A gas explosion-proof ball machine remote inspection method and system based on digital twinning

Abstract

The present application relates to the technical field of digital twinning, and particularly provides a gas explosion-proof ball machine remote inspection method and system based on digital twinning, which comprises the following steps: after time stamping and forcibly aligning video frames, gas concentration instantaneous values and temperature instantaneous values collected by the explosion-proof ball machine, a cylindrical coordinate mapping table with the gas pipeline center line as the reference is formed; a multi-modal anchor point data cluster is constituted; the change rate of the gas concentration value of each coordinate item before and after the action of the disturbance vector is calculated, and the arithmetic average of all the change rates is taken as the disturbance coefficient of the construction behavior; the local expansion potential of each point of the envelope surface is obtained; the three coordinate items with the highest local expansion potential are taken as the tracking target points of the ball machine holder; and the gas diffusion path after the envelope is broken is deduced according to the written disturbance vector value, with the risk response envelope as the boundary condition. The present application improves the sensitivity of risk early warning.

Description

Technical Field

[0001] This invention relates to the field of digital twin technology, and in particular to a remote inspection method and system for gas explosion-proof PTZ cameras based on digital twins. Background Technology

[0002] With the expansion of urban gas pipeline networks and the increasing complexity of the environment, traditional manual inspection methods face severe challenges in terms of real-time performance, coverage, and risk early warning capabilities. Current mainstream intelligent monitoring systems generally suffer from three limitations: a lack of spatiotemporal alignment mechanisms between multi-source sensor data, resulting in information silos between video surveillance, gas detection, and temperature monitoring; risk assessment relying on single parameter threshold triggers, making it difficult to identify early, gradual risks; and only qualitative assessments of the disturbance impact of surrounding construction activities, lacking quantitative correlation analysis. Especially in densely built-up areas, the coupling effect of pipeline corrosion, third-party damage, and other hidden dangers with complex environmental factors often leads to false alarms or missed reports by traditional monitoring systems.

[0003] Existing methods generally employ simple time synchronization or spatial interpolation, aligning sensor data through sliding time windows, but fail to consider the physical correlation between video grayscale gradients and gas concentration fields, resulting in feature fusion remaining superficial. Traditional risk envelope construction relies on static parameter thresholds, which cannot adapt to differences in material aging across different pipeline sections, nor can they respond to dynamic construction disturbances. Mainstream systems only implement a one-way monitoring-alarm process, lacking the ability to extrapolate risk evolution based on disturbance vectors, and resulting in insufficient matching between emergency response plans and real-time risk states. Summary of the Invention

[0004] To achieve the above objectives, the present invention adopts the following technical solution:

[0005] One aspect of the present invention provides a remote inspection method for gas explosion-proof PTZ cameras based on digital twins, comprising the following steps:

[0006] The video frames, instantaneous gas concentration values, and instantaneous temperature values ​​collected by the explosion-proof PTZ camera are forcibly aligned with timestamps to form a cylindrical coordinate mapping table based on the centerline of the gas pipeline; thus forming a multimodal anchor point data cluster.

[0007] Extract all coordinate entries in the multimodal anchor point data cluster whose gas concentration values ​​are higher than the background mean, and obtain the dynamic weight factor for each coordinate. Connect all coordinate entries whose dynamic weight factors are greater than a preset threshold into a non-closed surface, and define the surface as the risk response envelope. At the same time, convert the pre-entered construction behavior action library, including excavation impact, welding heat release rate, and mechanical vibration amplitude, into a superposition of spatial disturbance vectors. Apply the superposition vector to each coordinate entry in the multimodal anchor point data cluster, calculate the rate of change of the gas concentration value of each coordinate entry before and after the disturbance vector is applied, and take the arithmetic mean of all the rates of change as the disturbance coefficient of the construction behavior.

[0008] The local expansion potential of each point on the envelope surface is obtained; the three coordinate entries with the highest local expansion potential are used as the tracking target points of the PTZ camera; at the same time, the disturbance vector value corresponding to the construction behavior with the disturbance coefficient exceeding the warning threshold is written into the emergency exercise buffer of the remote command terminal. With the risk response envelope as the boundary condition, the gas diffusion path after the envelope rupture is deduced based on the written disturbance vector value, and the maximum diffusion radius obtained by deduction is back-calibrated as the upper bound of the dynamic weight factor in the next inspection.

[0009] Optionally, the explosion-proof PTZ camera integrates the pixel grayscale gradient in the video frame with the spatial weighting coefficient of the gas concentration measurement point to form a cylindrical coordinate mapping table based on the centerline of the gas pipeline. Each coordinate entry in the cylindrical coordinate mapping table simultaneously records the grayscale value, gas concentration value, and temperature value of the location, forming a multimodal anchor point data cluster.

[0010] Optionally, the temperature value of the coordinate entry is filtered by local mean and then multiplied by the gray value of the corresponding coordinate to obtain the dynamic weight factor of each coordinate; the excavation impact force, welding heat release rate, and mechanical vibration amplitude in the construction behavior action library are converted into spatial disturbance vectors and superimposed.

[0011] Optionally, the process of deducing the gas diffusion path after envelope rupture based on the written perturbation vector value includes the following steps:

[0012] The disturbance vector value written into the emergency drill buffer is used to calculate the fracture offset of each coordinate entry point by point according to the gray gradient direction of each coordinate entry in the cylindrical coordinate mapping table. Then, the fracture offset and the instantaneous temperature value of the coordinate entry are forcibly aligned with the timestamp and fused. After calibration by the observation tilt angle calculated by the current gimbal attitude angle of the explosion-proof PTZ camera, the fracture initiation coordinate cluster is obtained after the cylindrical mesh remapping of the fracture offset. The fracture initiation coordinate cluster constitutes the initial fracture point set starting from the risk response envelope surface.

[0013] Each point in the initial set of rupture points moves along the circumference and axial direction of the cylindrical coordinate mapping table, according to the moving step size constrained by the temperature value after filtering the local mean and the gas concentration value at the point, and jumps to the adjacent coordinate entries in sequence. After each jump, the correlation strength between the gray value of the coordinate entry and the background mean is used as the jump retention probability. When the jump retention probability is lower than the diffusion termination threshold, the jump stops. The coordinate entries passed through each jump, the current gas concentration value of the rupture point, and the jump retention probability are stored in an array in the order of the jumps to form the discretized propagation trajectory of the rupture point. The propagation trajectory forms a coordinate sequence of each rupture point carrying a concentration label.

[0014] In the discretized propagation trajectories of all rupture points, coordinate entries whose radial distance from the gas pipeline centerline exceeds the maximum radial distance of the current risk response envelope are concatenated into a spatial polyline in the order of forced alignment by the explosion-proof PTZ camera timestamps. The coordinate entry corresponding to the maximum concentration change rate recorded in the last inspection by the explosion-proof PTZ camera is embedded into the polyline as the trajectory anchor point. After binding the trajectory line and anchor point and matching it with the boundary conditions of the emergency drill buffer, the gas diffusion path after the envelope rupture is obtained. The farthest radial distance from the end of the path to the pipeline centerline is taken as the maximum diffusion radius. The gas diffusion path and the maximum diffusion radius are directly written into the simulation results of the display layer of the emergency drill buffer of the remote command terminal.

[0015] Optionally, the process of binding the trajectory line to the anchor point and matching it with the boundary conditions of the emergency drill buffer zone includes the following steps:

[0016] The gray-scale gradient direction of each coordinate entry in the embedded polyline of the trajectory anchor points is compared with the gray-scale gradient direction of the trajectory line and the anchor point. The coordinate entries with the angle between the two less than the minimum step angle of the explosion-proof PTZ camera are taken as binding nodes. After replacing the binding nodes with the coordinate values ​​of the trajectory anchor points and their maximum concentration change rate labels in the order of timestamps, the modified polyline with the historical risk feature identifier of the initial diffusion path is obtained after the anchor points are embedded.

[0017] The predicted gas concentration value of each coordinate entry on the initial diffusion path is compared with the historical concentration benchmark value of the coordinate entries pre-stored in the emergency drill buffer. The coordinate entries whose absolute value of the difference exceeds the concentration fluctuation tolerance corresponding to the disturbance vector magnitude are taken as valid path points. After removing invalid path points in the cylindrical circumferential order and reconnecting the remaining valid path points, a filtered path matching the boundary conditions of the emergency drill buffer is obtained. The filtered path forms a diffusion path skeleton that conforms to the geometric constraints of the current risk response envelope.

[0018] The radial distance of each coordinate entry in the diffusion path skeleton is compared with the predefined pipeline safety radius in the explosion-proof PTZ camera inspection digital twin. The coordinate entry whose radial distance is first less than the pipeline safety radius is taken as the path endpoint. After processing the output of all coordinate entries from the start point to the end point of the skeleton in the order of forced timestamp alignment, the gas diffusion path after envelope rupture is obtained.

[0019] Optionally, the process of forming a diffusion path skeleton that conforms to the current risk response envelope geometry after filtering includes the following steps:

[0020] The predicted gas concentration for each coordinate entry on the initial diffusion path is compared point-by-point with the historical concentration baseline value of the coordinate entries pre-stored in the emergency drill buffer. The absolute value of the difference is then taken. This absolute value is combined with the disturbance vector magnitude and the diffusion direction attenuation factor calculated from the current gimbal attitude angle of the explosion-proof PTZ camera to determine the final result.

[0021] High fluctuation points in the concentration fluctuation marker pairs of each coordinate entry are selected as candidate valid points. Then, the points are filtered according to the angle between the gray gradient direction corresponding to the point and the normal of the risk response envelope boundary: candidate points with an angle smaller than the minimum step angle of the explosion-proof PTZ camera are retained, and candidate points with an angle greater than or equal to the minimum step angle are removed. After traversing all candidate points in cylindrical grid order and recording the retained points, the set of valid path points is obtained. The set of valid path points constitutes the sparse coordinate sequence after removing interference points.

[0022] Points in the sparse coordinate sequence are sorted in ascending order of their circumferential angle in the cylindrical mapping table, and then sorted again by axial distance. Adjacent points are connected in sequence to form a polyline. During the connection, the cylindrical grid distance between adjacent points is checked to see if it exceeds the spatial resolution threshold corresponding to the optical zoom magnification of the explosion-proof PTZ camera. After filtering, the path forms a diffusion path skeleton that conforms to the geometric constraints of the current risk response envelope.

[0023] Optionally, if the absolute value of the difference is greater than the product of the magnitude of the disturbance vector and the attenuation factor, the point is marked as a high fluctuation point; otherwise, it is marked as a low fluctuation point. This yields a concentration fluctuation marker pair for each coordinate entry, and the concentration fluctuation marker pairs form a binary state set for determining whether a point should be retained.

[0024] Optionally, if the spatial resolution threshold corresponding to the optical zoom magnification of the explosion-proof PTZ camera is exceeded, a new coordinate point with the smaller of the predicted gas concentration values ​​of the two points is inserted at the midpoint of the line connecting adjacent points as the concentration label; after connection and interpolation point supplementation, a filtered path matching the boundary conditions of the emergency drill buffer zone is obtained.

[0025] Optionally, the process of constructing a sparse coordinate sequence after removing interference points from the effective path point set includes the following steps:

[0026] The gray-level gradient direction vector corresponding to each candidate valid point and the normal vector of the risk response envelope boundary at the point are respectively input into the direction angle calculation in the explosion-proof PTZ camera inspection digital twin. The direction angle value of each candidate valid point is obtained by querying the pre-stored gray-level gradient-normal angle mapping table.

[0027] The minimum step angle of the pan-tilt unit (PTZ) at the current optical zoom level of the explosion-proof PTZ camera is read in real time, which is the physical minimum rotation angle of the PTZ stepper motor. The directional angle value of each candidate valid point is compared with the minimum step angle point by point: if the angle value is less than the minimum step angle, it is marked as a trackable point; if the angle value is greater than or equal to the minimum step angle, it is marked as an untrackable point. The trackable points are then written into the valid point queue of the inspection digital twin buffer according to the circumferential scanning order of the cylindrical grid, thus obtaining the initial set of valid points of the intermediate point sequence after filtering by the PTZ tracking capability.

[0028] The cylindrical grid distance between two adjacent points in the intermediate point sequence is compared with the field of view coverage of the explosion-proof PTZ camera at maximum optical zoom. After traversing all adjacent points and completing the insertion process, an effective path point set that meets the requirements of continuous tracking of the PTZ is obtained. The effective path point set constitutes a sparse coordinate sequence that eliminates interference points and ensures the accessibility of the PTZ.

[0029] Another aspect of the present invention provides a remote inspection system for gas explosion-proof PTZ cameras based on digital twins, comprising:

[0030] The weighting coefficient fusion module is used to merge the video frames, instantaneous gas concentration values, and instantaneous temperature values ​​collected by the explosion-proof PTZ camera with the spatial weighting coefficients of the gas concentration measurement points after the video frames, gas concentration, and temperature values ​​are forcibly aligned with the timestamps. This forms a cylindrical coordinate mapping table with the center line of the gas pipeline as the reference. Each coordinate entry in the cylindrical coordinate mapping table records the gray value, gas concentration value, and temperature value of the location, forming a multimodal anchor point data cluster.

[0031] The disturbance coefficient acquisition module is used to extract all coordinate entries in the multimodal anchor point data cluster whose gas concentration values ​​are higher than the background mean. After local mean filtering of the temperature values ​​of the coordinate entries, the module multiplies them with the gray values ​​of the corresponding coordinates to obtain the dynamic weight factor of each coordinate. All coordinate entries with dynamic weight factors greater than a preset threshold are connected into a non-closed surface, which is defined as the risk response envelope. At the same time, the pre-entered construction behavior action library, including excavation impact, welding heat release rate, and mechanical vibration amplitude, is converted into a superposition of spatial disturbance vectors. The superposition vector is applied to each coordinate entry in the multimodal anchor point data cluster, and the rate of change of the gas concentration value of each coordinate entry before and after the disturbance vector is applied is calculated. The arithmetic mean of all the rates of change is taken as the disturbance coefficient of the construction behavior.

[0032] The diffusion path deduction module is used to perform a dot product between the boundary coordinate entries of the obtained risk response envelope and the spatial disturbance vector corresponding to the construction behavior with a disturbance coefficient greater than zero, to obtain the local expansion potential of each point on the envelope surface. The three coordinate entries with the highest local expansion potential are used as the tracking target points of the PTZ camera, and the optical zoom of the explosion-proof PTZ camera is triggered to the maximum focal length to capture the visual details of the target points in sequence. At the same time, the disturbance vector value corresponding to the construction behavior with a disturbance coefficient exceeding the warning threshold is written into the emergency exercise buffer of the remote command terminal. The buffer uses the risk response envelope as the boundary condition, and deduces the gas diffusion path after the envelope ruptures based on the written disturbance vector value. The maximum diffusion radius obtained from the deduction is back-calibrated as the upper bound of the dynamic weighting factor in the next inspection.

[0033] This invention achieves accurate modeling of the three-dimensional risk field of gas pipelines through spatiotemporal alignment and fusion of multimodal sensor data. The mapping table based on cylindrical coordinates overcomes the limitations of traditional two-dimensional monitoring, organically coupling visual features with physical field parameters to form a spatially continuous risk characterization system. Specifically, the synergistic mapping of pixel grayscale gradients and gas concentration effectively identifies early leakage characteristics caused by microscopic deformation of the pipe wall. Secondly, the innovative calculation mode of dynamic weighting factors significantly improves the sensitivity of risk warning. Through the product operation of local temperature field and grayscale features, weak signal features that are difficult to detect by traditional single-parameter monitoring are amplified. The dynamic evaluation mechanism based on construction disturbance vectors enables quantitative analysis of the impact of mechanical operations on pipeline risk status, providing a theoretical basis for proactive defense. The risk response envelope generation technology pioneers a new paradigm for non-contact monitoring. This surface boundary not only dynamically reflects the geometric characteristics of the current risk distribution, but its dot product operation with the disturbance vector reveals the energy accumulation law of potential risk points. The prediction method based on field strength calculation detects risk points earlier than traditional threshold alarm mechanisms. The collaborative optimization of the intelligent tracking and emergency simulation system has enabled the effective integration of the monitoring-early warning framework; the automatic locking capability of the optical zoom system for high-risk targets has greatly improved the efficiency of detailed data acquisition in emergency situations, while the diffusion path simulation based on the fluid dynamics model provides a scientific basis for emergency decision-making. Attached Figure Description

[0034] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0035] Figure 1 This is a flowchart of the remote inspection method for gas explosion-proof PTZ camera based on digital twin provided in Embodiment 1 of the present invention;

[0036] Figure 2 This is a schematic diagram of the remote inspection method for gas explosion-proof PTZ cameras based on digital twins provided in Embodiment 1 of the present invention;

[0037] Figure 3 This is a process diagram of constructing a multimodal anchor point data cluster provided in Embodiment 2 of the present invention;

[0038] Figure 4 This is a diagram illustrating the process of calculating the rate of change of the gas concentration values ​​of each coordinate entry before and after the disturbance vector is applied, as provided in Embodiment 4 of the present invention.

[0039] Figure 5 This is a process diagram of deduce the gas diffusion path after envelope rupture based on the written disturbance vector value provided in Embodiment 7 of the present invention;

[0040] Figure 6 This is a block diagram of the remote inspection system for gas explosion-proof PTZ cameras based on digital twins provided in Embodiment 11 of the present invention;

[0041] Figure 7 A block diagram of the electronic device provided by the present invention;

[0042] Figure 8 A block diagram of a computer-readable storage medium provided for this invention.

[0043] Reference numerals in the attached diagram: 1. Weighting coefficient fusion module; 2. Disturbance coefficient acquisition module; 3. Diffusion path deduction module; 4. Central processing unit / microprocessor / main control chip; 5. Storage medium; 6. Data bus; 7. Input / output bus / external bus / device bus; 8. Display; 9. Input / output device; 10. Computer-readable instructions; 11. Non-transitory computer-readable storage medium. Detailed Implementation

[0044] The technical solutions of the present invention will now be described with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0045] Hereinafter, the terms "first," "second," etc., are used for descriptive convenience only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "multiple" means two or more.

[0046] In this invention, unless otherwise explicitly specified and limited, the term "connection" should be interpreted broadly. For example, a connection can be a fixed mechanical connection, a detachable mechanical connection, or an integral part; or, a connection can be a direct connection or an indirect connection through an intermediate medium. Furthermore, unless otherwise explicitly specified and limited, the term "coupling" should be interpreted broadly. For example, coupling can be a direct electrical connection, such as physical contact and electrical conduction between two components; it can also be understood as an electrical connection between different components in a circuit structure through physical lines capable of transmitting electrical signals, such as copper foil or wires on a printed circuit board (PCB), to transmit electrical signals; or, coupling can be an indirect electrical connection between two components through an intermediate medium; or, coupling can be an electrical connection between two components in a non-contact manner, such as an electrical connection between two components using capacitive coupling to transmit electrical signals.

[0047] In this embodiment of the invention, directional terms such as up, down, left, and right may be defined relative to the orientation of the components shown in the accompanying drawings. It should be understood that these directional terms can be relative concepts, used for relative description and clarification, and can change accordingly depending on the orientation of the components in the accompanying drawings.

[0048] Example 1: As Figure 1 As shown in the figure, this embodiment of the invention provides a remote inspection method for gas explosion-proof PTZ cameras based on digital twins, specifically including the following steps:

[0049] Step S100: After the video frames, instantaneous gas concentration values, and instantaneous temperature values ​​collected by the explosion-proof PTZ camera are forcibly aligned with the timestamps, the pixel grayscale gradients in the video frames are fused with the spatial weighting coefficients of the gas concentration measurement points to form a cylindrical coordinate mapping table based on the centerline of the gas pipeline; each coordinate entry in the cylindrical coordinate mapping table simultaneously records the grayscale value, gas concentration value, and temperature value of the location, forming a multimodal anchor point data cluster;

[0050] Step S200: Extract all coordinate entries in the multimodal anchor point data cluster whose gas concentration values ​​are higher than the background mean. After local mean filtering of the temperature values ​​of the coordinate entries, multiply them with the gray values ​​of the corresponding coordinates to obtain the dynamic weight factor of each coordinate. Connect all coordinate entries whose dynamic weight factors are greater than the preset threshold into a non-closed surface, and define the surface as the risk response envelope. At the same time, convert the pre-entered construction behavior action library, including excavation impact, welding heat release rate, and mechanical vibration amplitude, into a superposition of spatial disturbance vectors. Apply the superposition vector to each coordinate entry in the multimodal anchor point data cluster, calculate the rate of change of the gas concentration value of each coordinate entry before and after the disturbance vector is applied, and take the arithmetic mean of all the rates of change as the disturbance coefficient of the construction behavior.

[0051] Step S300: Take the boundary coordinate entries of the obtained risk response envelope and perform a dot product with the spatial disturbance vector corresponding to the construction behavior with a disturbance coefficient greater than zero to obtain the local expansion potential of each point on the envelope surface; take the three coordinate entries with the highest local expansion potential as the tracking target points of the PTZ camera, and sequentially trigger the optical zoom of the explosion-proof PTZ camera to the maximum focal length to capture the visual details of the tracking target points; at the same time, write the disturbance vector value corresponding to the construction behavior with a disturbance coefficient exceeding the warning threshold into the emergency exercise buffer of the remote command terminal. The buffer uses the risk response envelope as the boundary condition, and deduces the gas diffusion path after the envelope ruptures based on the written disturbance vector value, and reverse-calibrates the maximum diffusion radius obtained from the deduction as the upper bound of the dynamic weight factor in the next inspection.

[0052] Among them, the explosion-proof PTZ camera is a spherical camera that meets explosion-proof electrical equipment standards and can operate safely in hazardous gas environments. It integrates optical zoom, pan-tilt control, and multi-sensor interfaces, responsible for video acquisition and also serving as the execution end, automatically tracking high-risk target points based on digital twin analysis results. Three types of real-time sensing data are used: video frames, instantaneous gas concentration values, and instantaneous temperature values. Video frames provide visual texture and grayscale gradients of the pipeline surface and surrounding environment; instantaneous gas concentration values ​​come from electrochemical or laser sensors, reflecting the intensity of gas leakage; instantaneous temperature values ​​are collected by infrared or contact temperature measuring elements. These three types of data are forcibly aligned using timestamps, solving the problem of asynchronous multi-source data and providing a time-consistent input for building a unified digital twin space. The cylindrical coordinate mapping table is a digital coordinate system established based on the centerline of the gas pipeline. It merges gradient-weighted planar video pixels with three-dimensional concentration and temperature measurement points, mapping them onto the pipeline cylindrical grid. Each coordinate entry records the grayscale, concentration, and temperature values ​​at that location, forming a multimodal anchor point data cluster, which is equivalent to establishing a calculable and queryable digital outer layer for the pipeline. The background mean is the baseline value of gas concentration statistically calculated along the pipeline cylindrical coordinates under leak-free or normal operating conditions. Environmental background fluctuations are filtered out, and only coordinate entries with concentrations significantly higher than the background are extracted, focusing on truly abnormal areas and avoiding false alarms. A preset threshold is used to determine whether the dynamic weighting factor is sufficient to constitute a numerical threshold for risk; its setting directly affects the tightness and sensitivity of the risk response envelope. The risk response envelope is an open surface formed by connecting all coordinate entries with dynamic weighting factors greater than the preset threshold in three-dimensional space; it dynamically delineates high-probability areas of gas accumulation or structural anomalies in the digital twin, which can be regarded as the risk shell that the current inspection system needs to focus on monitoring. Common construction behaviors pre-entered in the action library, such as excavation, welding, and mechanical vibration, and their corresponding physical action models, are quantified as excavation impact potential or energy, welding heat release rate or heat flow, and mechanical vibration amplitude or acceleration amplitude, respectively; the parameters are converted into the superposition of spatial disturbance vectors to simulate the interference of external operations on the pipeline and surrounding gas diffusion. The local expansion potential is the tendency of a coordinate point on the risk response envelope to deform or rupture under construction disturbance. The stress / mass transfer intensity of a point is obtained by taking the dot product of the envelope boundary coordinate entries and the disturbance vector corresponding to construction behaviors with a disturbance coefficient greater than zero. A higher value indicates that the point is more likely to become a leakage breakthrough or diffusion acceleration point. The tracking target points of the PTZ camera are the top three coordinate entries of the local expansion potential, converted back to their actual spatial positions, and used as the target points that the explosion-proof PTZ camera sequentially aligns with. The optical zoom is then triggered to its maximum focal length to obtain microscopic visual details of the target point, such as cracks, rust perforations, or the contact surface of construction machinery, thus achieving a closed loop of data analysis and visual confirmation.Visual details refer to high-resolution image information captured by the PTZ camera at maximum optical zoom, including but not limited to crack morphology on the pipe surface, corrosion pit edges, foreign object impact marks, or contact points of construction equipment; used for remote manual review or AI-assisted diagnosis to verify the authenticity of the digital twin judgment results. The emergency drill buffer is a dynamic simulation sandbox set up in the remote command terminal, using the current risk response envelope as the geometric boundary condition. When the disturbance coefficient exceeds the warning threshold, the corresponding disturbance vector is written into the buffer to quickly deduce the gas diffusion path after the envelope ruptures and calculate the maximum diffusion radius; the radius result is reverse-calibrated as an upper limit constraint for the dynamic weighting factor in the next inspection, forming an adaptive threshold adjustment mechanism; the preset threshold is used to determine whether the dynamic weighting factor of each coordinate item is sufficient to constitute a numerical threshold for risk; it can be derived from historical leak case data statistics or calibrated through a gas leak simulation model, and is a configurable parameter. Warning thresholds are used to determine whether the disturbance coefficient of a construction activity has reached the critical value that requires triggering an emergency drill warning. Warning thresholds are usually derived from historical accident data statistics, relevant safety regulations and standards, or gas leak simulation models. Preset thresholds are used to filter spatial coordinate points and construct risk envelopes. Warning thresholds are used to assess the hazard level of external construction activities. When the disturbance coefficient calculated for a certain construction activity exceeds this warning threshold, the activity is considered to have constituted a serious threat. The disturbance vector value corresponding to the activity will be written into the emergency drill buffer, the diffusion path simulation will be initiated, and an alarm may be sent to the remote command terminal.

[0053] The principles described in the above embodiments are referenced in the appendix. Figure 2This method involves collecting and processing multimodal data to generate a risk response envelope. Based on this envelope, target tracking and emergency plans are generated. This embodiment achieves accurate modeling of the three-dimensional risk field of gas pipelines through spatiotemporal alignment and fusion of multimodal sensor data. The mapping table based on cylindrical coordinates overcomes the limitations of traditional two-dimensional monitoring, organically coupling visual features with physical field parameters to form a spatially continuous risk characterization system. Specifically, the collaborative mapping of pixel grayscale gradients and gas concentration effectively identifies early leakage characteristics caused by microscopic deformation of the pipe wall. Furthermore, the innovative calculation mode of dynamic weighting factors significantly improves the sensitivity of risk warnings. The product operation of local temperature field and grayscale features amplifies weak signal features that are difficult to detect with traditional single-parameter monitoring. The dynamic evaluation mechanism based on construction disturbance vectors enables quantitative analysis of the impact of mechanical operations on pipeline risk status, providing a theoretical basis for proactive defense. The risk response envelope generation technology pioneers a new paradigm for non-contact monitoring. The curved boundary not only dynamically reflects the geometric characteristics of the current risk distribution, but its dot product with the disturbance vector also reveals the energy accumulation pattern of potential risk points. The prediction method based on field strength calculation detects risk points earlier than the traditional threshold alarm mechanism. The synergistic optimization of the intelligent tracking and emergency simulation system achieves effective integration of the monitoring-early warning framework. The automatic locking capability of the optical zoom system for high-risk targets greatly improves the efficiency of detailed data acquisition in emergency situations, while the diffusion path simulation based on the fluid dynamics model provides a scientific basis for emergency decision-making.

[0054] Example 2: Figure 3 As shown, based on Embodiment 1, the process of constructing a multimodal anchor data cluster in step S100 of this embodiment of the invention specifically includes the following steps:

[0055] Step S101: The grayscale gradient of each pixel in the video frame acquired by the explosion-proof dome camera is weighted and fused point by point with the spatial weight coefficient of the gas concentration sensor in the cylindrical coordinate system of the pipeline to obtain the grayscale-concentration association pair of each coordinate entry in the cylindrical coordinate mapping table; the grayscale-concentration association pair forms a semi-finished anchor point record containing only grayscale value and gas concentration value.

[0056] Among them, the spatial weight coefficient is a joint calculation of the radial projection distance of the sensor measuring point relative to the center line of the gas pipeline, the circumferential deflection angle, and the current gimbal attitude angle of the explosion-proof PTZ camera;

[0057] Step S102: After the instantaneous temperature values ​​with the same timestamp are forcibly aligned, the values ​​are assigned point by point to the corresponding coordinate entries in the semi-finished product anchor point record according to the precise spatial measurement point position of the temperature sensor in the cylindrical coordinate system. This allows each coordinate entry to simultaneously obtain the grayscale value, gas concentration value, and temperature value of the precise spatial measurement point position. The assigned coordinate entries form a complete multimodal anchor point record.

[0058] Among them, the precise spatial measurement point position is converted from the three-dimensional coordinates of the sensor dynamically calibrated by the laser ranging module built into the explosion-proof PTZ camera and the inclinometer into cylindrical coordinates;

[0059] Step S103: Record the index code, timestamp alignment flag, and sensor confidence label of the multimodal anchor points in the cylindrical mapping table, and combine and encapsulate them according to the predefined binary frame structure of the explosion-proof PTZ remote inspection digital twin to generate independent anchor point data blocks; all anchor point data blocks are spliced ​​together in the circumferential and axial order of the cylindrical grid to form a multimodal anchor point data cluster, which is stored in the inspection digital twin buffer of the explosion-proof PTZ camera.

[0060] In the above embodiments, after pixel-level weight calibration of the gray-level-concentration correlation pair in the cylindrical coordinate mapping table, the joint calculation of the spatial weight coefficients unifies the gimbal attitude, sensor radial projection, and circumferential deviation angle into the same coordinate frame. This ensures that every gray-level change on the imaging plane can be directly mapped to the corresponding cylindrical coordinate gas concentration field, eliminating spatial misalignment caused by parallax and sensor installation errors. The instantaneous temperature value, after dynamic calibration by a laser rangefinder-tilt meter, is converted to millimeter-level three-dimensional coordinates and injected into the semi-finished product anchor point record. The gray-level, concentration, and temperature data are locked in the same cylindrical grid, achieving spatiotemporal synchronization. The introduction of the temperature field compensates for the dependence of gas concentration on temperature, preventing a single concentration measurement from being misjudged as a leak due to ambient temperature drift, further improving the robustness of anomaly detection. Anchor data blocks encapsulate index codes, timestamps, and confidence labels through a predefined binary frame structure and align with the grid order of the digital twin; data clusters form a spatiotemporal tensor in the cache that can be directly accessed via a sliding window; the inspection algorithm does not require additional registration and can replay the grayscale-concentration-temperature joint field at any moment in the twin in real time, supporting millisecond-level leak location and backtracking analysis, while ensuring the verifiability of data traceability.

[0061] Example 3: Based on Example 2, the process of generating independent anchor point data blocks in step S103 of this embodiment of the invention specifically includes the following steps:

[0062] Step S1031: The index code, timestamp alignment flag, sensor confidence label, gray value, gas concentration value, and temperature value in the multimodal anchor point record are arranged bit-level according to the arrangement order and bit width of each field in the predefined binary frame structure of the explosion-proof PTZ remote inspection digital twin to obtain the unencapsulated raw binary data segment; the payload to be framed is formed from the raw binary data segment;

[0063] Step S1032: The payload to be framed is sequentially spliced ​​with the frame start flag, payload length indicator bit and cyclic redundancy check code of the predefined binary frame structure, and the complete binary frame body is obtained after field encapsulation processing; the binary frame body is an encapsulated frame with frame boundary and check protection.

[0064] Step S1033: The encapsulated frame is stored as an independent anchor point data block in the inspection digital twin buffer of the explosion-proof PTZ camera after a write operation, and is appended to the tail of the multimodal anchor point data cluster in the circumferential and axial order of the cylindrical grid.

[0065] In the above embodiments, the predefined field width and arrangement order of the binary frame structure ensure that the grayscale, concentration, temperature, and confidence labels have unique and constant bit offsets within the payload. This guarantees that the subsequent parser of the digital twin can directly extract the required channels bit by bit without dynamically inferring field boundaries, eliminating parsing ambiguity and reducing computational overhead. The frame start flag and payload length indicator define the physical boundaries of the data blocks, and the cyclic redundancy check code performs integrity verification at the frame level. Any bit flips, truncation, or misalignments during transmission or storage can be detected in real time, preventing corrupted data from entering the buffer area and maintaining zero pollution in the internal data chain of the twin. The write operation appends the anchor point data blocks to the tail of the multimodal anchor point data cluster according to the circumferential and axial order of the cylindrical grid, forming a spatiotemporal index isomorphic to the pipeline geometry. The subsequent inference algorithm can directly perform local neighborhood retrieval and sliding window batch processing according to the grid order without additional index reconstruction, achieving geometric consistency maintenance and efficient random access in the inspection buffer area.

[0066] Example 4: Figure 4 As shown, based on Example 1, the process of calculating the rate of change of the gas concentration value of each coordinate entry before and after the disturbance vector in step S200 of this embodiment of the invention specifically includes the following steps:

[0067] Step S201: The superimposed spatial disturbance vector is decomposed into vector components according to the spatial coordinates of each coordinate entry in the cylindrical mapping table, and applied to the original gas concentration value of the coordinate entry. After processing by the concentration response function based on the digital twin, the predicted gas concentration value of the coordinate entry after the disturbance is obtained.

[0068] Step S202: Subtract the original concentration value before the disturbance from the predicted concentration value after the disturbance of each coordinate entry to obtain the absolute amount of concentration change. Then divide it by the original concentration value before the disturbance. After processing by the predefined dynamic correction coefficient introduced into the explosion-proof PTZ camera inspection digital twin, the gas concentration change rate of the coordinate entry is obtained, forming the local change rate data of a single coordinate entry.

[0069] Step S203: Local rate of change data are sequentially written into the rate of change vector of the inspection digital twin buffer according to the circumferential and axial order of the cylindrical grid. After vector integrity verification and timestamp alignment flag binding, a standardized rate of change sequence that can participate in the arithmetic mean calculation is output.

[0070] In the above embodiments, in the calculation process of the rate of change before and after the disturbance vector, the spatial disturbance vector is first decomposed into components isomorphic to the cylindrical coordinate system, and then mapped to the gas concentration field through the concentration response function built into the digital twin, so that any external construction disturbance can reflect the corresponding physical coupling effect in the concentration prediction value. Subsequently, after the absolute amount of concentration change is normalized, a dynamic correction coefficient is introduced. This coefficient is adjusted in real time according to the real-time calibration status of the explosion-proof PTZ camera and the sensor confidence level, ensuring that the rate of change reflects the real physical fluctuations while eliminating artifacts caused by sensor drift or environmental noise. Finally, the local rate of change is serialized into a standardized vector and bound with a timestamp and integrity check bit, forming a rate of change sequence that can directly participate in the arithmetic mean calculation and has verifiability and traceability, providing zero-drift and zero-redundancy input for the aggregation calculation of the disturbance coefficient.

[0071] Example 5: Based on Example 4, the process of obtaining the predicted gas concentration value of the coordinate entry after the disturbance in step S201 of this embodiment of the invention specifically includes the following steps:

[0072] Step S2011: The radial, circumferential and axial components of the superimposed spatial disturbance vector after decomposition according to cylindrical coordinates are input together with the original gas concentration value of the coordinate entry into the linear transfer module of the predefined concentration response function of the digital twin. After vector-concentration coupling operation processing, the initial concentration response value without nonlinear correction is obtained.

[0073] Step S2012: Multiply the initial concentration response value by the gray value corresponding to the coordinate entry and the temperature value after local mean filtering, and then divide by the square of the spatial weight coefficient of the coordinate entry in the cylindrical mapping table. After processing by the nonlinear gain element in the concentration response function, the dynamic response value after gray-temperature weighted correction is obtained; the corrected concentration change estimate is formed from this dynamic response value.

[0074] Step S2013: Multiply the dynamic response value by the diffusion direction attenuation factor calculated from the current gimbal attitude angle of the explosion-proof PTZ camera, and add the background fluctuation component retained in the original gas concentration value due to forced alignment of the timestamp. After output limiting and smoothing filtering in the concentration response function, the predicted gas concentration value of the coordinate entry after the disturbance is obtained.

[0075] The nonlinear gain processing involves multiplying the initial concentration response value by the grayscale value of the coordinate entry and the temperature value after local mean filtering, then dividing by the square of the spatial weight coefficient of the coordinate entry in the cylindrical mapping table. A nonlinear mapping function based on the local curvature of the pipe and the sensor confidence level is introduced into the digital twin to amplify or compress the multiplication and division results. Specific technical features include:

[0076] Exponential weighting of the gray-temperature product: The product of the gray value reflecting the degree of micro-deformation of the pipe wall and the temperature value reflecting the local thermal anomaly is fed into an exponential function with the natural constant as the base. The exponential term is divided by the square of the radial projection distance in cylindrical coordinates to achieve a nonlinear response with high gain in the near-distance high-confidence region and low gain in the far-distance low-confidence region; The reciprocal of the square of the spatial weight coefficient is used as the gain basis: When the spatial weight coefficient calculated by the radial projection distance, circumferential deflection angle and gimbal attitude angle is less than the preset lower limit, the gain increases sharply to highlight weak signals; when the coefficient is greater than the upper limit, the gain is compressed to close to 1 to avoid excessive amplification of high-weight regions. The piecewise gain curve is integrated with the explosion-proof PTZ camera inspection scenario: A piecewise nonlinear curve is fitted beforehand through gas pipeline leak simulation and on-site calibration experiments. For example, the slope in the small signal region is >2, and the slope in the large signal region is <0.5. This significantly enhances the dynamic response value in the early stages of low-concentration leaks, maintains stability in known high-concentration areas, improves the sensitivity of early warning, and suppresses saturation distortion. The output is a dynamic response value corrected by grayscale-temperature weighting. Amplitude limiting and smoothing filtering are performed by multiplying the dynamic response value by the diffusion direction attenuation factor and adding the background fluctuation component, followed by amplitude limiting and sliding window smoothing to eliminate drastic jumps in predicted values ​​caused by sensor noise and transient disturbances, and to ensure the physical reasonableness of the concentration prediction value. Specific technical features include: Amplitude limiting processing: Based on the pressure resistance level of the current inspected pipe section and the maximum rate of increase of historical leakage data, a dynamic upper limit threshold is set, such as 3 times the current original concentration value, and a lower limit threshold, 0, or the lower limit of background fluctuation; any intermediate values ​​exceeding the range are forcibly clamped to the boundary value to prevent false high / low predictions due to numerical calculation overflow or sensor malfunction. Smoothing and Filtering: A weighted sliding window of length 3 or 5 is used to arithmetically average the amplitude-limited values ​​of the same coordinate item over multiple consecutive timestamps, such as three consecutive inspection cycles. The window length is correlated with the pan-tilt-zoom (PTZ) scanning step cycle of the explosion-proof PTZ camera, ensuring the filtering time constant is shorter than the gas diffusion characteristic time. This preserves the rapid responsiveness to concentration changes while filtering out high-frequency noise. Adaptive Binding of Background Fluctuation Components: Before smoothing, the amplitude-limited value is subtracted from the background fluctuation component (the concentration background value of the coordinate item under undisturbed conditions) retained due to forced timestamp alignment. Smoothing is then performed, and finally, the background component is added back. This prevents the background baseline from drifting during smoothing, ensuring the predicted value accurately reflects the net change caused by disturbances. The final gas concentration prediction is limited to a predefined physically reasonable range (e.g., 0-100% LEL) within the explosion-proof PTZ camera inspection digital twin and is simultaneously written to the confidence flag in the buffer. Amplification and smoothing filtering ensure that the gas concentration prediction has both rapid response to sudden disturbances and meets the digital twin's requirements for data stability and reliability.

[0077] In the above embodiments, this embodiment establishes a baseline proportional relationship between disturbance and concentration response, ensuring that external disturbances are mathematically mapped to quantifiable concentration increments. Through joint calculations of grayscale values, temperature values, and spatial weighting coefficients, optical texture, thermodynamic diffusion coefficients, and geometric weights are introduced into the gain model to complete nonlinear correction, filtering out spurious sensitization caused by illumination shifts and thermal inertia differences. This ensures that the dynamic response value retains the true concentration fluctuations caused by the disturbance while suppressing the coupling of environmental and sensor errors. The observation geometry is introduced into the concentration field model to compensate for signal attenuation caused by changes in viewing angle and distance. After superimposing the retained background fluctuation components, the result is further subjected to amplitude limiting and smoothing filtering to output a predicted gas concentration value with spatial continuity, temporal consistency, and controlled amplitude.

[0078] Example 6: Based on Example 5, the process of vector-concentration coupling operation in step S2011 provided in this embodiment of the invention specifically includes the following steps:

[0079] Step S20111: The radial, circumferential and axial components of the superimposed spatial disturbance vector after decomposition according to cylindrical coordinates are multiplied by the predefined radial coupling coefficient, circumferential coupling coefficient and axial coupling coefficient in the explosion-proof PTZ camera inspection digital twin, respectively. After linear weighted summation, the total disturbance intensity after coupling in the three directions is obtained.

[0080] Step S20112: Multiply the pure disturbance excitation amount by the original gas concentration value of the coordinate entry, and then divide by the square root of the spatial weight coefficient of the coordinate entry in the cylindrical mapping table. After the multiplication-scaling operation in the linear transfer module, the disturbance contribution value modulated by the original concentration is obtained.

[0081] Step S20113: The disturbance contribution value modulated by the original concentration is added to the original gas concentration value itself multiplied by a reference retention factor. The current pan-tilt attitude angle of the explosion-proof PTZ camera and the radius of curvature of the pipeline are jointly calculated. After addition and output buffering in the linear transmission module, the initial concentration response value without nonlinear correction is obtained.

[0082] In the above embodiments, this embodiment projects the radial, circumferential, and axial components onto coupling coefficients corresponding to the physical field of the pipeline, achieving direction-amplitude coupling. This ensures that the weights of disturbances in different spatial directions on the concentration field strictly correspond to their engineering significance, avoiding coupling distortion caused by directional confusion. The current concentration level is introduced as a scaling factor through original concentration modulation, allowing the disturbance contribution to adaptively amplify or attenuate with the true concentration gradient. This ensures compatibility with high and low concentration conditions within the same model framework and suppresses excessive amplification errors in low-concentration regions. The square root scaling of the spatial weight coefficients offsets differences in grid sampling density, ensuring a uniform response scale to disturbances at different cylindrical positions. The baseline retention factor preserves the original concentration as a zero-disturbance baseline. The joint solution of the gimbal attitude angle and pipeline curvature radius incorporates both the observation geometry and pipeline geometry into the linear accumulation term, compensating for signal attenuation or amplification caused by viewing angle and curvature. This ensures that the final output initial concentration response value possesses complete physical characteristics—directional resolution, concentration adaptation, geometric consistency, and observation compensability—before entering the nonlinear stage.

[0083] Example 7: Figure 5 As shown, based on Example 1, the process of deducing the gas diffusion path after envelope rupture according to the written disturbance vector value in step S300 of this embodiment of the invention specifically includes the following steps:

[0084] Step S301: Write the disturbance vector value into the emergency drill buffer, and calculate the fracture offset point by point according to the gray gradient direction of each coordinate entry in the cylindrical coordinate mapping table; then fuse the fracture offset with the instantaneous temperature value of the coordinate entry after time stamping, and calibrate it with the observation tilt angle calculated by the current gimbal attitude angle of the explosion-proof PTZ camera. After the cylindrical mesh remapping processing of the fracture offset, the fracture initiation coordinate cluster is obtained; the fracture initiation coordinate cluster constitutes the initial fracture point set starting from the risk response envelope surface;

[0085] Step S302: Each point in the initial set of rupture points moves along the circumference and axial direction of the cylindrical coordinate mapping table, according to the moving step size constrained by the temperature value after local mean filtering and the gas concentration value at the point, and jumps to the adjacent coordinate entries in sequence. After each jump, the correlation strength between the gray value of the coordinate entry and the background mean is used as the jump retention probability. When the jump retention probability is lower than the diffusion termination threshold, the jump stops. The coordinate entries passed through in each jump, the current gas concentration value of the rupture point, and the jump retention probability are stored in an array in the order of the jumps to form the discretized propagation trajectory of the rupture point. The propagation trajectory forms a coordinate sequence of each rupture point carrying a concentration label.

[0086] Step S303: In the discretized propagation trajectories of all rupture points, coordinate entries whose radial distance from the gas pipeline centerline exceeds the maximum radial distance of the current risk response envelope are concatenated into a spatial polyline in the order of forced alignment of the explosion-proof PTZ camera timestamps; and the coordinate entries corresponding to the maximum concentration change rate recorded in the last inspection by the explosion-proof PTZ camera are embedded into the polyline as trajectory anchor points. After binding the trajectory lines and anchor points and matching them with the boundary conditions of the emergency drill buffer zone, the gas diffusion path after the envelope rupture is obtained, and the farthest radial distance from the end of the path to the pipeline centerline is taken as the maximum diffusion radius; the gas diffusion path and the maximum diffusion radius are directly written into the simulation results of the emergency drill buffer zone display layer of the remote command terminal.

[0087] The diffusion termination threshold is a probability threshold used to determine whether the transition process continues. The calculated transition retention probability is compared to this threshold; if it falls below this threshold, the simulated diffusion process stops at that point and no longer continues outward. This acts as a stopping condition, defining the credible boundary of gas diffusion under the current physical and environmental conditions. The trajectory line is a spatial polyline formed by connecting the coordinates of all rupture points whose radial distance exceeds the maximum radial distance of the current risk response envelope, arranged in timestamp order. This polyline represents the preliminary outline of the possible diffusion path after the gas breaks through the risk envelope. Anchor points specifically refer to the coordinate entry with the largest concentration change rate in the last inspection record from the explosion-proof PTZ camera. These high-confidence historical risk feature points are embedded into the trajectory line. The embedding process involves directional consistency comparison, replacing points on the trajectory line with matching directions with anchor points and their corresponding maximum concentration change rate labels. Anchor points are key data for correcting and enhancing the trajectory line. By binding, the most historically anomalous information is injected into the currently simulated path, making the final diffusion path more credible and valuable for reference.

[0088] In the above embodiments, this embodiment couples the written disturbance vector with the gray-level gradient direction, introduces instantaneous temperature and gimbal attitude angle observation parameters, and completes the time-space-observation triple calibration of the rupture offset. This ensures that the rupture initiation coordinate cluster has a topological position in the cylindrical grid that is consistent with the geometry of the on-site observation, eliminating the initiation drift caused by viewpoint or time misalignment. The thermodynamic diffusion capability is bound to the local concentration gradient using a movement step size jointly constrained by temperature and concentration. The transition retention probability is used to suppress low-confidence paths in real time through gray-level-background correlation strength, ensuring that the propagation trajectory extends only in the reliable region and avoiding false extensions. Serialized records ensure that each trajectory carries a concentration label. Coordinate entries exceeding the radial boundary of the risk response envelope are concatenated in time stamp order to form a spatial polyline. The coordinate entries corresponding to the historical maximum concentration change rate are used as trajectory anchors embedded in the polyline to couple historical anomalies with the current calculation. The diffusion path and maximum diffusion radius output after the trajectory matches the boundary conditions are directly mapped to executable emergency information in the visualization layer for immediate use by the remote command terminal.

[0089] Example 8: Based on Example 7, the process of binding the trajectory line to the anchor point and matching it with the boundary conditions of the emergency drill buffer zone in step S303 of this embodiment of the invention specifically includes the following steps:

[0090] Step S3031: The gray-scale gradient direction of each coordinate entry in the embedded polyline of the trajectory anchor points is compared with the gray-scale gradient direction of the trajectory line and the anchor points. The coordinate entries with an angle between the two less than the minimum step angle of the explosion-proof PTZ camera are taken as binding nodes. After replacing the binding nodes with the coordinate values ​​of the trajectory anchor points and their maximum concentration change rate labels in the order of timestamps, a corrected polyline with historical risk characteristics indicating the initial diffusion path is obtained after the anchor points are embedded.

[0091] Step S3032: The predicted gas concentration value of each coordinate entry on the initial diffusion path is compared with the historical concentration benchmark value of the coordinate entries pre-stored in the emergency drill buffer. The coordinate entries whose absolute value of the difference exceeds the concentration fluctuation tolerance corresponding to the disturbance vector magnitude are taken as valid path points. After removing invalid path points in the cylindrical circumferential order and reconnecting the remaining valid path points, a filtered path matching the boundary conditions of the emergency drill buffer is obtained. The filtered path forms a diffusion path skeleton that conforms to the geometric constraints of the current risk response envelope.

[0092] Step S3033: The radial distance of each coordinate entry in the diffusion path skeleton is compared with the predefined pipeline safety radius in the explosion-proof PTZ camera inspection digital twin. The coordinate entry whose radial distance is first less than the pipeline safety radius is taken as the path endpoint. After processing the output of all coordinate entries from the start point to the end point of the skeleton in the order of forced timestamp alignment, the gas diffusion path after envelope rupture is obtained.

[0093] In the above embodiments, this embodiment utilizes grayscale gradient direction consistency comparison to embed the historical maximum concentration change rate label carried by the trajectory anchor point into the current spatial polyline, forcing alignment with historical anomalies at both the geometric topology and concentration feature levels, thereby improving the credibility of the diffusion path and reducing the probability of false extensions. Through dual verification using difference calculation and concentration fluctuation tolerance threshold, coordinate entries that do not match the historical benchmark of the buffer zone or cannot be explained by the current disturbance are eliminated, ensuring that the path maintains a strict match with the boundary conditions of the emergency drill buffer zone, forming a geometrically continuous and physically interpretable diffusion path skeleton. Using the pipeline safety radius as the termination criterion, the skeleton is truncated before the safety radius is first met, ensuring that the output path completely covers the risk area in an engineering sense without excessive extrapolation; combined with forced alignment of the output using timestamps, the final diffusion path has traceable spatiotemporal consistency.

[0094] Example 9: Based on Example 8, the process of forming a diffusion path skeleton that conforms to the current risk response envelope geometric constraints in step S3032 of this embodiment of the invention specifically includes the following steps:

[0095] Step S30321: The predicted gas concentration value of each coordinate entry on the initial diffusion path is compared with the historical concentration benchmark value of the coordinate entry stored in the emergency drill buffer point by point, and the absolute value of the difference is taken. The absolute value is combined with the disturbance vector magnitude and the diffusion direction attenuation factor calculated by the current gimbal attitude angle of the explosion-proof PTZ camera to determine whether the point is retained. If the absolute value of the difference is greater than the product of the disturbance vector magnitude and the attenuation factor, the point is marked as a high fluctuation point; otherwise, it is marked as a low fluctuation point. The concentration fluctuation mark pairs of each coordinate entry are obtained, and the concentration fluctuation mark pairs form a binary state set for determining whether the point is retained.

[0096] Step S30322: High fluctuation points in the concentration fluctuation marker pairs of each coordinate entry are taken as candidate valid points, and then the direction is filtered according to the angle between the gray gradient direction corresponding to the point and the normal of the risk response envelope boundary: candidate points with an angle smaller than the minimum step angle of the explosion-proof PTZ camera are retained, and candidate points with an angle greater than or equal to the minimum step angle are removed. After traversing all candidate points in the order of cylindrical grid and recording the retained points, the effective path point set is obtained; the effective path point set constitutes the sparse coordinate sequence after removing interference points;

[0097] Step S30323: The points in the sparse coordinate sequence are arranged in ascending order of their circumferential angle in the cylindrical mapping table, and then sorted again by axial distance. Adjacent points are connected sequentially to form a polyline. During the connection, the cylindrical grid distance between adjacent points is checked to see if it exceeds the spatial resolution threshold corresponding to the optical zoom magnification of the explosion-proof PTZ camera. If it exceeds the spatial resolution threshold corresponding to the optical zoom magnification of the explosion-proof PTZ camera, a new coordinate point is inserted at the midpoint of the line connecting adjacent points, with the smaller value of the gas concentration prediction between the two points as the concentration label. After connection and interpolation point supplementation, a filtered path matching the boundary conditions of the emergency drill buffer zone is obtained. The filtered path forms a diffusion path skeleton that conforms to the geometric constraints of the current risk response envelope.

[0098] Among them, the historical concentration benchmark value is the historical normal value or baseline value of gas concentration for each coordinate entry, which is pre-stored in the emergency drill buffer. It is obtained through long-term statistical learning under normal operating conditions without leakage or construction disturbance. In steps S3032 and S30321, the difference between the predicted gas concentration value on the current simulation path and this historical concentration benchmark value is calculated. The absolute value of the difference, combined with other factors such as the disturbance vector magnitude and attenuation factor, is used to determine whether the point is a valid path point, that is, a point with significant concentration fluctuation, thereby eliminating invalid points caused by noise or model error and ensuring the physical authenticity of the simulation path. The disturbance vector magnitude is the size or intensity of the spatial disturbance vector, a scalar value. The spatial disturbance vector is a three-dimensional vector (with direction and magnitude), comprehensively reflecting the type and intensity of the construction activity, such as excavation or welding. The concentration fluctuation tolerance is the product of the disturbance vector magnitude and the diffusion direction attenuation factor; it is a dynamic threshold representing how much change in the gas concentration at a coordinate point can be reasonably explained under the current construction disturbance intensity and observation direction. Exceeding this tolerance is considered a truly noteworthy abnormal fluctuation. The diffusion direction attenuation factor is a coefficient calculated from the current pan-tilt attitude angle of the explosion-proof PTZ camera, typically ranging from 0 to 1. The observation direction (attitude angle) of the PTZ camera affects the signal strength acquired by the sensor. For example, the signal is strong when the sensor is directly facing the leak source, but attenuates when it is angled or facing away from the source. The factor is used to quantify the degree of signal attenuation caused by the observation geometry. In the above embodiments, this embodiment uses a threshold determination by combining the concentration prediction difference with the perturbation vector magnitude and the diffusion direction attenuation factor derived from the gimbal attitude angle. This allows the labeling results to simultaneously reflect the impact of perturbation intensity and observation direction on concentration confidence, generating a binary state set that can be directly used for screening. This state set provides a unique and reproducible decision basis for eliminating low-confidence coordinates. An angle constraint between the gray-scale gradient direction and the envelope boundary normal is superimposed on high-fluctuation points to eliminate points that only satisfy concentration fluctuation but whose direction deviates from the risk response boundary, ensuring that the retained points have both significant concentration changes and conform to the envelope geometry. The retained points are recorded sequentially by traversing the cylindrical grid, forming a sparse coordinate sequence, achieving data dimensionality reduction and noise suppression while maintaining topological continuity. The monotonicity and path uniqueness of the polyline are ensured by using circumferential-axial dual-order sorting. By detecting the spatial resolution threshold corresponding to optical zoom and inserting concentration interpolation points, the spatial resolution of the final polyline is matched with the observation system capability, avoiding geometric gaps in the path. The resulting diffusion path skeleton satisfies both the buffer boundary conditions and has geometric constraints consistent with the risk response envelope, and can be directly used for maximum diffusion radius calculation and visualization.

[0099] Example 10: Based on Example 9, the process of constructing a sparse coordinate sequence after removing interference points from the effective path point set in step S30322 of this embodiment of the invention specifically includes the following steps:

[0100] Step S303221: The gray-level gradient direction vector corresponding to each candidate valid point and the normal vector of the risk response envelope boundary at the point are respectively input into the direction angle calculation in the digital twin of the explosion-proof PTZ camera inspection. By querying the pre-stored gray-level gradient-normal angle mapping table, which is generated by the calibration experiment of the explosion-proof PTZ camera on the surface texture of the pipeline under different gimbal attitude angles, the direction angle value of each candidate valid point is obtained.

[0101] Step S303222: Real-time reading of the minimum step angle of the gimbal at the current optical zoom level of the explosion-proof PTZ camera, i.e., the minimum physical rotation angle of the gimbal stepper motor; comparing the directional angle value of each candidate valid point with the minimum step angle point by point: if the angle value is less than the minimum step angle, it is marked as a trackable point; if the angle value is greater than or equal to the minimum step angle, it is marked as an untrackable point; and writing the trackable points into the valid point queue of the inspection digital twin buffer according to the circumferential scanning order of the cylindrical grid, to obtain the initial set of valid points that form the intermediate point sequence after filtering by gimbal tracking capability.

[0102] Step S303223: The cylindrical grid distance between two adjacent points in the intermediate point sequence is compared with the field of view coverage of the explosion-proof PTZ camera at maximum optical zoom. If the distance exceeds the field of view coverage, a transition point generated by the preset position interpolation algorithm of the explosion-proof PTZ camera is inserted between the two points according to the trend of gray-level gradient direction. The coordinates of the transition point are determined by the weighted sum of the coordinates of the two points and their respective gray-level gradient direction vectors. After traversing all adjacent points and completing the insertion process, an effective path point set that meets the continuous tracking requirements of the PTZ camera is obtained. The effective path point set constitutes a sparse coordinate sequence that eliminates interference points and ensures the reachability of the PTZ camera.

[0103] The grayscale gradient-normal angle mapping table is a pre-stored lookup table that records the correspondence between the grayscale gradient direction of a point on the pipe surface and the angle between the normal direction of the point risk response envelope boundary under different pan-tilt angles of the explosion-proof PTZ camera. Generation method: It is obtained through prior calibration experiments on the pipe surface texture; for example, the PTZ camera is used to photograph the standard pipe texture at different angles, and the mathematical relationship between the texture direction or grayscale gradient and the geometric normal angle is analyzed and recorded. The mapping table allows the system to quickly and accurately deduce the geometric normal direction or its angle, which is difficult to measure directly, from the observable grayscale gradient direction; it provides reliable basic data for direction selection and realizes the coupling of optical and geometric features. The inspection digital twin cache area is located in a dedicated memory area on the explosion-proof PTZ camera or its edge computing device; it is used to temporarily store intermediate and final data required to build the digital twin; specifically, it includes: storing multimodal anchor point data clusters, stored in cylindrical grid order to form a directly callable spatiotemporal tensor; storing a normalized rate of change sequence, storing the calculated local concentration change rate; and storing a valid point queue, storing the intermediate point sequence after being filtered by the PTZ tracking capability. The intermediate point sequence is the ordered sequence of points obtained by writing the trackable points in the initial set of valid points into the cache queue in the circumferential scanning order of the cylindrical grid; it represents the points that have passed the dual screening of concentration fluctuation and directional angle, and are theoretically points that the PTZ camera can stably track because the angle is less than the minimum step angle; it still needs to go through the next step of field of view coverage check (step S303223), and may add transition points through interpolation, before it can finally become a valid path point set that meets the requirements of continuous PTZ tracking, i.e., a sparse coordinate sequence; it is a key intermediate step in generating the final tracking trajectory.

[0104] In the above embodiments, this embodiment directly converts the experimental calibration results into angle thresholds through a gray-level gradient-normal mapping table, coupling the texture observation capability of the explosion-proof PTZ camera with the geometric normal constraint of the risk response envelope, ensuring that all screening is based on the same set of physically measurable indicators; point-by-point judgment is performed using the minimum step angle as a hard threshold, eliminating all points that cannot be stably tracked by the PTZ, so that the effective point queue can be locked by the PTZ camera in real time at the physical control level, avoiding command lag or loss of tracking; when the distance between adjacent points exceeds the maximum optical zoom field of view, a preset position interpolation algorithm is used to insert transition points along the gray-level gradient direction, so that the geometric resolution of the entire polyline is always lower than the limit of the PTZ camera's field of view, ensuring that the target does not leave the frame during continuous zoom; the final output sparse coordinate sequence not only eliminates interference points that deviate from the envelope or have insufficient tracking capability, but also forms a trajectory that can be smoothly transitioned by the PTZ camera between any adjacent nodes, meeting the combined requirements of digital twins for real-time visualization, continuous tracking, and geometric consistency.

[0105] Example 11: As Figure 6As shown, based on Embodiments 1-10, the remote inspection system for gas explosion-proof PTZ cameras based on digital twins provided in this embodiment of the invention includes:

[0106] The weighting coefficient fusion module 1 is used to merge the video frames, instantaneous gas concentration values, and instantaneous temperature values ​​collected by the explosion-proof PTZ camera with the spatial weighting coefficients of the gas concentration measurement points after the video frames, gas concentration instantaneous values, and temperature instantaneous values ​​are forcibly aligned with the timestamps. This forms a cylindrical coordinate mapping table with the center line of the gas pipeline as the reference. Each coordinate entry in the cylindrical coordinate mapping table simultaneously records the gray value, gas concentration value, and temperature value of the location, forming a multimodal anchor point data cluster.

[0107] The disturbance coefficient acquisition module 2 is used to extract all coordinate entries in the multimodal anchor point data cluster whose gas concentration values ​​are higher than the background mean. After local mean filtering of the temperature values ​​of the coordinate entries, the module multiplies them with the gray values ​​of the corresponding coordinates to obtain the dynamic weight factor of each coordinate. All coordinate entries with dynamic weight factors greater than a preset threshold are connected into a non-closed surface, which is defined as the risk response envelope. At the same time, the pre-entered construction behavior action library, including excavation impact, welding heat release rate, and mechanical vibration amplitude, is converted into a superposition of spatial disturbance vectors. The superposition vector is applied to each coordinate entry in the multimodal anchor point data cluster, and the rate of change of the gas concentration value of each coordinate entry before and after the disturbance vector is applied is calculated. The arithmetic mean of all the rates of change is taken as the disturbance coefficient of the construction behavior.

[0108] The diffusion path deduction module 3 is used to perform a dot product between the boundary coordinate entries of the obtained risk response envelope and the spatial disturbance vector corresponding to the construction behavior with a disturbance coefficient greater than zero, to obtain the local expansion potential of each point on the envelope surface. The three coordinate entries with the highest local expansion potential are used as the tracking target points of the PTZ camera, and the optical zoom of the explosion-proof PTZ camera is triggered to the maximum focal length to capture the visual details of the target points in sequence. At the same time, the disturbance vector value corresponding to the construction behavior with a disturbance coefficient exceeding the warning threshold is written into the emergency exercise buffer of the remote command terminal. The buffer uses the risk response envelope as the boundary condition, and deduces the gas diffusion path after the envelope ruptures based on the written disturbance vector value. The maximum diffusion radius obtained from the deduction is back-calibrated as the upper bound of the dynamic weight factor in the next inspection.

[0109] In the above embodiments, this embodiment realizes a physical sensing array composed of an explosion-proof PTZ camera, a gas concentration sensor, and a temperature sensor, which is completely mapped to a real-time volume data field within a digital twin. The weighting coefficient fusion module locks the gray-level gradient, instantaneous concentration value, and temperature field into the same cylindrical coordinate system, so that each pixel carries three-dimensional coupled information of concentration and temperature. The disturbance coefficient acquisition module uses areas with high concentration as seeds, generates dynamic weighting factors through temperature filtering and gray-level weighting, and constructs a non-closed risk response envelope; the envelope defines the geometric boundary of the potential leakage area and also retains local intensity differences; after the construction behavior action library is converted into a spatial disturbance vector, it is superimposed on the same coordinate grid, and the concentration change rate is calculated in real time and the disturbance coefficient is output; the instantaneous impact intensity of external operations on the pipeline safety status is quantified, providing quantifiable indicators for risk classification. The diffusion path deduction module performs a dot product between the envelope boundary and the disturbance vector to obtain the local expansion potential. The coordinate entries with the highest expansion potential are locked as gimbal tracking targets, triggering optical zoom to capture high-resolution details and enabling proactive acquisition of event-level visual evidence. At the same time, vectors with disturbance coefficients higher than the warning threshold are written into the emergency drill buffer. Using the risk response envelope as the initial boundary, the gas diffusion path after the envelope breaks is deduced. The deduction results inversely constrain the upper bound of the dynamic weight factor for the next round of inspection, forming a closed-loop parameter tuning mechanism, so that the system's response threshold to construction disturbances adaptively converges with the evolution of the scenario.

[0110] Figure 7 A block diagram of an exemplary electronic device suitable for implementing embodiments of the present invention is shown.

[0111] The electronic device may include a central processing unit / microprocessor / main control chip 4; and a storage medium 5 coupled to the central processing unit / microprocessor / main control chip 4 and storing computer-executable instructions therein for performing the steps of various methods of embodiments of the present invention when executed by the processor.

[0112] The central processing unit / microprocessor / main control chip 4 may include, but is not limited to, one or more processors or microprocessors.

[0113] Storage medium 5 may include, but is not limited to, random access memory (RAM), read-only memory (ROM), flash memory, EPROM memory, EEPROM memory, registers, computer storage media (e.g., hard disk, floppy disk, solid-state drive, removable disk, CD-ROM, DVD-ROM, Blu-ray disc, etc.).

[0114] In addition, the electronic device may include (but is not limited to) a data bus 6, an input / output bus / external bus / device bus 7, a display 8, and input / output devices 9 (e.g., keyboard, mouse, speaker, etc.).

[0115] The central processing unit / microprocessor / main control chip 4 can communicate with the display 8 and input / output devices 9 via wired or wireless networks (not shown) through the input / output bus / external bus / device bus 7.

[0116] The storage medium 5 may also store at least one computer-executable instruction for performing the steps of various functions and / or methods in the embodiments described herein when the central processing unit / microprocessor / main control chip 4 is running.

[0117] In one embodiment, the at least one computer-executable instruction may also be compiled into or comprise a software product, wherein one or more computer-executable instructions are executed by a processor to perform the steps of the various functions and / or methods in the embodiments described herein.

[0118] Figure 8 A schematic diagram of a computer-readable storage medium according to an embodiment of the present invention is shown.

[0119] like Figure 8 As shown, the non-transitory computer-readable storage medium 11 stores instructions, such as computer-readable instructions 10. When the computer-readable instructions 10 are executed by a processor, the various methods described above can be performed. The non-transitory computer-readable storage medium includes, but is not limited to, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory. Non-transitory non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. For example, the non-transitory computer-readable storage medium 11 can be connected to a computing device such as a computer, and then, when the computing device executes the computer-readable instructions 10 stored on the non-transitory computer-readable storage medium 11, the various methods described above can be performed.

[0120] In the embodiments provided by this invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between systems or units may be electrical, mechanical, or other forms.

[0121] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0122] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0123] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for executing all or part of the steps of the methods of the various embodiments of this invention through a computer device (which may be a personal computer, server, or network device, etc.). The aforementioned storage medium includes: USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, optical disks, and other media capable of storing program code.

[0124] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A remote inspection method for gas explosion-proof PTZ cameras based on digital twins, characterized in that, Includes the following steps: The video frames, instantaneous gas concentration values, and instantaneous temperature values ​​collected by the explosion-proof PTZ camera are forcibly aligned with timestamps to form a cylindrical coordinate mapping table based on the centerline of the gas pipeline; thus forming a multimodal anchor point data cluster. Extract all coordinate entries in the multimodal anchor point data cluster whose gas concentration values ​​are higher than the background mean. After local mean filtering of the temperature values ​​of the coordinate entries, multiply them by the gray values ​​of the corresponding coordinates to obtain the dynamic weight factor of each coordinate. Connect all coordinate entries whose dynamic weight factors are greater than a preset threshold into a non-closed surface, which is defined as the risk response envelope. At the same time, convert the pre-entered construction behavior action library, including excavation impact, welding heat release rate, and mechanical vibration amplitude, into a superposition of spatial disturbance vectors. Apply the superposition vector to each coordinate entry in the multimodal anchor point data cluster, calculate the rate of change of the gas concentration value of each coordinate entry before and after the disturbance vector is applied, and take the arithmetic mean of all the rates of change as the disturbance coefficient of the construction behavior. The boundary coordinate entries of the obtained risk response envelope are multiplied by the spatial disturbance vector corresponding to the construction behavior with a disturbance coefficient greater than zero to obtain the local expansion potential of each point on the envelope surface. The three coordinate entries with the highest local expansion potential are used as the tracking target points of the PTZ camera. At the same time, the disturbance vector values ​​corresponding to the construction behavior with a disturbance coefficient exceeding the warning threshold are written into the emergency exercise buffer of the remote command terminal. Using the risk response envelope as the boundary condition, the gas diffusion path after the envelope rupture is deduced based on the written disturbance vector values, and the maximum diffusion radius obtained from the deduction is back-calibrated as the upper bound of the dynamic weighting factor in the next inspection.

2. The remote inspection method for gas explosion-proof PTZ cameras based on digital twins as described in claim 1, characterized in that, The explosion-proof PTZ camera integrates the pixel grayscale gradient in the video frame with the spatial weighting coefficient of the gas concentration measurement point to form a cylindrical coordinate mapping table based on the centerline of the gas pipeline. Each coordinate entry in the cylindrical coordinate mapping table simultaneously records the grayscale value, gas concentration value, and temperature value of the location, forming a multimodal anchor point data cluster.

3. The remote inspection method for gas explosion-proof PTZ cameras based on digital twins as described in claim 1, characterized in that, The process of deducing the gas diffusion path after envelope rupture based on the written perturbation vector value includes the following steps: The disturbance vector value written into the emergency drill buffer is used to calculate the fracture offset of each coordinate entry point by point according to the gray gradient direction of each coordinate entry in the cylindrical coordinate mapping table. Then, the fracture offset and the instantaneous temperature value of the coordinate entry are forcibly aligned with the timestamp and fused. After calibration by the observation tilt angle calculated by the current gimbal attitude angle of the explosion-proof PTZ camera, the fracture initiation coordinate cluster is obtained after the cylindrical mesh remapping of the fracture offset. The fracture initiation coordinate cluster constitutes the initial fracture point set starting from the risk response envelope surface. Each point in the initial set of rupture points moves along the circumference and axial direction of the cylindrical coordinate mapping table, according to the moving step size constrained by the temperature value after filtering the local mean and the gas concentration value at the point, and jumps to the adjacent coordinate entries in sequence. After each jump, the correlation strength between the gray value of the coordinate entry and the background mean is used as the jump retention probability. When the jump retention probability is lower than the diffusion termination threshold, the jump stops. The coordinate entries passed through each jump, the current gas concentration value of the rupture point, and the jump retention probability are stored in an array in the order of the jumps to form the discretized propagation trajectory of the rupture point. The propagation trajectory forms a coordinate sequence of each rupture point carrying a concentration label. In the discretized propagation trajectories of all rupture points, coordinate entries whose radial distance from the gas pipeline centerline exceeds the maximum radial distance of the current risk response envelope are concatenated into a spatial polyline in the order of forced alignment by the explosion-proof PTZ camera timestamps. The coordinate entry corresponding to the maximum concentration change rate recorded in the last inspection by the explosion-proof PTZ camera is embedded into the polyline as the trajectory anchor point. After binding the trajectory line and anchor point and matching it with the boundary conditions of the emergency drill buffer, the gas diffusion path after the envelope rupture is obtained. The farthest radial distance from the end of the path to the pipeline centerline is taken as the maximum diffusion radius. The gas diffusion path and the maximum diffusion radius are directly written into the simulation results of the display layer of the emergency drill buffer of the remote command terminal.

4. The remote inspection method for gas explosion-proof PTZ cameras based on digital twins as described in claim 3, characterized in that, The process of binding the trajectory line to the anchor point and matching it with the boundary conditions of the emergency drill buffer zone includes the following steps: The gray-scale gradient direction of each coordinate entry in the embedded polyline of the trajectory anchor points is compared with the gray-scale gradient direction of the trajectory line and the anchor point. The coordinate entries with the angle between the two less than the minimum step angle of the explosion-proof PTZ camera are taken as binding nodes. After replacing the binding nodes with the coordinate values ​​of the trajectory anchor points and their maximum concentration change rate labels in the order of timestamps, the modified polyline with the historical risk feature identifier of the initial diffusion path is obtained after the anchor points are embedded. The predicted gas concentration value of each coordinate entry on the initial diffusion path is compared with the historical concentration benchmark value of the coordinate entries pre-stored in the emergency drill buffer. The coordinate entries whose absolute value of the difference exceeds the concentration fluctuation tolerance corresponding to the disturbance vector magnitude are taken as valid path points. After removing invalid path points in the cylindrical circumferential order and reconnecting the remaining valid path points, a filtered path matching the boundary conditions of the emergency drill buffer is obtained. The filtered path forms a diffusion path skeleton that conforms to the geometric constraints of the current risk response envelope. The radial distance of each coordinate entry in the diffusion path skeleton is compared with the predefined pipeline safety radius in the explosion-proof PTZ camera inspection digital twin. The coordinate entry whose radial distance is first less than the pipeline safety radius is taken as the path endpoint. After processing the output of all coordinate entries from the start point to the end point of the skeleton in the order of forced timestamp alignment, the gas diffusion path after envelope rupture is obtained.

5. The remote inspection method for gas explosion-proof PTZ cameras based on digital twins as described in claim 4, characterized in that, The process of forming a diffusion path skeleton that conforms to the geometric constraints of the current risk response envelope after filtering includes the following steps: The predicted gas concentration for each coordinate entry on the initial diffusion path is compared point-by-point with the historical concentration baseline value of the coordinate entries pre-stored in the emergency drill buffer. The absolute value of the difference is then taken. This absolute value is combined with the disturbance vector magnitude and the diffusion direction attenuation factor calculated from the current gimbal attitude angle of the explosion-proof PTZ camera to determine the final result. High fluctuation points in the concentration fluctuation marker pairs of each coordinate entry are selected as candidate valid points. Then, the points are filtered according to the angle between the gray gradient direction corresponding to the point and the normal of the risk response envelope boundary: candidate points with an angle smaller than the minimum step angle of the explosion-proof PTZ camera are retained, and candidate points with an angle greater than or equal to the minimum step angle are removed. After traversing all candidate points in cylindrical grid order and recording the retained points, the set of valid path points is obtained. The set of valid path points constitutes the sparse coordinate sequence after removing interference points. Points in the sparse coordinate sequence are sorted in ascending order of their circumferential angle in the cylindrical mapping table, and then sorted again by axial distance. Adjacent points are connected in sequence to form a polyline. During the connection, the cylindrical grid distance between adjacent points is checked to see if it exceeds the spatial resolution threshold corresponding to the optical zoom magnification of the explosion-proof PTZ camera. After filtering, the path forms a diffusion path skeleton that conforms to the geometric constraints of the current risk response envelope.

6. The remote inspection method for gas explosion-proof PTZ cameras based on digital twins as described in claim 5, characterized in that, If the absolute value of the difference is greater than the product of the magnitude of the disturbance vector and the attenuation factor, the point is marked as a high fluctuation point; otherwise, it is marked as a low fluctuation point. This yields a concentration fluctuation marker pair for each coordinate entry, and the concentration fluctuation marker pairs form a binary state set for determining whether a point should be retained.

7. The remote inspection method for gas explosion-proof PTZ cameras based on digital twins as described in claim 5, characterized in that, If the spatial resolution threshold corresponding to the optical zoom magnification of the explosion-proof PTZ camera is exceeded, a new coordinate point is inserted at the midpoint of the line connecting adjacent points, using the smaller of the predicted gas concentration values ​​of the two points as the concentration label; after connection and interpolation point supplementation, a filtered path matching the boundary conditions of the emergency drill buffer zone is obtained.

8. The remote inspection method for gas explosion-proof PTZ cameras based on digital twins as described in claim 7, characterized in that, The process of constructing a sparse coordinate sequence after removing interference points from the effective path point set includes the following steps: The gray-level gradient direction vector corresponding to each candidate valid point and the normal vector of the risk response envelope boundary at the point are respectively input into the direction angle calculation in the explosion-proof PTZ camera inspection digital twin. The direction angle value of each candidate valid point is obtained by querying the pre-stored gray-level gradient-normal angle mapping table. The minimum step angle of the pan-tilt unit (PTZ) at the current optical zoom level of the explosion-proof PTZ camera is read in real time, which is the physical minimum rotation angle of the PTZ stepper motor. The directional angle value of each candidate valid point is compared with the minimum step angle point by point: if the angle value is less than the minimum step angle, it is marked as a trackable point; if the angle value is greater than or equal to the minimum step angle, it is marked as an untrackable point. The trackable points are then written into the valid point queue of the inspection digital twin buffer according to the circumferential scanning order of the cylindrical grid, thus obtaining the initial set of valid points of the intermediate point sequence after filtering by the PTZ tracking capability. The cylindrical grid distance between two adjacent points in the intermediate point sequence is compared with the field of view coverage of the explosion-proof PTZ camera at maximum optical zoom. After traversing all adjacent points and completing the insertion process, an effective path point set that meets the requirements of continuous tracking of the PTZ is obtained. The effective path point set constitutes a sparse coordinate sequence that eliminates interference points and ensures the accessibility of the PTZ.

9. A remote inspection system for gas explosion-proof PTZ cameras based on digital twins, used to implement the remote inspection method for gas explosion-proof PTZ cameras based on digital twins as described in any one of claims 1 to 8, characterized in that, include: The weighting coefficient fusion module is used to merge the video frames, instantaneous gas concentration values, and instantaneous temperature values ​​collected by the explosion-proof PTZ camera with the spatial weighting coefficients of the gas concentration measurement points after the video frames, gas concentration, and temperature values ​​are forcibly aligned with the timestamps. This forms a cylindrical coordinate mapping table with the center line of the gas pipeline as the reference. Each coordinate entry in the cylindrical coordinate mapping table records the gray value, gas concentration value, and temperature value of the location, forming a multimodal anchor point data cluster. The disturbance coefficient acquisition module is used to extract all coordinate entries in the multimodal anchor point data cluster whose gas concentration values ​​are higher than the background mean. After local mean filtering of the temperature values ​​of the coordinate entries, the module multiplies them with the gray values ​​of the corresponding coordinates to obtain the dynamic weight factor of each coordinate. All coordinate entries with dynamic weight factors greater than a preset threshold are connected into a non-closed surface, which is defined as the risk response envelope. At the same time, the pre-entered construction behavior action library, including excavation impact, welding heat release rate, and mechanical vibration amplitude, is converted into a superposition of spatial disturbance vectors. The superposition vector is applied to each coordinate entry in the multimodal anchor point data cluster, and the rate of change of the gas concentration value of each coordinate entry before and after the disturbance vector is applied is calculated. The arithmetic mean of all the rates of change is taken as the disturbance coefficient of the construction behavior. The diffusion path deduction module is used to perform a dot product between the boundary coordinate entries of the obtained risk response envelope and the spatial disturbance vector corresponding to the construction behavior with a disturbance coefficient greater than zero, to obtain the local expansion potential of each point on the envelope surface. The three coordinate entries with the highest local expansion potential are used as the tracking target points of the PTZ camera, and the optical zoom of the explosion-proof PTZ camera is sequentially triggered to the maximum focal length to capture the visual details of the target points. At the same time, the disturbance vector value corresponding to the construction behavior with a disturbance coefficient exceeding the warning threshold is written into the emergency exercise buffer of the remote command terminal. The buffer uses the risk response envelope as the boundary condition, and deduces the gas diffusion path after the envelope ruptures based on the written disturbance vector value. The maximum diffusion radius obtained from the deduction is then reverse-calibrated as the upper bound of the dynamic weighting factor in the next inspection.

Images