifying key temporal nodes in the structural evolution of each dynamic unit within the associated network, and calculating the evolutionary stability boundary of the structural contour information within each time interval based on the time intervals divided by the key temporal nodes, includes: Extract the structural contour information of all instances of this dynamic unit in the associated network and arrange them in chronological order; Calculate the difference measure of structural contour information between adjacent temporal instances, which integrates the changes in size, position and orientation angle of the outer envelope 3D cube; On the time series curve of the difference measure, the local maximum point is located. The time series corresponding to the local maximum point is marked as the key time series node, which means that the unit structure has undergone significant evolution at this time. The time interval between two adjacent key time nodes is defined as an evolutionary stage; Within each evolutionary stage, the structural contour information of all instances of the dynamic unit is summarized, and the numerical distribution range of each geometric parameter within that stage is calculated. The numerical distribution range is the evolutionary stability boundary corresponding to that evolutionary stage.

[0013] Preferably, the step of parsing the real-time structural contour information and determining its conformity with the statistical invariance interval or evolutionary stability boundary of the corresponding unit when receiving real-time point cloud data includes: Real-time structural contour information of dynamic units is extracted from real-time point cloud data; The real-time structural contour information is compared with the evolutionary stability boundary of each evolution stage of the dynamic unit. Determine which evolutionary stability boundary the real-time structural contour information falls within; If it successfully falls within a certain evolutionary stability boundary, the real-time structural profile information of the dynamic unit is determined to be true. If it does not fall within any evolutionary stability boundary, the conformity is considered false.

[0014] Preferably, the step of selecting qualified structural contour information from stable and dynamic units based on the conformity judgment results and combining them to form a structured digital skeleton for describing the blasting site conditions includes: For entity units that pass the conformity test, their real-time structural contour information, which is parsed from the real-time point cloud data, is incorporated into the structured digital skeleton. For entity units that are deemed false in compliance, the median or center value is extracted from the statistical invariance interval corresponding to the unit or the evolutionary stability boundary of the current nearest evolutionary stage to generate alternative structural contour information, and the alternative structural contour information is incorporated into the structured digital skeleton. All qualified structural outline information that is included is organized according to the spatial topological relationship of the entity units it represents at the blasting site, forming a structured digital skeleton that fully describes the scene structure.

[0015] Preferably, the present invention further includes a fusion point cloud-based digital management system for blasting sites, used to implement the above-described fusion point cloud-based digital management method for blasting sites, the system comprising: The point cloud acquisition module is used to collect multi-temporal point cloud data from the blasting site; The contour information generation module is used to process the multi-temporal point cloud data, extract the surface geometric information of entity units with specific geological structures or artificial structures in the blasting scene, and generate structural contour information corresponding to each entity unit based on the surface geometric information of the entity unit. The association network construction module is used to spatiotemporally associate entity units with adjacent time series in the multi-time series point cloud data, establish an association network of entity units along the time dimension, and traverse the association network to filter out stable units whose geometric properties remain constant within the association network, as well as dynamic units whose geometric properties evolve within the association network. The attribute interval calculation module is used to calculate the statistical invariance interval of the structural profile information of each stable unit over the entire time dimension, and to identify the key time nodes of the structural evolution of each dynamic unit in the associated network. Based on the time intervals divided by the key time nodes, the module calculates the evolution stability boundary of the structural profile information in each time interval. The real-time conformity judgment module is used to parse real-time structural contour information when receiving real-time point cloud data and judge its conformity with the statistical invariance interval or evolutionary stability boundary of the corresponding unit. The digital skeleton synthesis module is used to select qualified structural contour information from stable and dynamic units based on the compliance judgment results, and combine them to form a structured digital skeleton for describing the blasting site conditions.

[0016] Compared with the prior art, the beneficial effects of the present invention are: By spatiotemporally associating entity units in multi-temporal point clouds and constructing an association network, the system can automatically classify entities into stable and dynamic units based on the constancy or evolution of geometric attributes over time. This technology abandons the isolated comparison of traditional discrete temporal sequences, enabling the tracking and pattern determination of the complete temporal behavior trajectory of entities. This allows for the automatic separation of background structures expected to remain static from moving targets expected to change within a complex scene, providing a clear and semantically defined object classification foundation for subsequent analysis. This eliminates the problem of data registration errors, environmental interference, and the confusion between real physical evolution caused by uniformly processing all entities, improving the accuracy and object-specificity of state analysis.

[0017] This method calculates the statistical invariance interval across the entire time series for stable units and identifies the temporal nodes of structural abrupt changes and divides the internal stages for dynamic units, then calculates the evolutionary stability boundary within each stage. This differentiated modeling method establishes adaptive state descriptions and evaluation criteria for entities with different behavioral patterns. The description of stable structures is globally consistent and rigorous, while the description of dynamic processes is staged and conforms to their discontinuous evolutionary laws. In real-time monitoring, it can make more refined and physically accurate judgments about the current state of an entity based on its category and corresponding specific boundaries, thereby supporting the hierarchical and object-specific precise evaluation and management of the blasting process and results. Attached Figure Description

[0018] Figure 1This is a schematic diagram illustrating the working principle of the digital management method for blasting sites based on fused point clouds as described in this invention. Figure 2 A flowchart for extracting surface geometry information of solid elements; Figure 3 A flowchart for establishing the spatiotemporal correlation network of entity units; Figure 4 Distribution diagram of key nodes for measuring the comprehensive time-series differences between adjacent dynamic units in a burst-type reactor; Figure 5 The time sequence node identification curve is used to measure the structural evolution differences of the dynamic unit (bursting pile A). Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see Figure 1 This invention provides a digital management method for blasting sites based on point cloud fusion. The method includes: firstly, by collecting multi-temporal point cloud data from the blasting site, extracting the surface geometric information of various entity units with specific geological or artificial structures in the blasting scene, and generating structural contour information corresponding to each entity unit based on this information. Subsequently, the entity units in adjacent temporal sequences from these multi-temporal point cloud data are spatiotemporally correlated to construct a correlation network describing the existence and changes of entity units along the time dimension. Traversing this correlation network allows for the selection of stable units whose geometric properties remain constant over time, as well as dynamic units whose geometric properties change. For stable units, the statistical invariance interval of their structural contour information over the entire time dimension is calculated; for dynamic units, key temporal nodes of their structural evolution are identified, and the evolutionary stability boundary of the structural contour information within each interval is calculated based on the time intervals divided by the nodes. Upon receiving real-time point cloud data, the real-time structural contour information of the entity units is parsed, and its conformity is judged against the pre-calculated statistical invariance interval or evolutionary stability boundary of the corresponding unit. Finally, based on the conformity judgment results, qualified structural contour information is selected from stable and dynamic units and combined according to their spatial relationships in the scene to form a structured digital skeleton to describe the current state of the blasting site.

[0021] In one embodiment of the present invention, see [reference] Figure 2Geometric feature identification is performed on point cloud data for each time series. This identification aims to pinpoint locally high-density regions exhibiting planar, linear, or edge-like distributions. Boundary fitting is then performed on each identified locally high-density region to form closed or semi-closed 3D geometric surfaces. From these 3D geometric surfaces, a set of geometric parameters characterizing the spatial occupancy and surface morphology of the entity unit is extracted. This set of parameters includes at least the surface's center position, normal vector direction, area, perimeter, and the 3D coordinates of key boundary points. Subsequently, spatial clustering analysis is performed on the geometric parameter sets of multiple 3D geometric surfaces belonging to the same entity unit, merging spatially adjacent surfaces with consistent normal vector directions. For the merged geometric surface group, its enclosing 3D cube is calculated. The dimensions, spatial position, and orientation angles of this enclosing 3D cube together constitute the structural contour information of the entity unit.

[0022] In practical implementation, a specific example scenario is a rock step surface at a blasting site. The collected single-frame point cloud data contains hundreds of thousands of disordered 3D points, which record the spatial positions of the step surface, blast hole openings, and surrounding fixed equipment. Geometric feature recognition is performed on this time-series point cloud data. The processing first calculates the local curvature and normal vector of each point in the point cloud. Then, a region growing algorithm is used to cluster adjacent points with curvature below a set threshold and consistent normal vector directions. These clustered regions are identified as locally high-density regions with a planar distribution, corresponding to the complete slope of the step or the flat surface of the equipment. At the same time, an edge detection operator is used to find continuous point sequences with abrupt changes in point cloud density and normal vector. These point sequences are identified as locally high-density regions with a linear and edge-like distribution, corresponding to the top line, bottom line, or boundary of the structure of the step.

[0023] In some embodiments, boundary fitting is performed on each identified local high-density region. For a local high-density region identified as having a planar distribution, a best three-dimensional plane equation is fitted using a random sampling consensus algorithm. For this fitted three-dimensional plane, its projection boundary points in the point cloud are extracted, and the Alpha Shape Algorithm is used to connect these boundary points to form a closed polygonal three-dimensional geometric surface. For a local high-density region identified as having a linear distribution, a three-dimensional B-spline curve is used to fit its discrete points to form a three-dimensional spatial curve, which is regarded as a semi-closed geometric contour. In its implementation, the Alpha Shape Algorithm begins with spatial relationship analysis of the boundary point set extracted from the point cloud. This algorithm defines the effective neighborhood of the point set by introducing an adjustable alpha radius parameter, thereby controlling the precision of the generated geometry. The algorithm iterates through all boundary points, calculates the Euclidean distance between each pair of points, and constructs a triangular mesh structure based on the alpha radius value. Only triangles with side lengths less than or equal to the alpha radius are retained, thus filtering out the set of triangles representing the outer contour of the point set. For the formed triangular mesh, the algorithm further identifies and extracts its outer boundary edges. The vertex sequence of these outer boundaries constitutes a closed two-dimensional polygonal contour. Subsequently, this two-dimensional polygonal contour is mapped three-dimensionally according to the spatial coordinate system of the original point cloud. By assigning it depth information and associating it with elevation data from the point cloud, a closed polygonal three-dimensional geometric surface is finally generated. This surface accurately fits the true boundaries of local high-density areas in the point cloud and ensures the continuity and integrity of the geometric contour.

[0024] In practice, a set of geometric parameters is extracted from the obtained three-dimensional geometric surface. For a closed polygonal three-dimensional geometric surface representing a stepped slope generated by fitting, the extracted set of geometric parameters includes the coordinates of the geometric center of the polygonal three-dimensional geometric surface, the direction of the average normal vector of the polygonal three-dimensional geometric surface, the projected area of ​​the polygonal three-dimensional geometric surface, the perimeter of the polygonal three-dimensional geometric surface, and the three-dimensional coordinates of all vertices of the polygonal three-dimensional geometric surface. From the three-dimensional B-spline curve representing the slope crest line, the extracted set of geometric parameters includes the coordinates of the curve's starting point, ending point, midpoint, and the total length of the curve.

[0025] Optionally, when identifying local high-density regions in the geometric feature recognition step, the following formulaic criteria are used to analyze each point in the point cloud and its local neighborhood:

[0026] in: Point The aggregation degree evaluation value, It is a point of Number of nearest neighbors and They are points and its first The unit normal vector of the nearest neighboring points, symbol This represents the dot product operation of vectors. This represents the absolute value of the cosine of the angle between the normal vectors of two points. It is a point The eigenvalues ​​of the covariance matrix of local neighborhood points and satisfying ,ratio It describes the degree of concentration of a local point set in the direction of the minimum eigenvalue.

[0027] In practice, spatial clustering analysis is performed on the geometric parameter sets of multiple three-dimensional geometric surfaces belonging to the same entity unit to identify six three-dimensional geometric surfaces. Their normal vector directions are calculated to fluctuate slightly around the vector (0.87, 0, 0.50), and their center positions are closely arranged in the same direction in space. Thresholds for the included angle of the normal vectors and the distance between the center points are set. When the included angle of the normal vectors of two three-dimensional geometric surfaces is less than 5 degrees and the distance between their center points is less than 1.5 meters, the three-dimensional geometric surfaces are determined to be spatially adjacent and have the same normal vector direction. Then, the six three-dimensional geometric surfaces are merged into a geometric surface group. This geometric surface group together describes a large rock mass structure surface.

[0028] In some embodiments, an outer envelope 3D cube is calculated for the merged geometric surface group, the 3D coordinates of all points in the geometric surface group are obtained, and the maximum and minimum values ​​of these points in the X, Y, and Z coordinate axes are calculated. Using these six extreme values, a minimum cuboid with its axis parallel to the global coordinate system is constructed. Furthermore, the size, spatial position, and attitude angle of the outer envelope 3D cube constitute the structural contour information of the entity unit. In this specific case, the attitude angle is (0,0,0) because the cuboid is axially aligned with the global coordinate system. In a specific implementation, the principal direction of the geometric surface group is determined by principal component analysis during the calculation process, and the axis of the outer envelope 3D cube is aligned with the principal direction, thereby obtaining an outer envelope 3D cube with a rotation angle in a more general case. At this time, the attitude angle of the outer envelope 3D cube is described by a rotation matrix or Euler angles.

[0029] In one embodiment of the present invention, see [reference] Figure 3Based on the spatial position of the outer 3D cube in the structural contour information of entity units, element matching is performed between point cloud data from different time series. If the spatial overlap of the outer 3D cubes of two entity units from different time series exceeds a preset threshold, they are determined to be instances of the same entity unit at different time series. Directed connections are established between successfully matched entity unit instances, with the starting point of the connection being an instance from an earlier time series and the ending point being an instance from a later time series. All instances and their directed connections together constitute an association network describing the existence and evolution of entity units. For each entity unit in the association network, the size change sequence of the outer 3D cube in the structural contour information of all its instances is statistically analyzed. The standard deviation of this size change sequence is calculated and compared with a preset global size change threshold. If the calculated standard deviation is less than the global size change threshold, the entity unit is determined to be a stable unit; if the standard deviation is greater than or equal to the global size change threshold, the entity unit is determined to be a dynamic unit.

[0030] In practical implementation, a specific example scenario involves processing point cloud data from three consecutive time periods (T1, T2, and T3) at a blasting site, specifically an area containing steps and blast piles. Point cloud data from time period T1 generates structural outline information for five entity units, time period T2 generates six, and time period T3 generates seven. Each entity unit's structural outline information includes an outer envelope 3D cube, defined by its center coordinates, size, and attitude angle. In practice, unit matching is performed between data from different time periods based on the spatial position of the outer envelope 3D cube within the entity unit's structural outline information. During operation, an entity unit from time period T1 is selected, whose outer envelope 3D cube's center coordinates are (…). , , Then, iterate through all entity units in time series T2, and calculate the coordinates of the outer envelope 3D cube center of each entity unit in time series T2 and ( , , The Euclidean distance between the two outer envelope 3D cubes is calculated, and the volume overlap rate of the two outer envelope 3D cubes in space is calculated simultaneously.

[0031] In some embodiments, a preset threshold is set as follows: the center point distance does not exceed 0.8 meters and the volume overlap exceeds 60%. In a specific implementation, if there is a solid unit in time series T2, the coordinates of the center of its outer envelope three-dimensional cube are ( +0.2, +0.1, The distance between the entity unit in time series T1 and the center point of unit T2 is 0.224 meters. By calculating the ratio of the intersection volume to the union volume of the two enclosing 3D cubes, the volume overlap is found to be 85%. This volume overlap exceeds a preset threshold, therefore, the entity unit in time series T1 and the entity unit in time series T2 are determined to be instances of the same entity unit in different time series. In specific implementation, directed connections are established between successfully matched entity unit instances, creating a directed connection from the entity unit instance in time series T1 to the entity unit instance in time series T2. This matching and connection operation is performed on entity unit instances of all time series. All successfully matched entity unit instances and the directed connections between them together constitute an association network describing the existence and evolution process of entity units.

[0032] Optionally, in the matching process, in addition to the spatial overlap, the attitude angles of the outer envelope 3D cubes can also be used for joint judgment. If the distance between the center positions of two outer envelope 3D cubes meets the preset threshold, but the difference in attitude angles exceeds 15 degrees, they are not determined to be the same entity unit instance. This helps to distinguish different objects that are close in position but whose attitudes have been drastically reversed.

[0033] In practical implementation, for each entity unit in the associated network, the size change sequence of the outer envelope 3D cube in the structural contour information of all its instances is statistically analyzed. For example, an entity unit identified as "slope A" in the associated network has instances in time series T1, T2, and T3. The length values ​​of the outer envelope 3D cube of these three instances are extracted to form a size change sequence. , , =[20.5 m, 20.3 m, 20.4 m]. Calculate the standard deviation of the size change sequence and compare it with a preset global size change threshold. In the specific implementation, the preset global size change threshold is 0.3 m. The standard deviation of the length size change sequence of "Slope A" is calculated to be 0.082 m. This standard deviation is less than the global size change threshold of 0.3 m, so the "Slope A" entity unit is determined to be a stable unit. Another entity unit identified as "Explosive Pile B" has a length size change sequence of [5.1 m, 8.7 m, 12.3 m] for time series T1, T2, and T3. The standard deviation of its length size change sequence is calculated to be 2.97 m. This standard deviation is greater than the global size change threshold of 0.3 m, so the "Explosive Pile B" entity unit is determined to be a dynamic unit.

[0034] In some embodiments, the formula for calculating the standard deviation of the size change sequence is:

[0035] in: The standard deviation of the size variation series, This represents the total number of instances of an entity unit in the association network. It is a time-series index of the instance. Indicates the first The size value of the outer envelope 3D cube of a time-series instance in a certain dimension. This represents the arithmetic mean of all size values ​​in the size variation sequence.

[0036] Optionally, the global size change threshold is not a fixed value and can be dynamically set based on the statistical distribution of the initial sizes of all entity units. For example, it can be set to 5% of the average initial size of all entity units to adapt to the sensitivity requirements of different scale blasting scenarios. It can be understood that by traversing the entire interconnected network and applying the above judgment logic, all entity unit instances in the network can be classified into their respective stable or dynamic unit categories, laying the foundation for subsequent differentiated statistical analysis.

[0037] In one embodiment of the present invention, structural contour information of all instances of the stabilizing unit in the associated network is collected. Probability density estimation is performed on the size, position, and attitude angle parameters of the outer envelope 3D cube in the structural contour information of all instances, thereby obtaining the distribution curve of each parameter. For each parameter's distribution curve, the range of parameter values ​​corresponding to the region with the highest probability density is selected; this range constitutes the statistical invariance interval of that parameter. Finally, the statistical invariance intervals of size, position, and attitude angle are combined to form a multidimensional invariance interval describing the geometrically inherent properties of the stabilizing unit.

[0038] In practical implementation, a specific example scenario involves a fixed slope structure identified as a stable unit in an interconnected network. This stable unit exists in all ten consecutive data acquisition times from T1 to T10. The structural contour information of all ten instances of this stable unit in the interconnected network is collected. Each structural contour includes the size parameters, spatial position parameters, and attitude angle parameters of the outer envelope 3D cube. In practice, probability density estimation is performed on each parameter of the outer envelope 3D cube in the structural contour information of all these instances. The probability density distribution curve of the length parameter is obtained using the kernel density estimation method.

[0039] In some embodiments, probability density estimation employs a Gaussian kernel function. The bandwidth of the kernel density estimation is adaptively determined based on the standard deviation of the sample data to smoothly fit the distribution pattern of the parameter values, thereby obtaining the distribution curve for each parameter. Peak region analysis is performed on the distribution curve of each parameter. For example, on the distribution curve of the length parameter, the region with the highest probability density is observed to correspond to the range from 20.35 meters to 20.55 meters. The parameter value range of 20.35 meters to 20.55 meters corresponding to this region with the highest probability density is selected, and this parameter value range constitutes the statistical invariance interval of the length parameter. In a specific implementation, the region with the highest probability density is determined using the following formulaic method:

[0040] in: This represents the interval of statistical invariance. This represents the numerical value of the input parameter. Represents the parameters obtained through kernel density estimation. The probability density value at that location. Represents the probability density function over the entire domain. The maximum value, It is a proportionality coefficient, ranging from 0.6 to 0.95. The formula means selecting probability density values ​​that are not lower than the maximum density value multiplied by the coefficient. All parameter values The set is used as the statistical invariance interval.

[0041] It is understandable that the above process is repeated for the position and attitude angle parameters of the outer envelope 3D cube of the stable unit. For example, the numerical distribution curve of the center point X-coordinate shows that its probability density is highest between 102.1 meters and 102.3 meters, and the probability density of the yaw angle is highest between -1.5 degrees and 1.5 degrees. These ranges are respectively determined as the statistical invariance intervals of each parameter. Combining the statistical invariance intervals of the nine parameters—length, width, height, X-coordinate, Y-coordinate, Z-coordinate, yaw angle, pitch angle, and roll angle—forms a nine-dimensional multidimensional invariance interval describing the geometrically inherent properties of this stable unit. Optional, the scaling factor... The selection of this parameter affects the tolerance of the statistical invariance interval. In practice, it can be set according to the accuracy of on-site measurements or the strictness of management requirements. For example, in scenarios requiring higher fault tolerance, it can be set as follows: Setting it to 0.8 will be beneficial in scenarios requiring high consistency. The value is set to 0.95. In practice, for some stable elements with minimal changes in attitude angle, the probability density distribution curve of their attitude angle parameters exhibits a very sharp peak. In this case, the calculated statistical invariance interval will be very narrow, accurately characterizing the stability of the element's spatial attitude.

[0042] In one embodiment of the present invention, the structural contour information of all instances of the dynamic unit in the associated network is extracted and arranged in chronological order. A difference metric is calculated between adjacent temporal instances, which integrates changes in the size, position, and orientation angle of the enclosing 3D cube. Local maxima are located on the time-series curve of the difference metric; the temporal sequences corresponding to these local maxima are marked as key temporal nodes, indicating that the unit structure has undergone significant evolution at this point. The time interval between two adjacent key temporal nodes is defined as an evolutionary stage. Within each evolutionary stage, the structural contour information of all instances of the dynamic unit is summarized, and the numerical distribution range of each geometric parameter within that stage is calculated; this numerical distribution range is the evolutionary stability boundary corresponding to that evolutionary stage. The real-time structural contour information of the dynamic unit is parsed from the real-time point cloud data, and the parsed real-time structural contour information is compared with the evolutionary stability boundaries of each evolutionary stage of the dynamic unit. The system determines which evolutionary stable boundary the real-time structural profile information falls within. If it successfully falls within a certain evolutionary stable boundary, the real-time structural profile information of the dynamic unit is deemed to have true conformity; otherwise, it is deemed to have false conformity.

[0043] In a practical implementation, a specific example scenario involves a burst of data identified as a dynamic unit in an interconnected network. This dynamic unit has instances at all eight consecutive acquisition times from T1 to T8. The structural contour information of all eight instances of this dynamic unit in the interconnected network is extracted and arranged in time sequence T1, T2, ..., T8. The structural contour information of each instance includes the size, position, and orientation angle of its outer envelope 3D cube. In practice, a difference metric is calculated between the structural contour information of adjacent time-series instances. This difference metric integrates the changes in size, position, and orientation angle of the outer envelope 3D cube. For example, the difference metric between the instance at time T1 and the instance at time T2 is calculated by weighted fusion of the normalized changes in all parameters. See Table 1.

[0044] Table 1: Calculation Table for Adjacent Timing Differences in Dynamic Units of a Bursting Reactor

[0045] In practical implementation, the following formula is used to calculate the comprehensive difference measure between adjacent time series t and t+1:

[0046] in: This represents the overall difference measure from time series t to time series t+1. and These represent the length values ​​of the outer envelope 3D cube for instances at time t and time t+1, respectively. This indicates that the maximum length among all instances of the dynamic unit is used for normalization. and These are the center coordinate vectors of the time series t and the time series t+1 instance, respectively. This represents the calculation of the Euclidean norm of a vector. The scene's reference scale is used to normalize the displacement. and These are the attitude angle vectors for instances at time t and time t+1, respectively. These are weighting coefficients assigned to changes in size, position, and orientation, and satisfy the following conditions: On the time series curve of the difference measure Δ, its local maxima are located. Analyzing the data in Table 1, the difference measure Δ shows significant peaks at time series T2->T3 and T4->T5. These two peaks are local maxima, and their corresponding time series T3 and T5 are marked as key time series nodes, which means that the structure of the burst reactor unit has undergone significant evolution at time series T3 and T5.

[0047] In some embodiments, the time interval between two adjacent key time nodes is defined as an evolutionary stage. Based on the marked key time nodes T3 and T5, three evolutionary stages can be divided: Stage 1, Stage 2, and Stage 3. Within each evolutionary stage, the structural contour information of all instances of the dynamic unit is summarized. For example, in Stage 2, the structural contour information of instances at time points T3, T4, and T5 is summarized. The numerical distribution range of each geometric parameter within this stage is calculated. For example, in Stage 2, the value of the length L of the outer envelope 3D cube is [7.5m, 8.2m, 10.3m]. Therefore, the numerical distribution range of length L in this evolutionary stage is 7.5 meters to 10.3 meters, and this numerical distribution range is the evolutionary stability boundary corresponding to this evolutionary stage. Similarly, the center coordinates within Stage 2 are calculated. , , The numerical distribution range of all parameters, including attitude angle, together constitute the evolutionary stability boundary of stage two.

[0048] It is understandable that the evolutionary stability boundary describes the reasonable limit of geometric parameter fluctuations of a dynamic unit within a specific evolutionary stage. The real-time structural outline information of this burst-type dynamic unit is parsed from real-time point cloud data; for example, the real-time parsed structural outline information shows a length of 9.5 meters and center coordinates... The length is 150.2 meters. The real-time structural outline information is compared with the evolutionary stability boundaries of each evolution stage of the dynamic unit of the blast reactor. The real-time length value of 9.5 meters and the center coordinate value of 150.2 meters are compared with the boundary ranges of stage one, stage two, and stage three. It is determined which evolutionary stage's stability boundary the real-time structural outline information falls within. The real-time length of 9.5 meters falls within the length boundary of stage two (7.5 meters to 10.3 meters), and the real-time center coordinate of 150.2 meters also falls within the center coordinate range of stage two. If all other parameters within the boundary also satisfy the boundary conditions of Stage Two, then the real-time structural profile information is determined to have successfully fallen within the evolutionary stability boundary of Stage Two, and the real-time structural profile information of this dynamic unit is considered to have true conformity. If any parameter exceeds the boundary range of all evolutionary stages, for example, if the real-time length is 12.5 meters, exceeding the upper limit of the length boundary of Stage Two (10.3 meters), and does not fall within the boundary of Stage One or Stage Three, then the conformity is determined to be false.

[0049] Optional, weighting coefficient The settings can be configured according to the degree of attention paid to changes in size, position, and posture in different scenarios. For example, in scenarios that emphasize volume changes, settings can be configured as follows: .

[0050] In some embodiments, the determination of whether real-time data conforms to a certain evolutionary stability boundary can adopt a "one-vote veto" system for all parameters, meaning that if any parameter exceeds the boundary, it is judged as non-conforming. Alternatively, a weighted scoring system can be used, allowing individual parameters to slightly exceed limits while the overall score remains within an acceptable range. It can be understood that by dividing dynamic units into evolutionary stages and calculating evolutionary stability boundaries, an adaptive, phased evaluation standard is provided for real-time monitoring, which can effectively distinguish between normal intra-stage fluctuations and abnormal inter-stage evolution or erroneous data.

[0051] See Figure 4In the structural evolution analysis of explosive reactor dynamic units, the extraction of temporal features for comprehensive difference measurement relies on the multi-parameter fusion calculation of adjacent instances. Specifically, the structural contour information of an explosive reactor dynamic unit is represented by a multi-dimensional parameter space composed of the size, position, and attitude angles of the enclosing three-dimensional cube. The comprehensive difference measurement between adjacent temporal instances is achieved by weighted fusion of normalized size changes, center displacement, and attitude angle changes: length changes are normalized using the maximum length of the unit instance, center displacement is normalized using the scene reference scale, and attitude angle changes are directly quantified and fused into a comprehensive difference measurement value according to preset weight coefficients. The degree of difference between two adjacent temporal states is reflected by the temporal fluctuation characteristics of this comprehensive value, and the location of its local maxima is used to identify key temporal nodes in structural evolution. During parameter configuration, the stable difference threshold is based on the comprehensive value of low-fluctuation temporal sequences (e.g., T1->T2), and the determination of local maxima is based on twice the increment of the difference between adjacent temporal sequences.

[0052] In one embodiment of the present invention, for entity units whose compliance determination is true, their real-time structural contour information parsed from real-time point cloud data is incorporated into a structured digital skeleton. For entity units whose compliance determination is false, the median or center value is extracted from the statistical invariance interval corresponding to the unit or the evolutionary stability boundary of the current nearest evolutionary stage to generate alternative structural contour information, which is then incorporated into the structured digital skeleton. All the incorporated qualified structural contour information is organized according to the spatial topological relationship of the entity units they represent at the blasting site to form a complete structured digital skeleton describing the scene structure.

[0053] In practical implementation, a specific example scenario involves processing real-time point cloud data containing both stable and dynamic units. After parsing, the real-time point cloud data yields the real-time structural contour information of five entity units. The conformity assessment results for these five units are as follows: one stable unit labeled "fixed slope" and two dynamic units labeled "blast pile A" and "blast pile B" are deemed to have true conformity; one stable unit labeled "loose rock" and one dynamic unit labeled "flying rock" are deemed to have false conformity. For entity units with true conformity assessments, such as "fixed slope," "blast pile A," and "blast pile B," their real-time structural contour information parsed from the real-time point cloud data is directly incorporated into the structured digital skeleton. This real-time structural contour information includes the accurate outer envelope 3D cube size, position, and orientation of these units at the latest moment.

[0054] In some embodiments, for entity units whose compliance assessment is false, numerical values ​​are extracted from the statistical invariance interval corresponding to the unit or the evolutionary stability boundary of the current nearest evolutionary stage to generate alternative structural contour information. For example, for a stable unit of "loose rock block" whose compliance assessment is false, its pre-calculated statistical invariance interval is queried, and the statistical invariance interval for the length parameter is... The statistical invariance interval of the X-coordinate of the center point is: From these intervals, median or center values ​​are extracted to generate alternative structural contour information with a length of 1.0 meter and a center point X coordinate of 50.6 meters. This alternative structural contour information is then incorporated into the structured digital skeleton. For "flying stone" dynamic units that are deemed false in the conformity assessment, since their real-time data does not fall within the evolutionary stability boundary of any evolutionary stage, the nearest current evolutionary stage to which their real-time moment belongs is queried. For example, if the unit is closest in time to the end of stage two, the center value is extracted from the evolutionary stability boundary of stage two to generate alternative structural contour information.

[0055] In practice, the operation of extracting the median or center value from an interval or boundary can be uniformly described by the following formula:

[0056] in: This represents the specific value of a geometric parameter generated from alternative structural contour information. This indicates the lower limit of the parameter's value within the statistical invariance interval on which it is based or the evolutionary stability boundary of the current nearest evolutionary stage. This represents the upper limit of the parameter's value within its statistical invariance interval or the evolutionary stability boundary of the current nearest evolutionary stage. The formula implies that the arithmetic mean of the upper and lower limits of the parameter's value range is used as a substitute value for that parameter. All qualified structural contour information included, including real contours from real-time analysis and substitute contours generated based on intervals, is organized according to the spatial topological relationships of the entity units they represent at the blasting site. For example, if "fixed slope" is determined to be located on the north side of the scene, and "blast pile A" is piled in front of "fixed slope," all units are associated and indexed according to their spatial adjacency and relative positional relationships, forming a complete three-dimensional structured digital skeleton describing the scene structure.

[0057] Optionally, when generating alternative structural contour information, if the parameter distribution exhibits significant asymmetry, the median within the interval can be used instead of the arithmetic mean to better represent the central tendency of the parameters. It can be understood that the structured digital skeleton integrates reliable real-time observation data with reasonable estimates based on historical statistics, enabling it to construct a coherent digital representation of the on-site spatial structure even when some real-time data is abnormal or missing.

[0058] See Figure 5 In the structural evolution analysis of the dynamic unit (bursting pile A), the temporal changes of the difference metric value characterize the evolution process of its structural contour information, and the identification of key temporal nodes depends on the local extremum analysis of the difference metric sequence. In specific operation, the difference metric value integrates the changes in size, position and attitude angle of the three-dimensional cube enveloping the bursting pile A, and its time series curve is divided into three evolution stages: In evolution stage 1 (0~20 seconds), the difference metric value rises rapidly from the initial value to a local peak (corresponding to the first key temporal node), reflecting that the bursting pile A structure undergoes significant morphological expansion in this stage; In evolution stage 2 (20~30 seconds), the difference metric value shows a fluctuating downward trend until the second key temporal node, indicating that the structural evolution rate slows down and tends to stabilize in this stage; In evolution stage 3 (30~50 seconds), the difference metric value rises again, corresponding to the bursting pile A structure entering a new rapid evolution stage. The location of key time nodes is based on the screening of local maxima of the difference measurement sequence. The corresponding time points are the boundary points where the burst pile A structure undergoes significant morphological transformation, providing a basis for the division of time intervals for the subsequent calculation of the evolutionary stability boundary.

[0059] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for digital management of a blasting site by fusing point clouds, characterized in that, The method comprises the following steps: Collecting multi-time point cloud data of the blasting site, and extracting surface geometric information of entity units with specific geological structure or artificial structure in the blasting scene; generating structure contour information corresponding to each entity unit based on the surface geometric information of the entity units; Correlating adjacent time entity units in the multi-time point cloud data in space-time, and establishing an associated network of the entity units along the time dimension; Traversing the associated network, screening out stable units whose geometric properties remain constant within the associated network, and dynamic units whose geometric properties evolve within the associated network; Calculating the statistical invariance interval of the structure contour information of each stable unit in the entire time dimension; identifying the key time nodes of the structure evolution of each dynamic unit in the associated network, and calculating the evolution stable boundary of the structure contour information in each time interval according to the time interval divided by the key time nodes; When receiving real-time point cloud data, analyzing real-time structure contour information, and judging its compliance with the statistical invariance interval or the evolution stable boundary of the corresponding unit; According to the compliance judgment result, selecting qualified structure contour information from the stable units and the dynamic units to form a structured digital skeleton for describing the state of the blasting site.

2. The method for digital management of blasting site according to claim 1, wherein, The extraction of the surface geometric information of the entity units with specific geological structure or artificial structure in the blasting scene comprises, Performing geometric feature recognition on each time point cloud data to identify the local high-density areas in the point cloud that present face-like, line-like and edge-like distribution; Boundary fitting is performed on each identified local high-density area to form a closed or semi-closed three-dimensional geometric surface; From the three-dimensional geometric surface, a set of geometric parameters capable of representing the spatial occupation and surface morphology of the entity unit is extracted, and the set of geometric parameters at least includes the center position, normal vector direction, area, perimeter and three-dimensional coordinates of the boundary key points of the surface.

3. The method for digital management of blasting site according to claim 2, characterized in that, The generation of the structure contour information corresponding to each entity unit based on the surface geometric information of the entity units comprises, Spatial clustering analysis is performed on the geometric parameter sets of multiple three-dimensional geometric surfaces belonging to the same entity unit, and geometric surfaces with adjacent spatial positions and consistent normal vector directions are merged; The outer envelope three-dimensional cube of the merged geometric surface group is calculated, and the size, spatial position and attitude angle of the outer envelope three-dimensional cube constitute the structure contour information of the entity unit. The space-time correlation of adjacent time entity units in the multi-time point cloud data to establish the associated network of the entity units along the time dimension comprises, 4. The method for digital management of blasting site according to claim 1, characterized in that, Based on the spatial position of the outer envelope three-dimensional cube in the structure contour information of the entity unit, the unit matching is performed between the data at different times; If the spatial position coincidence degree of the outer envelope three-dimensional cubes of the two entity units at different times exceeds the preset threshold, it is determined that the same entity unit exists at different times; A directed connection is established between the matched entity unit instances, the starting point of the connection is the earlier time instance, and the end point is the later time instance, and all instances and connections together constitute the associated network describing the existence and evolution process of the entity unit. ​ 5. The method for digital management of blasting site according to claim 4, characterized in that, The traversing the association network, screening out stable units with constant geometric properties in the association network and dynamic units with evolving geometric properties in the association network, comprises, For each entity unit in the association network, the size variation sequence of the outer envelope three-dimensional cube in the structure profile information of all instances thereof is counted; The standard deviation of the size variation sequence is calculated and compared with a preset global size variation threshold; If the standard deviation is less than the global size variation threshold, the entity unit is determined as a stable unit; If the standard deviation is greater than or equal to the global size variation threshold, the entity unit is determined as a dynamic unit.

6. The method for digital management of blasting site according to claim 1, wherein, The calculation of the statistical invariance interval of the structure profile information of each stable unit in the whole time dimension comprises, The structure profile information of all instances of the stable unit in the association network is collected; The size, position and attitude angle parameters of the outer envelope three-dimensional cube in the structure profile information of all instances are respectively subjected to probability density estimation, to obtain the distribution curve of each parameter; For the distribution curve of each parameter, the parameter value range corresponding to the region with the highest probability density is selected, and the parameter value range constitutes the statistical invariance interval of the parameter. The statistical invariance intervals of the size, position and attitude angle are combined to form a multi-dimensional invariance interval describing the geometric inherent properties of the stable unit.

7. The method for digital management of blasting site according to claim 1, wherein, The identification of the key time sequence nodes of the structure evolution of each dynamic unit in the association network and the calculation of the evolution stable boundary of the structure profile information in each time interval divided by the key time sequence nodes comprise, The structure profile information of all instances of the dynamic unit in the association network is extracted and arranged in time sequence; The difference measure of the structure profile information between adjacent time sequence instances is calculated, which comprehensively considers the changes in size, position and attitude angle of the outer envelope three-dimensional cube; The local maximum points on the time sequence curve of the difference measure are located, and the time sequence corresponding to the local maximum points is marked as a key time sequence node, which means that the unit structure significantly evolves at this time; The time interval between two adjacent key time sequence nodes is defined as an evolution stage. In each evolution stage, the structure profile information of all instances of the dynamic unit is summarized, and the numerical distribution range of each geometric parameter in the stage is calculated, which is the evolution stable boundary corresponding to the evolution stage.

8. The method for digital management of blasting site according to claim 7, characterized in that, The analysis of real-time structure profile information from real-time point cloud data and the judgment of the compliance of the real-time structure profile information with the statistical invariance interval or the evolution stable boundary of the corresponding unit comprise, The real-time structure profile information of the dynamic unit is parsed from the real-time point cloud data; The real-time structure profile information is compared with the evolution stable boundary of each evolution stage of the dynamic unit; It is determined that the real-time structure profile information falls within which evolution stable boundary; If the real-time structure profile information successfully falls within a certain evolution stable boundary, it is determined that the real-time structure profile information of the dynamic unit is compliant; If the real-time structure profile information does not fall within any evolution stable boundary, it is determined that the compliance is false.

9. The method for digital management of blasting site according to claim 1, wherein, The structural profile information of the stable units and the dynamic units is selected according to the conformity judgment result to form a structured digital skeleton for describing the state of the blasting site. For the entity unit with the true conformity judgment, the real-time structural profile information parsed from the real-time point cloud data is included in the structured digital skeleton. For the entity unit with the false conformity judgment, the median value or the center value is extracted from the statistical invariance interval or the evolution stability boundary of the current nearest evolution stage corresponding to the unit to generate alternative structural profile information, and the alternative structural profile information is included in the structured digital skeleton. All the included qualified structural profile information is organized according to the spatial topological relationship of the entity unit in the blasting site to form a structured digital skeleton for describing the scene structure.

10. A digital management system of blasting site with fusion point cloud, used to implement a digital management method of blasting site with fusion point cloud according to any one of claims 1 to 9, characterized in that, The point cloud acquisition module is configured to acquire multi-time point cloud data of the blasting site. The profile information generation module is configured to process the multi-time point cloud data, extract surface geometric information of an entity unit with a specific geological structure or artificial structure in the blasting scene, and generate structural profile information corresponding to each entity unit based on the surface geometric information of the entity unit. The association network construction module is configured to perform spatio-temporal association on adjacent time entity units in the multi-time point cloud data, establish an association network of the entity units along the time dimension, and traverse the association network to screen out stable units with constant geometric attributes in the association network and dynamic units with evolved geometric attributes in the association network. The attribute interval calculation module is configured to calculate a statistical invariance interval of the structural profile information of each stable unit in the entire time dimension, identify a key time node of the structural evolution of each dynamic unit in the association network, and calculate an evolution stability boundary of the structural profile information in each time interval according to a time interval divided by the key time node. The real-time conformity judgment module is configured to parse real-time structural profile information when receiving real-time point cloud data, and judge the conformity of the real-time structural profile information with the statistical invariance interval or the evolution stability boundary of the corresponding unit. The digital skeleton synthesis module is configured to select qualified structural profile information from the stable units and the dynamic units according to the conformity judgment result to form a structured digital skeleton for describing the state of the blasting site. ​

Images