Distributed multi-source heterogeneous sensor data processing method and system

Abstract

The invention relates to the technical field of data processing, in particular to a distributed multi-source heterogeneous sensor data processing method and system. The method comprises the following steps: collecting environmental parameters of a leakage area in real time through a distributed sensor array; performing coordinate system unification and timestamp alignment on the environmental parameters of the leakage area to generate a standardized leakage situation data set; extracting gas diffusion characteristics and thermal imaging characteristics of the standardized leakage situation data set to calculate a gas cloud cluster diffusion path probability; performing risk decision instruction generation on the gas cloud cluster diffusion path probability based on a preset leakage level classification neural network to obtain a risk decision instruction; calling a matched emergency plan based on the risk decision instruction; and analyzing the implementation steps of the emergency plan and performing instruction conversion to generate an emergency plan instruction. According to the invention, through real-time data acquisition, intelligent risk assessment and automatic emergency response, the timeliness responsiveness of data processing is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method and system for processing distributed multi-source heterogeneous sensor data. Background Art

[0002] Initially, sensor data processing mainly focused on the data acquisition and analysis of a single type of sensor, and the processing architecture was mostly centralized. The disadvantage of this method is that the system is vulnerable to single-point failures, and there are problems with data compatibility and processing efficiency for different types of sensors. With the rapid development of cloud computing and edge computing, distributed architectures can effectively process sensor data from different sources and different types. This process usually involves multiple links such as data acquisition, transmission, fusion, and analysis, and key factors such as the unification of heterogeneous data formats, real-time transmission, and efficient processing need to be considered in each link. With the continuous development of artificial intelligence and big data technologies, deep learning and machine learning algorithms have gradually been applied to the field of sensor data processing, improving the accuracy and real-time performance of data analysis. However, currently, the prediction of gas diffusion paths usually relies on a single data source, making it difficult to effectively fuse gas monitoring data and thermal imaging data. At the same time, traditional leakage risk decisions usually rely on manual judgment, with low efficiency and susceptibility to human factors, resulting in low timeliness and responsiveness in data processing. Summary of the Invention

[0003] Based on this, it is necessary to provide a method and system for processing distributed multi-source heterogeneous sensor data to solve at least one of the above technical problems.

[0004] To achieve the above object, a method for processing distributed multi-source heterogeneous sensor data includes the following steps:

[0005] Step S1: Real-time collect environmental parameters of the leakage area through a distributed sensor array; unify the coordinate system and align the timestamps of the environmental parameters of the leakage area to generate a standardized leakage situation data set;

[0006] Step S2: Extract the gas diffusion characteristics and thermal imaging characteristics of the standardized leakage situation data set to calculate the probability of the gas cloud diffusion path; generate a risk decision instruction based on a preset leakage level classification neural network for the probability of the gas cloud diffusion path to obtain a risk decision instruction;

[0007] Step S3: Invoke a matching emergency plan based on the risk decision instruction; analyze the implementation steps of the emergency plan and perform instruction conversion to generate an emergency plan instruction; sort the priorities of the emergency plan instructions to generate a sequential execution instruction set;

[0008] Step S4: Use the Internet of Things gateway to distribute the time-series execution instruction set to the intelligent explosion-proof robot terminal and optimize it dynamically to perform explosion-proof operations.

[0009] In the present invention, environmental parameters of the leakage area are collected in real time through a distributed sensor array, and the data is unified in the coordinate system and aligned with timestamps to generate a standardized leakage situation dataset, improving the consistency and timeliness of the data. Subsequently, gas diffusion features and thermal imaging features are extracted based on this dataset to calculate the diffusion path probability of the gas cloud, and a leakage level classification neural network is used for risk assessment to automatically generate risk decision instructions, ensuring the accuracy and reliability of the judgment. On this basis, the system automatically matches an appropriate emergency plan according to the risk decision instructions, analyzes the implementation steps of the plan for instruction conversion, generates emergency plan instructions, and sorts the instructions by priority to ensure the rationality and execution efficiency of the emergency response. Finally, the optimized time-series execution instruction set is distributed to the intelligent explosion-proof robot terminal through the Internet of Things gateway, and the robot performs explosion-proof operations according to the dynamic optimization strategy, realizing precise and efficient accident handling, reducing manual intervention, and improving safety. This process integrates technologies such as artificial intelligence, the Internet of Things, and automatic control, greatly enhancing the intelligent and automatic response capabilities to leakage accidents and being applicable to high-risk scenarios such as petrochemical industry and dangerous goods storage and transportation. Therefore, the present invention improves the timeliness and responsiveness of data processing through real-time data collection, intelligent risk assessment, and automatic emergency response.

[0010] Preferably, step S1 includes the following steps:

[0011] Step S11: Use a fixed multi-gas detector to synchronously obtain concentration gradient data of hydrogen sulfide, chlorine, and benzene series substances; scan the thermal radiation distribution of the leakage source through an airborne infrared spectral imaging device of a drone to obtain thermal radiation distribution data; perform heterogeneous data integration on the concentration gradient data and the thermal radiation distribution data through a distributed sensor array to obtain environmental parameters of the leakage area;

[0012] Step S12: Perform coordinate system unification processing on the environmental parameters of the leakage area to generate unified leakage environment data;

[0013] Step S13: Align the timestamps of the unified leakage environment data to generate aligned leakage environment data;

[0014] Step S14: Perform standardized conversion on the aligned leakage environment data to generate a standardized leakage situation dataset.

[0015] Through multi-source data fusion and standardized processing, the present invention realizes accurate perception and efficient modeling of the environment in the leakage area. First, a fixed multi-gas detector is used to synchronously collect the concentration gradient data of hydrogen sulfide, chlorine, and benzene series compounds, and an airborne infrared spectral imaging device is combined to scan the thermal radiation distribution of the leakage source to obtain high-precision thermal radiation data. Subsequently, a distributed sensor array fuses these heterogeneous data to form complete environmental parameters of the leakage area. Then, through unified processing of the coordinate system, spatial alignment of the data is achieved to ensure the compatibility and comparability of multi-source data. On this basis, the system performs timestamp alignment on the data to further improve the timeliness and synchronization of the data. Finally, through standardized conversion, a standardized leakage situation dataset in a unified format is generated. This process significantly improves the integrity, accuracy, and real-time nature of the leakage monitoring data, providing high-quality input data for subsequent gas diffusion modeling, risk assessment, and emergency response. At the same time, this method combines fixed sensors, UAV inspections, and intelligent data fusion technology, greatly improving the detection accuracy and response speed of hazardous chemical leaks, and is widely applicable to high-risk scenarios such as petrochemical industry and hazardous chemical storage and transportation.

[0016] Preferably, the extraction of the gas diffusion characteristics and thermal imaging characteristics of the standardized leakage situation dataset to calculate the gas cloud diffusion path probability in step S2 includes:

[0017] Perform gas concentration gradient analysis on the standardized leakage situation dataset to generate gas diffusion gradient data;

[0018] Perform thermal imaging spectral analysis on the standardized leakage situation dataset, extract the thermal distribution characteristics, and generate thermal imaging characteristic data;

[0019] Perform spatio-temporal correlation analysis based on the gas diffusion gradient data and the thermal imaging characteristic data to generate gas cloud diffusion trend data;

[0020] Perform probability modeling on the gas cloud diffusion trend data, calculate the gas cloud diffusion path probability under different wind speeds and environmental conditions, and generate gas diffusion path probability data.

[0021] The present invention realizes the accurate prediction of the diffusion path of the gas cloud through gas concentration gradient analysis, thermal imaging spectrum analysis, spatio-temporal correlation analysis and probability modeling. Its advantages lie in constructing a diffusion trend model by combining multi-dimensional data to ensure high-precision prediction of the diffusion path. At the same time, probability modeling is used to dynamically simulate the diffusion under different wind speeds and environmental conditions, so that the prediction results can adapt to complex and changeable actual working conditions. Based on the diffusion path probability data, the system can identify high-risk areas in advance, provide a scientific basis for emergency decision-making, thereby reducing the risks of personal injury and environmental pollution. In addition, this method provides high-quality data support for subsequent risk assessment and emergency plan optimization, improves the speed and accuracy of emergency response, and reduces accident losses. Generally speaking, this method improves the intelligent level of leakage monitoring, enhances the safety protection ability in high-risk scenarios, and is widely applicable to fields such as petrochemical industry and hazardous chemical storage and transportation.

[0022] Preferably, the thermal imaging spectrum analysis of the standardized leakage situation data set includes:

[0023] Import the standardized leakage situation data set into FLIR ResearchIR for spectral analysis parameter setting: set the spectral range to 8μm - 14μm, the frame rate to 60Hz, the temperature range to -20°C - 650°C, the emissivity is set to 0.95, and the Beijing temperature compensation is set to the average temperature of the on-site environment to obtain spectral analysis setting data;

[0024] Perform non-uniformity correction on the standardized leakage situation data set through the spectral analysis setting data, and select Gaussian 3×3 filtering to remove high-frequency noise to obtain spectral analysis filtering data;

[0025] Based on a preset high-temperature area threshold, perform hot spot analysis on the spectral analysis filtering data to generate thermal imaging feature data.

[0026] Through parameter optimization, non-uniformity correction, and high-frequency noise removal of the professional spectral analysis software FLIR ResearchIR, the present invention realizes high-precision thermal imaging feature extraction. First, for the standardized leakage situation dataset, the spectral range, frame rate, temperature range, and emissivity are accurately set, and temperature compensation is performed in combination with the average temperature of the on-site environment to ensure the accuracy and adaptability of spectral analysis. Subsequently, the non-uniformity correction of the thermal imaging data is performed using the spectral analysis setting data to eliminate the sensor response difference, and the Gaussian 3×3 filter is used to remove high-frequency noise to improve the signal-to-noise ratio of the data. Finally, hotspot analysis is performed through the preset high-temperature area threshold to accurately identify the thermal anomaly features in the leakage area and generate high-quality thermal imaging feature data. This method effectively improves the thermal imaging detection accuracy of the leakage source, ensures the reliability of the gas cloud diffusion modeling, provides accurate thermal imaging support for subsequent risk assessment and emergency response, and is widely applicable to the intelligent monitoring and early warning of high-risk scenarios such as petrochemical industry and hazardous chemical storage and transportation.

[0027] Preferably, the spatio-temporal correlation analysis based on the gas diffusion gradient data and the thermal imaging feature data includes:

[0028] Performing spatial interpolation analysis on the gas diffusion gradient data to generate continuous gas concentration field data;

[0029] Calculating the temperature gradient of the thermal imaging feature data to obtain the thermal convection influence data;

[0030] Performing multi-dimensional spatio-temporal matching on the continuous gas concentration field data and the thermal convection influence data to generate a gas diffusion-thermal imaging correlation matrix;

[0031] Performing trend fitting based on the gas diffusion-thermal imaging correlation matrix to generate gas cloud diffusion trend data.

[0032] Through spatial interpolation analysis, temperature gradient calculation, multi-dimensional spatio-temporal matching, and trend fitting, the present invention realizes the accurate correlation between gas diffusion characteristics and thermal imaging characteristics, improving the accuracy and reliability of diffusion trend prediction. First, continuous gas concentration field data is constructed through spatial interpolation analysis to eliminate data discreteness and improve the integrity of gas diffusion characteristics. Subsequently, the temperature gradient of thermal imaging characteristic data is calculated to generate heat convection influence data to quantify the effect of the thermal environment on gas diffusion. Then, through multi-dimensional spatio-temporal matching, gas diffusion data and heat convection data are fused to construct a gas diffusion-thermal imaging correlation matrix to realize in-depth correlation analysis of gas diffusion and environmental thermal characteristics. Finally, trend fitting is performed based on the correlation matrix to generate gas cloud diffusion trend data, providing accurate prediction support for gas leakage risk assessment and emergency response. This method effectively improves the fineness of gas diffusion simulation, making the prediction results more in line with actual environmental changes, and is applicable to leakage monitoring and warning systems in high-risk scenarios such as petrochemical industry and hazardous chemical storage and transportation.

[0033] Preferably, the construction process of the preset leakage level classification neural network in step S2 includes:

[0034] Obtain meteorological data;

[0035] Extract the leakage source location coordinates of the standardized leakage situation dataset to obtain leakage source location data;

[0036] Merge the meteorological data, leakage source location data, and gas cloud diffusion path probability to generate neural network input data;

[0037] Perform feedforward calculation on the neural network input data, and calculate the probability distribution of the leakage risk level through a multi-layer perceptron, and use the calculation result as the output of the model to obtain leakage risk level risk probability data;

[0038] Post-process the leakage risk level risk probability data to generate the final leakage risk level decision data;

[0039] Construct a leakage level classification neural network based on the neural network input data, leakage risk level risk probability data, and the final leakage risk level decision data to obtain the preset leakage level classification neural network.

[0040] Through the integration of meteorological data, leakage source location data, and the probability of gas cloud diffusion paths, the present invention constructs a leakage level classification neural network based on a multi-layer perceptron, achieving accurate determination and intelligent classification of leakage risk levels. First, meteorological data and leakage source location data are used to enhance the model's perception of environmental factors and improve its adaptability to gas diffusion behaviors under different working conditions. Second, through data merging and feed-forward calculations, the neural network can fully learn gas diffusion characteristics and generate a probability distribution of leakage risk levels, providing high-precision data support for leakage level assessment. Subsequently, through post-processing optimization, the stability and reliability of decision results are ensured, and finally, accurate leakage risk level decision data are generated. This method effectively improves the automation level and decision-making efficiency of leakage level classification, making risk assessment more intelligent and refined, capable of quickly responding to sudden leakage incidents, providing a scientific basis for the implementation of emergency plans, and being widely applicable to leakage monitoring and early warning systems in high-risk fields such as petrochemical and hazardous chemical storage and transportation.

[0041] Preferably, the invocation of the matching emergency plan based on the risk decision instruction in step S3 includes:

[0042] Analyze the leakage risk level in the risk decision instruction to obtain leakage risk level data;

[0043] Extract the emergency response requirements from the risk decision instruction according to the leakage risk level data to obtain emergency response requirement data;

[0044] Perform a query operation on the leakage risk level data, match the corresponding emergency plan from the preset emergency plan library, and generate emergency plan template matching data;

[0045] Conduct a risk assessment on the selected emergency plan template, and adjust the plan steps of the emergency plan template matching data according to the risk assessment results to obtain the matching emergency plan.

[0046] The present invention realizes the intelligent emergency response to leakage accidents, improves the accuracy and adaptability of emergency plans, by analyzing risk decision instructions, matching the emergency plan library, and making dynamic adjustments. First, based on the leakage risk level data, the emergency response requirements are accurately extracted to ensure the pertinence of the emergency plan. Subsequently, through the intelligent query and matching mechanism, the most suitable emergency plan can be quickly located from the preset emergency plan library to improve the response efficiency. At the same time, combined with the risk assessment results, the selected emergency plan is optimized and adjusted to make it more in line with the current leakage situation and environmental conditions, enhancing the feasibility and effectiveness of emergency disposal. Finally, this method can generate a highly matching emergency plan, realize fast and accurate decision support, reduce accident losses, and improve the intelligent level of emergency response, and is widely applicable to the emergency management systems of high-risk industries such as petrochemical and hazardous chemical storage and transportation.

[0047] Preferably, step S4 includes the following steps:

[0048] Step S41: Initialize and configure the IoT gateway;

[0049] Step S42: Send a timing execution instruction set to the intelligent explosion-proof robot terminal through the configured IoT gateway to activate the explosion-proof mode, and collect execution process data of the intelligent explosion-proof robot to obtain robot execution data;

[0050] Step S43: Extract acoustic vibration characteristic data from the robot execution data; construct a three-dimensional obstacle map based on the acoustic vibration characteristic data;

[0051] Step S44: Dynamically optimize the robot execution data according to the three-dimensional obstacle map to perform explosion-proof operations.

[0052] Through the collaborative work of the IoT gateway and the intelligent explosion-proof robot terminal, the present invention realizes the dynamic optimization and precise execution of intelligent explosion-proof operations. First, the initialization configuration of the IoT gateway ensures the stability and reliability of the system, providing a basic guarantee for subsequent instruction transmission. Subsequently, the intelligent explosion-proof robot is activated through the timing execution instruction set, and the execution process of the robot is monitored through real-time data collection to generate detailed robot execution data for subsequent analysis and optimization. Based on the acoustic vibration characteristics in the execution data, the system can construct a three-dimensional obstacle map to realize environmental perception and obstacle recognition, thereby providing more efficient path planning and decision-making support for explosion-proof operations. Finally, through dynamic instruction optimization, the system can adjust the robot operation in real time according to the actual execution situation, making the explosion-proof operation more accurate and safe, and avoiding potential risks. This method effectively improves the intelligence level of explosion-proof operations, ensures the safety and efficiency during the operation process, and is widely applicable to automated explosion-proof operations in high-risk fields such as petrochemical industry and hazardous material handling.

[0053] Preferably, step S41 includes the following steps:

[0054] Step S411: Configure the gateway hardware parameters of the IoT gateway: the gateway power input range is 12V ± 5% DC, set the working temperature range of the gateway to -20°C to +60°C, configure the communication interface of the gateway as a dual-band of 2.4GHz and 5GHz, and set the signal transmission rate to 150Mbps or 433Mbps;

[0055] Step S412: Configure the communication protocol of the IoT gateway: set the data transmission protocol as the MQTT protocol, and set the message quality service level as QoS1;

[0056] Step S413: Configure the network connection and bandwidth management of the IoT gateway: Configure the static IP address as 192.168.1.10 and limit the bandwidth usage of the gateway to 10 Mbps.

[0057] The present invention ensures the stability, reliability, and efficient data transmission ability of the system by finely configuring the hardware parameters, communication protocols, and network connections of the IoT gateway. First, through hardware parameter configuration, it ensures that the gateway operates normally within a specific power input and temperature range, optimizes the communication frequency band and signal transmission rate, and provides a stable wireless connection ability. Second, through communication protocol configuration, the MQTT protocol and an appropriate quality of service level (QoS1) are adopted, making data transmission more reliable, reducing the risk of data loss, and improving communication efficiency. Finally, through network connection and bandwidth management configuration, it ensures that the gateway can operate stably under the specified static IP address, while restricting bandwidth usage, optimizing the allocation of network resources, and avoiding the impact of bandwidth overload on system performance. Overall, this configuration method provides efficient hardware support and communication management for the IoT system, ensures the stable operation and real-time data transmission of the explosion-proof operation system, and is widely applicable to intelligent monitoring and emergency response systems in fields such as petrochemical and hazardous material storage and transportation.

[0058] In this specification, a distributed multi-source heterogeneous sensor data processing system is provided for implementing the above-mentioned distributed multi-source heterogeneous sensor data processing method. The distributed multi-source heterogeneous sensor data processing system includes:

[0059] A data acquisition module for real-time collecting environmental parameters of the leakage area through a distributed sensor array; unifying the coordinate system and aligning the timestamps of the environmental parameters of the leakage area to generate a standardized leakage situation data set;

[0060] A risk decision module for extracting the gas diffusion characteristics and thermal imaging characteristics of the standardized leakage situation data set to calculate the probability of the gas cloud diffusion path; generating a risk decision instruction based on a preset leakage level classification neural network for the probability of the gas cloud diffusion path to obtain a risk decision instruction;

[0061] An instruction generation module for calling a matching emergency plan based on the risk decision instruction; analyzing the implementation steps of the emergency plan and performing instruction conversion to generate an emergency plan instruction; sorting the priorities of the emergency plan instructions to generate a time-sequential execution instruction set;

[0062] An instruction distribution module for distributing and dynamically optimizing the time-sequential execution instruction set to an intelligent explosion-proof robot terminal through the IoT gateway to perform explosion-proof operations.

[0063] The beneficial effects of the present invention are as follows: through the collaborative work of modules such as data acquisition, risk decision-making, instruction generation, and instruction distribution, accurate assessment of leakage risks, intelligent emergency response, and efficient explosion-proof operations are achieved. First, the data acquisition module real-time monitors the environmental parameters of the leakage area through a distributed sensor array, and unifies and aligns them with timestamps through a coordinate system to generate a standardized leakage situation dataset, providing basic data support for subsequent decision-making. Next, the risk decision-making module extracts gas diffusion characteristics and thermal imaging characteristics, and combines a preset leakage level classification neural network to calculate path probabilities, thereby generating accurate risk decision-making instructions, which provide a scientific basis for the invocation and execution of emergency plans. Subsequently, the instruction generation module matches the emergency plan, conducts risk assessment and step adjustment to ensure the pertinence and operability of the emergency plan; at the same time, it prioritizes the emergency plan instructions to ensure that the most urgent tasks are processed first. Finally, the instruction distribution module, through the collaborative effect of the Internet of Things gateway and the intelligent explosion-proof robot terminal, distributes and optimizes the execution instructions in real time to achieve dynamic adjustment and accurate execution of explosion-proof operations. Overall, the system improves the efficiency and accuracy of emergency response to leakage accidents, ensures the safety and intelligence of explosion-proof operations, and is applicable to automated monitoring and emergency management in high-risk industries such as petrochemical and hazardous chemical storage and transportation. Therefore, the present invention improves the timeliness and responsiveness of data processing through real-time data acquisition, intelligent risk assessment, and automated emergency response. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 It is a schematic diagram of the step flow of a distributed multi-source heterogeneous sensor data processing method;

[0065] Figure 2 For Figure 1 It is a schematic diagram of the detailed implementation step flow of step S1 in

[0066] Figure 3 For Figure 1 It is a schematic diagram of the detailed implementation step flow of step S4 in

[0067] The realization, functional characteristics, and advantages of the object of the present invention will be further described in conjunction with the embodiments with reference to the drawings. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0068] The technical method of the present invention will be clearly and completely described below with reference to the drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0069] In addition, the attached drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus repeated descriptions thereof will be omitted. Some of the block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor methods and / or microcontroller methods.

[0070] It should be understood that although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are only used to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0071] To achieve the above object, please refer to Figures 1 to 3 , a distributed multi-source heterogeneous sensor data processing method, the method comprising the following steps:

[0072] Step S1: Real-time collect the environmental parameters of the leakage area through a distributed sensor array; unify the coordinate system and align the timestamps of the environmental parameters of the leakage area to generate a standardized leakage situation data set;

[0073] Step S2: Extract the gas diffusion characteristics and thermal imaging characteristics of the standardized leakage situation data set to calculate the probability of the gas cloud diffusion path; generate a risk decision instruction based on a preset leakage level classification neural network for the probability of the gas cloud diffusion path to obtain a risk decision instruction;

[0074] Step S3: Invoke a matching emergency plan based on the risk decision instruction; analyze the implementation steps of the emergency plan and perform instruction conversion to generate an emergency plan instruction; sort the priorities of the emergency plan instructions to generate a time-sequential execution instruction set;

[0075] Step S4: Use the Internet of Things gateway to distribute and dynamically optimize the time-sequential execution instruction set to the intelligent explosion-proof robot terminal to perform explosion-proof operations.

[0076] The present invention collects environmental parameters of the leakage area in real time through a distributed sensor array, unifies the coordinate system and aligns the timestamps of the data, generates a standardized leakage situation dataset, and improves the consistency and timeliness of the data. Subsequently, based on this dataset, gas diffusion features and thermal imaging features are extracted to calculate the diffusion path probability of the gas cloud, and a leakage level classification neural network is used for risk assessment to automatically generate risk decision instructions, ensuring the accuracy and reliability of the judgment. On this basis, the system automatically matches an appropriate emergency plan according to the risk decision instructions, analyzes the implementation steps of the plan for instruction conversion, generates emergency plan instructions, and at the same time sorts the instructions by priority to ensure the rationality and execution efficiency of the emergency operation. Finally, the optimized time-series execution instruction set is distributed to the intelligent explosion-proof robot terminal through the Internet of Things gateway, and the robot performs explosion-proof operations according to the dynamic optimization strategy, realizing accurate and efficient accident handling, reducing manual intervention, and improving safety. This process integrates technologies such as artificial intelligence, the Internet of Things, and automatic control, greatly improving the intelligent and automatic response capabilities to leakage accidents and is applicable to high-risk scenarios such as petrochemical industry and dangerous goods storage and transportation. Therefore, the present invention improves the timeliness and responsiveness of data processing through real-time data collection, intelligent risk assessment, and automatic emergency response.

[0077] In the embodiment of the present invention, with reference to Figure 1 shown, it is a schematic flow chart of the steps of a distributed multi-source heterogeneous sensor data processing method of the present invention. In this example, the distributed multi-source heterogeneous sensor data processing method includes the following steps:

[0078] Step S1: Collect environmental parameters of the leakage area in real time through a distributed sensor array; unify the coordinate system and align the timestamps of the environmental parameters of the leakage area to generate a standardized leakage situation dataset;

[0079] In the embodiments of the present invention, a distributed sensor array is arranged in the leakage risk area of a chemical plant, including gas concentration sensors (such as NDIR CO2 sensors, PID VOC sensors), temperature and humidity sensors, wind speed and direction sensors, barometric pressure sensors, etc. Each sensor collects data at a set time interval (such as 1 second or shorter) and transmits the data to the edge computing unit. The sensor nodes perform local data cleaning, such as eliminating sensor misreadings, filling in short-term missing data, and reducing noise through low-pass filtering. Each sensor in the leakage area has been calibrated with geographical coordinates (such as GPS coordinates or the local coordinate system of the plant) during installation, and the data is uniformly converted to the same coordinate system after collection. Since the timestamps of the data from each sensor are not synchronized, the Network Time Protocol (NTP) or GPS time synchronization technology is used to unify the timestamps and align the data through interpolation. All the collected data is stored in the standard JSON format or CSV format, including time, coordinates, gas concentration, temperature, humidity, wind speed and direction, etc. The standardized data is stored in the local database or the cloud data center to support subsequent modeling and analysis.

[0080] Step S2: Extract the gas diffusion characteristics and thermal imaging characteristics of the standardized leakage situation dataset to calculate the probability of the gas cloud diffusion path; generate a risk decision instruction based on the preset leakage level classification neural network for the probability of the gas cloud diffusion path to obtain a risk decision instruction;

[0081] In the embodiments of the present invention, gas concentration data (such as CO2, VOC, NH3, etc.) at different times and different positions are extracted from the standardized leakage situation dataset. Calculate the concentration change rate (gradient) to form a gas diffusion trend curve: where G(x, y, t) is the gas concentration change rate, C(x, y, t) is the gas concentration at the position (x, y) at time t, and t is the time. The Gaussian diffusion model is used to estimate the propagation range of the gas: where Q is the leakage source intensity, σ x and σ y are diffusion coefficients, and x, y are position coordinates. Combining the wind speed and direction data, the gas diffusion model is corrected to form a more accurate wind field corrected diffusion model: C′(x, y, t) = C(x, y, t) · e -αv ; where v is the wind speed and α is the correction coefficient. A thermal infrared camera (such as a FLIR infrared thermal imager) is used to photograph the leakage area to obtain thermal imaging data. Perform pseudo-color mapping to extract the temperature anomaly area (the temperature of the leakage point is significantly different from the background temperature). Set the background temperature baseline and calculate the temperature difference: ΔT = T leak -T background, if ΔT > θ (preset threshold), it is marked as a high-risk area. Neural networks such as YOLOv5 and ResNet are used to extract features from the thermal imaging images, detect the hot spot distribution in the leakage area, and extract the feature vector: F thermal = f CNN (I thermal ), where I thermal is the input thermal imaging data, and F thermal is the output feature vector. Combining the gas concentration distribution, wind speed and direction, and thermal imaging features, a gas diffusion probability model is constructed: P(x, y, t) = w1G(x, y, t) + w2C′(x, y, t) + w3F thermal ; where w1, w2, and w3 are weighting coefficients. Using time series neural networks such as LSTM and Transformer, inputting gas diffusion data and historical leakage data, and predicting the future diffusion path: P next = f LSTM (P(x, y, t)), calculating the probability of the high-risk area, and generating a heat map of the diffusion path of the gas cloud. Using a classification neural network (such as a Transformer classifier) to classify the risk of the gas cloud diffusion path: low risk (0 - 30%), medium risk (30 - 60%), high risk (above 60%). According to the classification results, corresponding emergency instructions are generated: low risk: closely monitor and give early warning. Medium risk: start local exhaust and alarm. High risk: immediately evacuate personnel and start emergency response.

[0082] Step S3: Invoke the matching emergency plan based on the risk decision instruction; analyze the implementation steps of the emergency plan and perform instruction conversion to generate emergency plan instructions; sort the priorities of the emergency plan instructions to generate a time series execution instruction set;

[0083] In the embodiments of the present invention, a contingency plan database is constructed by adopting knowledge graph technology, and the relevance between different leakage scenarios (such as chemical plants and underground pipelines) and contingency plans is established through an entity relationship network. It includes a relationship chain of "leakage type → gas property → affected area → contingency plan type → key action steps". Natural language processing (NLP) is used to parse risk decision instructions, and the best contingency plan is automatically matched through graph database query (such as Neo4j). Cosine similarity or Jaccard similarity is used to calculate the optimal matching contingency plan. Dependency Parsing is used to decompose the instructions in the contingency plan. For example, "activate the gas neutralization system" is disassembled into "operation object = gas neutralization system, operation = activate". Key actions (such as "blockade", "evacuation", "detection") are extracted based on named entity recognition (NER) and classified into a standardized instruction template. The parsed instructions are converted into a structured format suitable for subsequent sorting and execution. Key indicators for emergency response are defined (such as personnel safety, leakage control, environmental impact). The AHP method is used to calculate the relative weight of each indicator. For example: personnel safety (weight 0.5), leakage control (weight 0.3), equipment protection (weight 0.2). The comprehensive score is calculated according to the indicators involved in the instructions and sorted. State nodes are defined, and each emergency action is regarded as a state, such as [leakage detection] → [blockade] → [personnel evacuation] → [gas neutralization] → [environmental monitoring]. Through Boolean logic or conditional triggering (such as "the neutralization system can only be activated after the blockade is completed"), ensure that the order of instructions meets the emergency requirements. Based on real-time feedback (such as sensor data), adjust the execution order of sequential instructions. The Petri net is used to model the dependency relationship between each instruction to ensure that subsequent steps are executed after the completion of the previous tasks.

[0084] Step S4: Use the IoT gateway to distribute the sequential execution instruction set to the intelligent explosion-proof robot terminal and dynamically optimize it to perform explosion-proof operations.

[0085] In the embodiments of the present invention, by analyzing environmental parameters according to operation requirements, such as temperature, humidity, concentration of combustible gas, etc., the task type of the explosion-proof robot is determined, such as inspection, fire extinguishing, leakage detection, risk elimination treatment, etc. The task is disassembled into multiple execution steps, and the key nodes are clarified. The specific actions performed by the robot are determined, including forward movement, rotation, scanning, detection, grasping, etc. The instruction set includes task number, execution order, action type, execution parameters and priority. A time window for task execution is set to ensure that critical tasks are completed first. Ensure that the explosion-proof robot is properly connected to the IoT gateway and supports multiple communication protocols, such as 5G, WiFi, LoRa, etc. Configure data encryption and identity authentication mechanisms to prevent instructions from being tampered with. The IoT gateway parses the task plan and distributes the task instructions to the explosion-proof robot in the order of priority. After receiving the instructions, the robot performs verification and returns a confirmation message. According to the received sequential instructions, the robot sequentially performs various actions, such as moving, detecting, avoiding obstacles, operating the robotic arm, etc. During the execution process, the robot continuously monitors the environmental data. In case of any abnormal situation, the task can be paused and an alarm can be sent. The robot regularly uploads the execution status to the IoT gateway, including task progress, abnormal situations, real-time monitoring data, etc. The IoT gateway adjusts the instruction issuing rhythm according to the feedback to adapt to the actual operation environment. The IoT gateway analyzes the data returned by the robot to judge whether the operation progress meets the expectations. Combine AI algorithms to optimize the task execution order, such as adjusting the inspection route in high-temperature areas to reduce risk exposure. If the robot encounters an obstacle or detects abnormal gas, the IoT gateway can immediately adjust the instructions, modify the action path or execute a new task. After the task is completed, the system automatically generates an operation report and stores it in the cloud for subsequent analysis and optimization. After the robot finishes the execution, it sends a task completion signal to the IoT gateway. The IoT gateway records all execution data and performs closed-loop management on the task.

[0086] As an example of the present invention, refer to Figure 2 As shown, in this example, step S1 includes:

[0087] Step S11: Use a fixed multi-gas detector to synchronously obtain the concentration gradient data of hydrogen sulfide, chlorine, and benzene series; scan the thermal radiation distribution of the leakage source through an unmanned aerial vehicle (UAV)-borne infrared spectral imaging device to obtain thermal radiation distribution data; perform heterogeneous data integration on the concentration gradient data and the thermal radiation distribution data through a distributed sensor array to obtain the environmental parameters of the leakage area;

[0088] Step S12: Perform coordinate system unification processing on the environmental parameters of the leakage area to generate unified leakage environment data;

[0089] Step S13: Align the timestamps of the unified leakage environment data to generate aligned leakage environment data;

[0090] Step S14: Perform standardized conversion on the leakage environment alignment data to generate a standardized leakage situation data set.

[0091] In the embodiment of the present invention, by installing fixed multi-gas detectors in and around potential leakage areas, it is ensured that key points are covered, such as emission outlets, downstream areas in the wind direction, low-lying areas, etc. The concentration gradients of hydrogen sulfide, chlorine, and benzene series are synchronously detected to obtain real-time concentration data and spatial distribution. The collected data includes time, concentration values, and location information, forming a gas concentration gradient data table. Set the UAV flight path to cover the leakage source area and scan it with an infrared spectral imaging device. Use infrared imaging technology to detect the leakage point and the surrounding temperature anomaly area, generating thermal radiation distribution data. The infrared image data includes time, spatial coordinates, and temperature distribution matrix. Use a distributed sensor network to transmit the gas concentration gradient data and thermal radiation distribution data to the central processing unit. Adopt a heterogeneous data integration algorithm to fuse the multi-gas concentration data and infrared spectral data based on spatial position matching and environmental feature association, generating environmental parameters of the leakage area. Include complete environmental parameters such as gas concentration, temperature distribution, sensor location information, and time stamps. Uniformly convert the spatial coordinates of different data sources (such as GPS coordinates, relative position coordinates, UAV image coordinates) to a unified geographic coordinate system (WGS84 or UTM coordinates). Based on the spatial interpolation algorithm, perform position matching on the sensor data and UAV data to ensure that all data uses the same coordinate reference. Construct a spatial parameter matrix of the leakage area so that all data points can be analyzed in the same coordinate framework. Adopt a unified time reference (such as UTC time) to ensure that the time of all data sources (gas sensors, UAV images, sensor arrays) is consistent. Use the time interpolation algorithm to correct the time of non-synchronously collected data so that each data point can be matched according to the same time step. Store the aligned data in time series to form a time series leakage environment data set. Detect time stamp anomalies or data missing situations, and use interpolation or prediction completion techniques to ensure time continuity, and finally generate leakage environment alignment data. Normalize data such as concentration values, temperature values, and time series so that the data values fall into a unified interval (such as the 0-1 range). Use statistical analysis methods (such as mean deviation detection) to remove outliers and avoid the influence of incorrect data on the analysis results, generating a standardized leakage situation data set.

[0092] Preferably, extracting the gas diffusion characteristics and thermal imaging characteristics of the standardized leakage situation data set to calculate the probability of the gas cloud diffusion path in step S2 includes:

[0093] Perform gas concentration gradient analysis on the standardized leakage situation data set to generate gas diffusion gradient data;

[0094] Perform thermal imaging spectral analysis on the standardized leakage situation dataset, extract thermal distribution features, and generate thermal imaging feature data;

[0095] Conduct spatio-temporal correlation analysis based on the gas diffusion gradient data and the thermal imaging feature data to generate gas cloud diffusion trend data;

[0096] Perform probability modeling on the gas cloud diffusion trend data, calculate the gas cloud diffusion path probabilities under different wind speeds and environmental conditions, and generate gas diffusion path probability data.

[0097] In the embodiments of the present invention, by collecting the standardized leakage situation dataset, including the concentration data of gas sensors (such as gas detectors, infrared sensors) at different positions. Discretize the data of all measurement points in space to generate grid-like data (such as each grid size is 1m×1m). Allocate the measured gas concentration to the corresponding grid positions. Use the Finite Difference Method or gradient calculation to calculate the concentration difference between adjacent grid points, and obtain the gradient of the gas concentration (unit: ppm / m or mg / m 3 ·m). Formula: Where C i and C i+1are the gas concentrations at two adjacent grid points respectively, and Δx is the distance between these two grid points. The calculated gas diffusion gradient data is saved as a data matrix, representing the gas concentration change trend at different positions. Use a thermal imaging instrument (such as an infrared thermal imager) to collect thermal images of the leakage area. Ensure that the thermal imaging data is synchronized with the gas concentration data in time and space. Perform image processing on the thermal imaging data to extract the thermal distribution characteristics. Usually, use thermal image analysis methods to focus on areas with significant temperature changes, especially the temperature fluctuations near the leakage source. Calculate the average temperature of each grid area and identify temperature anomalies related to gas leakage (such as the gas cloud changing the surrounding thermal distribution due to heat absorption or heat release). Adopt a threshold-based detection method, set a threshold temperature, and identify the areas in the thermal image that exceed this temperature as potential leakage areas, generating thermal imaging feature data, representing the areas with temperature anomalies and their characteristic values (such as the temperature, range, duration, etc. of the area). Fuse the gas diffusion gradient data and the thermal imaging feature data. Consider the change in gas concentration and the spatial distribution of the temperature anomaly area, and conduct spatio-temporal correlation analysis. Adopt the Kalman Filter or the Mutual Information method to combine these two data sources in order to more accurately predict the diffusion path and trend of the gas cloud. Based on the correlation between the gas diffusion gradient and the thermal imaging characteristics, use the Particle Tracking Method to simulate the diffusion trend of the gas cloud. Calculate the spatial position of the gas cloud at different time steps according to the direction, speed of gas diffusion and the influence of temperature anomalies. Simulate the diffusion path of the gas cloud, consider environmental factors such as wind speed and wind direction, estimate the diffusion area of the gas cloud, and generate gas cloud diffusion trend data, including information such as diffusion path, speed, direction, etc., representing the diffusion dynamics of the gas cloud. According to the gas cloud diffusion trend data, consider different environmental factors such as wind speed, wind direction, temperature and humidity, and calculate the diffusion path probability of the gas cloud through Monte Carlo simulation or Bayesian Network. Use a stochastic process model (such as a random walk model) for simulation to calculate the diffusion probability of the gas cloud under different conditions. Evaluate the probability of each diffusion path and train and optimize the gas diffusion model according to historical data. For each diffusion path, calculate the occurrence probability of this path at different time points and different environmental conditions. For example, factors such as wind speed, wind direction, temperature and humidity all affect the path selection. Adopt the Maximum Likelihood Estimation (MLE) or the Least Squares method to fit the model to ensure that the probability model can accurately reflect the diffusion characteristics of the gas cloud, generating gas diffusion path probability data for subsequent risk assessment, emergency response planning, and determination of the leakage source.

[0098] Preferably, the thermal imaging spectral analysis of the standardized leakage situation dataset includes:

[0099] Import the standardized leakage situation dataset into FLIR ResearchIR for spectral analysis parameter settings: set the spectral range to 8μm - 14μm, the frame rate to 60Hz, the temperature range to -20°C - 650°C, the emissivity to 0.95, and the Beijing temperature compensation to the average temperature of the on-site environment to obtain spectral analysis setting data;

[0100] Perform non-uniformity correction on the standardized leakage situation dataset through the spectral analysis setting data, and select Gaussian 3×3 filtering to remove high-frequency noise to obtain spectral analysis filtering data;

[0101] Based on the preset high-temperature area threshold, perform hotspot analysis on the spectral analysis filtering data to generate thermal imaging feature data.

[0102] In the embodiments of the present invention, a standardized leakage situation dataset is imported into the FLIR ResearchIR software. This dataset should include thermal imaging data and relevant time and space coordinate information. Set the spectral analysis band range to 8μm - 14μm. This range belongs to the long-wave infrared (LWIR) band, which is suitable for detecting thermal imaging changes caused by gas leakage. Set the frame rate to 60Hz to ensure that the time resolution of data acquisition is high enough to monitor rapidly changing thermal distributions. Set the temperature range to -20°C to 650°C to adapt to temperature changes caused by different gas leaks, covering the range from low to high temperatures. Set the emissivity to 0.95, which is a relatively common emissivity for the surface of an object under normal circumstances and is applicable to most leaking objects. Set the Beijing temperature compensation to the average temperature of the on-site environment, and perform compensation according to the on-site environmental temperature to ensure the accuracy of the thermal imaging data. The software generates spectral analysis setting data according to the above settings, including information such as spectral range, frame rate, and temperature range, to prepare for subsequent spectral analysis. Use the spectral analysis setting data to perform non-uniformity correction on the standardized leakage situation dataset. This correction step is to correct the spatial non-uniform response of the detector of the thermal imaging device, thereby improving the accuracy of the thermal image. In FLIR ResearchIR, select the corresponding correction function, and the software will automatically correct the non-uniformity of the thermal imaging image. Select the Gaussian 3×3 filtering algorithm and apply it to the spectral analysis data for noise removal. Gaussian filtering reduces high-frequency noise by smoothing the image, ensuring that the thermal image is clearer and avoiding unnecessary image interference. In FLIR ResearchIR, select the filter setting as Gaussian 3×3 and apply it to the dataset for smoothing. After filtering, the noise in the image is effectively removed, and spectral analysis filtered data is obtained. At this time, the thermal imaging data has been corrected for non-uniformity and noise removal, presenting a more real and clear thermal distribution. Set the threshold for the high-temperature region according to the actual situation. This threshold should be able to distinguish the temperature change region near the leakage source. For example, set the threshold to a certain temperature point, such as 50°C, and all regions exceeding this temperature value will be regarded as the heat source region. In FLIR ResearchIR, use the threshold tool to set this temperature point. Based on the spectral analysis filtered data, perform hotspot analysis through the software to identify the regions in the thermal image that exceed the set threshold, that is, the hotspot regions. These regions usually indicate the positions of gas leakage sources or gas clouds. In FLIR ResearchIR, use the hotspot analysis tool, and the software will automatically identify the regions with abnormal temperatures and draw a hotspot map. According to the hotspot analysis results, generate thermal imaging feature data, including information such as the spatial distribution, temperature value, and duration of the hotspot regions. Extract the characteristic values (such as temperature, position, duration, etc.) of the hotspot regions to provide a basis for subsequent analysis and decision-making.

[0103] Preferably, the spatio-temporal correlation analysis based on the gas diffusion gradient data and the thermal imaging feature data includes:

[0104] Performing spatial interpolation analysis on the gas diffusion gradient data to generate continuous gas concentration field data;

[0105] Calculating the temperature gradient of the thermal imaging feature data to obtain heat convection influence data;

[0106] Performing multi-dimensional spatio-temporal matching on the continuous gas concentration field data and the heat convection influence data to generate a gas diffusion-thermal imaging correlation matrix;

[0107] Based on the gas diffusion-thermal imaging correlation matrix, performing trend fitting to generate gas cloud diffusion trend data.

[0108] In the embodiment of the present invention, by ensuring that the gas diffusion gradient data is ready, it contains gas concentration gradient information at multiple spatial positions. The data is usually collected by gas sensors at different positions and times. In order to generate continuous gas concentration field data, spatial interpolation techniques such as Kriging Interpolation or Inverse Distance Weighting (IDW) are used to interpolate the gas diffusion gradient data. Kriging Interpolation is based on the variogram to estimate the gas concentration at each position. First, calculate the spatial correlation between each position, and then use these correlations to estimate the concentration values in the blank areas. The Inverse Distance Weighting (IDW) method generates the gas concentration in the blank areas by weighted averaging of the data at the surrounding known positions. Using the interpolation algorithm, calculate and generate a continuous gas concentration field data set, representing the gas concentration values at each spatial position. Perform temperature gradient analysis on the thermal imaging feature data, calculate the temperature change rate at each spatial position, and then obtain the heat convection influence data. Use the finite difference method or the Gradient Operator to process the thermal imaging image: where T is the temperature, is the temperature gradient, and x, y, and z respectively represent the positions of the image on the spatial coordinate axes. The calculated data on the influence of heat convection is saved as a data matrix, representing the temperature change trend in the thermal imaging area. The continuous gas concentration field data and the data on the influence of heat convection are integrated into a unified multi-dimensional data structure. Each data point contains gas concentration and temperature change information and has spatio-temporal dimensions. The data includes: spatial dimension (position coordinates) and time dimension (gas concentration and temperature data at different time points). Spatio-temporal matching is performed on the continuous gas concentration field data and the data on the influence of heat convection, and techniques such as Cross-Correlation or Dynamic Time Warping (DTW) are used to find the correlation between the two. Cross-Correlation evaluates the relationship between the gas concentration change and the temperature gradient by calculating the correlation coefficient between the two. Dynamic Time Warping (DTW), when dealing with time series data, can evaluate the spatio-temporal correlation between them by comparing the matching degree of the gas concentration and the temperature gradient at different time points. The result of spatio-temporal matching is stored as a gas diffusion-thermal imaging correlation matrix, which reflects the correlation between the gas diffusion process and the thermal imaging data. Each element in the matrix represents the degree of correlation between the gas concentration and the temperature change at a certain spatio-temporal position. According to the gas diffusion-thermal imaging correlation matrix, a suitable fitting algorithm is selected for trend modeling. This matrix provides the correlation information between gas diffusion and thermal imaging and can be used as input data. Methods such as polynomial fitting, exponential fitting, or Support Vector Regression (SVR) are used to model the diffusion trend of the gas cloud. The Least Squares Method is used to fit the gas diffusion-thermal imaging correlation matrix to find the best trend curve or trend function. For non-linear relationships, curve fitting methods can be used for more accurate trend modeling. Formula: f(x) = a0 + a1x + a2x 2 +…+a n x n ; where x is a time or space variable, and a n are fitting parameters. Through the fitted trend model, gas cloud diffusion trend data is generated, representing the diffusion trend of the gas cloud over time and space. This data is used to predict the future behavior of the gas cloud and its potential risks.

[0109] Of particular importance is that the multi-dimensional spatio-temporal matching of the continuous gas concentration field data and the data on the influence of heat convection further includes:

[0110] Performing field discretization processing on the continuous gas concentration field data to generate grid-based concentration data; extracting temporal sequence features from the grid-based concentration data to generate concentration change feature data;

[0111] Extract the convection characteristics of the data on the influence of thermal convection to generate thermal convection characteristic data; perform spatial distribution analysis on the data of concentration change characteristics to generate diffusion pattern data;

[0112] Perform flow field structure analysis on the thermal convection characteristic data to generate flow field structure data; synchronize the diffusion pattern data in time to generate time series alignment data; perform spatial registration on the flow field structure data to generate spatial mapping data;

[0113] Analyze the spatio-temporal correlation between the time series alignment data and the spatial mapping data to generate multi-dimensional correlation data; perform matrix operations on the multi-dimensional correlation data to generate a gas diffusion-thermal imaging correlation matrix.

[0114] In the embodiments of the present invention, by obtaining continuous gas concentration field data, the data should include gas concentration values at different time and space positions. The gas concentration field is subjected to grid processing, and the entire gas concentration field is discretized into multiple grid cells. Each grid cell represents the gas concentration at a spatial position. Regular grid processing or adaptive grid processing methods can be used: Regular grid processing is based on a predetermined grid resolution (such as each grid unit being 1m×1m), and the gas concentration field is divided into regular grids. Adaptive grid processing automatically adjusts the grid resolution according to the severity of the gas concentration change, using smaller grid cells in areas with larger gas concentration changes. The generated grid-concentrated data includes the gas concentration values within each grid cell, providing basic data for subsequent analysis. The concentration values of each grid cell at multiple time points are extracted from the grid-concentrated data to form time series data. The time series data of each grid cell is analyzed to extract the following time series features: Mean value: representing the average concentration, reflecting the overall concentration of the gas during a certain period. Standard deviation: representing the degree of concentration fluctuation, reflecting the variability of the gas concentration. Rate of change: calculating the rate of change of the concentration between adjacent time points, reflecting the rate of change of the concentration over time. Periodic analysis: analyzing the periodic changes in the concentration time series to determine whether the gas diffusion shows periodic fluctuations. Through time series analysis, concentration change feature data is generated, and these feature data describe the trend and fluctuations of the gas concentration over time. Obtain heat convection influence data, usually obtained through thermal imaging data or temperature gradient data. Calculate heat convection characteristics, mainly including: calculating the temperature gradient at each position to understand how heat flows from high-temperature areas to low-temperature areas. Based on the temperature gradient, infer the direction and speed of the heat flow and judge the flow pattern of the heat convection. Evaluate the intensity of the heat flow, usually represented by calculating the rate of temperature change, and generate heat convection feature data to describe information such as the temperature gradient, flow direction, and intensity in the heat convection area. Obtain concentration change feature data, which is derived from the time series feature extraction process of the grid-concentrated data. Use spatial interpolation methods (such as Kriging interpolation, IDW, etc.) to perform spatial interpolation on the concentration change feature data to generate a continuous concentration change distribution map. Analyze the spatial distribution pattern of the concentration change, including: the distribution of high-concentration areas and low-concentration areas, and the spatial range of gas diffusion. Based on spatial analysis, generate diffusion pattern data to describe the spatial distribution trend of gas diffusion and identify diffusion paths and hot spots. Obtain heat convection feature data, which contains information such as the direction and intensity of the heat flow. Use fluid mechanics analysis methods (such as flow field simulation or particle tracking) to analyze the flow field structure in the heat convection area. The analysis objectives include: Stability of the flow field: judging whether the flow field is stable and whether there are eddy currents or unstable areas. Expandability of the flow field: predicting whether the heat convection will expand outward, and generating flow field structure data to describe the flow structure and pattern of the heat convection, providing basic data for subsequent spatio-temporal matching.Time synchronization of diffusion mode data is performed by interpolation or dynamic time warping (DTW) to ensure data comparison in the same time dimension. Interpolation is performed on the diffusion mode data according to the time interval between data to generate continuous time series data. Alignment of irregular time series is carried out to ensure the synchronization of different time series data. Output time series aligned data, which synchronizes the diffusion mode with time and can provide the correct time dimension for spatio-temporal analysis. Spatial registration of the flow field structure data is performed using spatial registration algorithms (such as rigid transformation, affine transformation, non-rigid transformation, etc.) and aligned with the gas concentration field. Translation, rotation, and scaling are performed on the flow field structure data to align it with the gas concentration field data. For complex flow fields, a more complex transformation model is used to register the flow field structure. Output spatial mapping data, which is used to describe the spatial relationship between the heat convection structure and the gas concentration distribution. The spatio-temporal correlation between the time series aligned data and the spatial mapping data is calculated through correlation analysis (such as Pearson correlation, mutual information method, etc.). Using methods such as weighted average or principal component analysis (PCA), these two types of data are fused to analyze their spatio-temporal dynamic characteristics. Output multi-dimensional correlation data, which describes the spatio-temporal correlation between gas diffusion and heat convection and provides a basis for predicting the gas diffusion path. Matrix operations are performed on the multi-dimensional correlation data, such as matrix multiplication, eigenvalue decomposition, or singular value decomposition (SVD), etc., to extract the main features from it. Using these features, a gas diffusion-thermal imaging correlation matrix is generated, which reflects the complex relationship between gas diffusion and thermal imaging data.

[0115] Preferably, the construction process of the preset leakage level classification neural network in step S2 includes:

[0116] Obtain meteorological data;

[0117] Extract the leakage source location coordinates of the standardized leakage situation data set to obtain leakage source location data;

[0118] Merge the meteorological data, leakage source location data, and gas cloud diffusion path probability to generate neural network input data;

[0119] Perform a feed-forward calculation on the neural network input data, and calculate and generate the probability distribution of the leakage risk level through a multi-layer perceptron, and use the calculation result as the output of the model to obtain the leakage risk level risk probability data;

[0120] Post-process the leakage risk level risk probability data to generate the final leakage risk level decision data;

[0121] Construct a leakage level classification neural network based on the neural network input data, leakage risk level risk probability data, and the final leakage risk level decision data to obtain the preset leakage level classification neural network.

[0122] In the embodiments of the present invention, by obtaining real-time or historical meteorological data, including but not limited to air temperature, wind speed, wind direction, humidity, atmospheric pressure, etc., these data can be obtained from weather stations, satellite data, weather forecasting systems or meteorological APIs. The meteorological data can be stored in CSV, JSON or API response formats to ensure that the data includes timestamps and various meteorological parameters. Clean the obtained meteorological data, process missing values and outliers, and ensure that the timestamps of the data are consistent with other data for subsequent fusion. Obtain a standardized leakage situation dataset, which includes relevant information of the leakage source (such as location, leakage volume, leakage type, etc.). Extract the leakage source location coordinates from the leakage situation dataset, and these locations are usually values in longitude, latitude or a spatial coordinate system. If the data contains multiple leakage sources, the coordinate extraction needs to be performed for each leakage source to generate the corresponding leakage source location data. Standardize the extracted leakage source location data and map it to a format suitable for the input of the neural network (for example, normalize the coordinate values to the 0-1 interval). Align these data according to timestamps or spatial coordinates: merge the meteorological data and the leakage source location data with the gas diffusion path probability data. The merged dataset can use time series data or a spatial coordinate system as the basis for matching. If the data formats are inconsistent, interpolation or filling methods can be used to align different sources of data at a unified time point or location, generating the merged neural network input data, including meteorological information, leakage source location, diffusion path probability, etc., and preparing it as the input of the neural network. Design a multi-layer perceptron (MLP) neural network architecture, which usually includes: Input layer: Accept the merged input data. Hidden layer: At least include one or more layers of neurons, and adjust the number of layers and neurons according to the task complexity and the amount of training data. Output layer: Used to output the probability distribution of the leakage risk level. Select a suitable activation function (such as ReLU, Sigmoid or Softmax) to increase the non-linear ability. Input the neural network input data generated in step S23 into the neural network model. The neural network performs calculations through the forward process: the data enters from the input layer, passes through the neurons in the hidden layer for weighted summation, activation, and is transmitted to the output layer. The output layer generates the probability distribution of the leakage risk level, that is, the occurrence probability of each risk level. The neural network outputs the probability distribution of the leakage risk level and provides the corresponding probability value for each risk level. Post-process the output leakage risk level probability. Common methods include: determining the final leakage risk level according to the set threshold; selecting the risk level with the highest probability as the final classification result; if there are high probability values for multiple risk levels, a weighted average method can be used for comprehensive decision-making. Select a probability threshold or make a multi-class decision on the leakage risk level risk probability data.Ensure that the final result meets the actual requirements. For example, if the risk level is divided into three levels: low, medium, and high, then when the probability exceeds a certain threshold, it is determined as that risk level, and the final leakage risk level decision data is generated, that is, the risk level selected according to the model output and the threshold. Use the input data, output data, and post-processing results in steps S23 and S24 to train the neural network and construct a leakage level classification neural network. Select an appropriate loss function (such as the cross-entropy loss function) to optimize the model to ensure that the neural network can effectively classify different leakage risk levels. Train the model through the backpropagation algorithm, continuously adjust the weight parameters of the neural network until the loss function reaches the minimum. Use historical data for training and evaluate the performance of the model using the cross-validation method. Evaluate indicators such as the accuracy, recall rate, and F1 score of the model based on the validation data during the training process, and select the best model. The constructed and trained leakage level classification neural network can perform leakage risk assessment in real-time or offline.

[0123] Preferably, the step of invoking the matching emergency plan based on the risk decision instruction in step S3 includes:

[0124] Analyze the leakage risk level in the risk decision instruction to obtain leakage risk level data;

[0125] Extract the emergency response requirements from the risk decision instruction according to the leakage risk level data to obtain emergency response requirement data;

[0126] Perform a query operation on the leakage risk level data, match the corresponding emergency plan from the preset emergency plan library, and generate emergency plan template matching data;

[0127] Conduct a risk assessment on the selected emergency plan template, and adjust the plan steps of the emergency plan template matching data according to the risk assessment result to obtain the matching emergency plan.

[0128] In the embodiments of the present invention, the risk decision-making instruction is the leakage risk level obtained through a neural network model based on real-time or historical data. This instruction contains information related to the leakage event, including the leakage source location, gas diffusion path, meteorological conditions, leakage risk level, etc. By parsing the risk decision-making instruction, the leakage risk level data therein is extracted. The leakage risk level usually includes three levels: low, medium, and high, or more detailed classifications. The data can be parsed through natural language processing technology or simple data extraction methods to extract the leakage risk level information and generate leakage risk level data, that is, the specific leakage risk level contained in the risk decision-making instruction. According to the leakage risk level data, the corresponding emergency response requirement data is extracted. The emergency response requirements are directly related to the risk level of the leakage: Low risk level: Less intervention is required, which is monitoring or minor treatment. Medium risk level: More emergency resources need to be deployed, such as personnel, equipment, and material preparation. High risk level: Immediate full response should be taken, including full evacuation, area blockade, and gas detection. Using a preset response requirement template or rule engine, map to the corresponding emergency response requirements according to different risk levels. For example, in the case of low risk, the response requirements only include local environmental monitoring and recording; in the case of high risk, the response requirements include full evacuation, evacuation route planning, and air quality detection, generating emergency response requirement data, including specific response tasks, resource requirements, time requirements, etc. The preset emergency plan library is a database containing emergency response plans for various risk levels constructed based on the analysis results of historical leakage events and simulation cases. The emergency plan library can be stored classified according to leakage risk level, leakage type, environmental conditions, etc. According to the leakage risk level data, a query operation is performed in the emergency plan library. Through the query interface, the emergency plan most suitable for this risk level is matched. If the risk level is low, a low-level emergency plan can be selected, and the operations involved are relatively simple. If the risk level is high, then query the emergency plan corresponding to the high risk level in the library, and the operations involved are relatively complex and require coordination of multiple resources. Through the set query algorithm or rule engine, the leakage risk level data is matched with the plan template in the emergency plan library. The matching process usually includes: screening the emergency plan template based on information such as risk level, leakage type, and meteorological conditions. It is necessary to further adjust the content of the matched plan according to a specific scenario (such as chemical leakage, oil and gas leakage, etc.) to generate emergency plan template matching data, that is, the emergency response plan template obtained from the emergency plan library according to the query result. Conduct a risk assessment on the selected emergency plan template to ensure the effectiveness and adaptability of the plan. The key points of the assessment include: assessing whether the plan matches the risk level of the current leakage event. Considering the impact of environmental changes on the implementation of the plan, such as the impact of changes in meteorological conditions (wind speed, wind direction, etc.) on the gas diffusion path. Assessing whether the required resources (personnel, equipment, materials, etc.) are sufficient and whether there is a potential risk of resource shortage.Adjust the emergency plan according to the risk assessment results: If it is found in the assessment that some steps in the plan are not suitable for the current situation, make adjustments and optimizations. According to the feedback of the assessment, it is necessary to modify the emergency response sequence, the deployment location of the response personnel, the arrangement of the response time, etc. Conduct simulation drills to verify whether the emergency plan can be executed smoothly, and further optimize the plan steps according to the drill results to generate a matching emergency plan, that is, based on the risk assessment and the adjusted emergency plan, ensure that it is applicable to the handling of the current leakage incident. Carry out the actual emergency response according to the matching emergency plan and execute various tasks in the plan. The specific tasks include: Conduct area blockade, personnel evacuation, etc. according to the leakage risk level and environmental conditions. Select appropriate fire extinguishing and leakage control equipment according to the leakage type. During the emergency response process, monitor the progress of the leakage incident and environmental changes in real time, and adjust the emergency plan in a timely manner to ensure the effectiveness of the emergency response. After the event, conduct a post-event assessment, check the implementation of the emergency plan, summarize the experience and lessons, and improve the emergency plan for the response to similar events in the future. Complete the emergency response and feedback the emergency response results.

[0129] As an example of the present invention, refer to Figure 3 As shown, in this example, step S4 includes:

[0130] Step S41: Initialize and configure the IoT gateway;

[0131] Step S42: Send a timing execution instruction set to the intelligent explosion-proof robot terminal through the configured IoT gateway to activate the explosion-proof mode, and collect the execution process data of the intelligent explosion-proof robot to obtain the robot execution data;

[0132] Step S43: Extract the acoustic vibration characteristic data from the robot execution data to obtain the acoustic vibration characteristic data; construct a three-dimensional obstacle map based on the acoustic vibration characteristic data;

[0133] Step S44: Dynamically optimize the robot execution data according to the three-dimensional obstacle map to perform explosion-proof operations.

[0134] In the embodiments of the present invention, by ensuring the normal operation of the hardware system of the Internet of Things gateway, connecting it to the network, and enabling it to communicate with the intelligent explosion-proof robot terminal. The Internet of Things gateway supports multiple protocols (such as MQTT, HTTP, CoAP, etc.) for data transmission. Set up the network connection between the gateway and the local or remote server, including configuring Wi-Fi or cellular network connections. According to specific circumstances, data encryption and communication security policies can be selected to ensure data security during the communication process. Conduct device authentication on the Internet of Things gateway to verify the identity of the gateway to ensure the security of the connection. Bind the gateway to the intelligent explosion-proof robot terminal and set corresponding device parameters, such as device ID, communication frequency, data transmission protocol, etc. After completing the initialization configuration, the gateway establishes communication with each device in the system (including the intelligent explosion-proof robot terminal) to ensure the stability and reliability of data transmission. According to the requirements of explosion-proof operations, design a timing execution instruction set, which includes a series of operation tasks (such as movement, positioning, sensor scanning, explosion-proof operation start, etc.), and set the execution timing and execution conditions for each task. Each instruction in the instruction set includes operation content, parameter settings, task time, and task trigger conditions. Send the designed timing execution instruction set to the intelligent explosion-proof robot terminal through the Internet of Things gateway. The gateway is responsible for transmitting the instructions to the robot terminal device to ensure the accurate transmission of the instructions by wireless or wired means. Each task in the instruction set will be executed in the predetermined timing after the robot terminal receives the instructions, and the robot activates the explosion-proof mode according to the instruction set. During the explosion-proof operation, the robot will collect the execution process data, including but not limited to: robot position and attitude data (through the built-in positioning system and inertial measurement unit IMU), various sensor data (such as temperature, gas concentration, vibration, etc.), operation process status (such as whether the operation is successful, whether an abnormality occurs, etc.). These data are transmitted to the monitoring system or server through the Internet of Things gateway in real time for processing and analysis to generate robot execution data, including various operation states and sensor feedback data of the robot, providing a basis for subsequent analysis. Extract the acoustic vibration data from the execution process data collected by the robot. This data is collected by the vibration sensors or acoustic sensors equipped on the robot. The vibration signals monitored by these sensors during the robot's explosion-proof operation can reveal the presence of obstacles or the interaction between the robot and the environment. Perform signal processing on the collected acoustic vibration data, including: denoising the data to remove environmental noise and sensor errors, using methods such as Fourier transform (FFT) or wavelet transform to convert the signal from the time domain to the frequency domain, extracting the vibration characteristics in different frequency ranges, calculating the amplitude, frequency and other characteristics of the acoustic vibration signal, and understanding the vibration intensity and propagation mode. These processing results help to identify the characteristics of obstacles (such as hard, soft, fixed or moving obstacles). Extract the acoustic vibration feature data, which will be used to construct a three-dimensional obstacle map and subsequent dynamic optimization instructions.Using acoustic wave vibration characteristic data, identify the presence and properties of obstacles through algorithms (such as threshold-based detection, machine learning models, etc.). Determine the type, location, size, and shape of obstacles based on the changes in vibration signals, and convert the acoustic wave vibration characteristic data into a three-dimensional obstacle map. This requires modeling the location information and spatial distribution of each obstacle, usually using point cloud data or gridded data to represent obstacles in space. Through the feedback of sensors and combined with the positioning information of the robot, update the obstacle map in real time, form a three-dimensional space model around the robot, generate a three-dimensional obstacle map, display the obstacle distribution around the robot, and provide data support for subsequent dynamic optimization instructions. According to the three-dimensional obstacle map, analyze the current movement path of the robot and the positions of surrounding obstacles. Determine the optimal path through path planning algorithms (such as the Dijkstra algorithm). Determine whether it is necessary to avoid certain obstacles or select a safer path to perform explosion-proof tasks. Dynamically adjust the operation instructions of the robot according to real-time environmental information and the execution status of the robot, which includes: modifying the speed, angle, or direction of the robot's movement to ensure avoiding obstacles. Adjusting the explosion-proof operation steps of the robot to cope with environmental changes or unexpected events on site. The optimized instructions will be sent to the robot terminal through the Internet of Things gateway to generate dynamically optimized execution instructions, ensuring that the robot can successfully complete the explosion-proof operation task while avoiding obstacles and coping with potential dangers.

[0135] Particularly importantly, step S44 further includes the following steps:

[0136] Step S441: Confirm the obstacle position according to the three-dimensional obstacle map; calculate the obstacle blocking height for the robot execution data based on the obstacle position to obtain the obstacle blocking height data;

[0137] Step S442: Perform primary obstacle avoidance on the robot execution data through the obstacle blocking height data to obtain the primary obstacle avoidance path data and the primary obstacle avoidance instruction;

[0138] Step S443: Use the primary obstacle avoidance optimization instruction to perform horizontal distance matching on the robot execution data and the obstacle avoidance path data. When the horizontal distances match, calculate the obstacle covering area for the three-dimensional obstacle map to obtain the obstacle covering area data;

[0139] Step S444: Perform secondary obstacle avoidance on the robot execution data based on the obstacle covering area data to obtain the secondary obstacle avoidance path optimization data and the secondary obstacle avoidance instruction;

[0140] Step S445: Optimize the steering of the robot execution data through a primary obstacle avoidance instruction and a secondary obstacle avoidance instruction to obtain the robot steering optimization data; optimize the movement of the robot execution data through the primary obstacle avoidance path data and the secondary obstacle avoidance path data to generate the robot movement optimization data;

[0141] Step S446: Optimize the pose dynamic instruction of the robot execution data according to the robot steering optimization data and the robot movement optimization data to perform explosion-proof operations.

[0142] In the embodiments of the present invention, the position coordinates of all obstacles are extracted from a three-dimensional obstacle map to identify the spatial positions and shapes of the obstacles. The precise positions and three-dimensional coordinates of the obstacles are obtained using a Light Detection and Ranging (LIDAR) or a stereo vision system to ensure high-precision obstacle information. The positions of the obstacles are marked on the map, and the volume, shape, and height of each obstacle are determined. The blocking height of each obstacle is calculated based on the position and morphology of the obstacle, that is, the height occupancy of the obstacle on the robot's travel path. Based on the geometric shape and grid model of the obstacle, the height range of the obstacle in a specific direction is calculated. Through simulation modeling or actual measurement, the height of the obstacle that the robot cannot cross is obtained to avoid the robot from touching or being blocked, generating obstacle blocking height data to provide data support for subsequent obstacle avoidance path planning. Based on the obstacle blocking height data, an obstacle avoidance calculation is performed on the robot execution data to generate a preliminary path to avoid the obstacles. In the path planning algorithm, classical path planning methods such as the Dijkstra algorithm are used to calculate the obstacle avoidance path according to the positions of the obstacles. The path considering obstacle avoidance should be the shortest while avoiding being too close to the obstacles. An obstacle avoidance instruction is generated according to the calculated obstacle avoidance path, and the instruction content includes the target direction, travel path, necessary turns, acceleration, and deceleration during obstacle avoidance, etc. This instruction is transmitted to the robot execution system to start the obstacle avoidance execution process of the robot, generating first-time obstacle avoidance path data and a first-time obstacle avoidance instruction to provide guidance for the next operation of the robot. Using the first-time obstacle avoidance path data and the robot execution data, the horizontal distance between the robot and the obstacle is matched. The horizontal distance between the robot and the obstacle is measured in real time through sensors (such as LIDAR, ultrasonic sensors, etc.) to adjust the forward direction of the robot. The route for the robot to bypass the obstacle is optimized in path planning to ensure the accuracy of obstacle avoidance. When the horizontal distances match, by calculating the footprint area of the obstacle, it is determined whether the robot can pass through or needs to re-avoid. Based on the intersection area between the actual footprint area of the obstacle and the robot path, the influence range of the obstacle is determined. This calculation considers whether the robot path overlaps with the footprint area of the obstacle to avoid inappropriate path selection, generating obstacle footprint area data to provide a basis for generating the next obstacle avoidance instruction. Based on the obstacle footprint area data, a secondary obstacle avoidance is performed on the robot execution data to generate a more optimized obstacle avoidance path. The space around the obstacle is evaluated more precisely to avoid misjudgment and re-plan the obstacle avoidance route. The Bellman-Ford algorithm or the fast marching algorithm is used for path optimization to make the obstacle avoidance path smoother and safer. According to the optimized path, a secondary obstacle avoidance instruction is generated, including the adjusted travel direction, speed, and bypass strategy. The instruction is transmitted to the robot through wireless communication, and the instruction triggers the execution system of the robot to make adjustments, generating secondary obstacle avoidance path optimization data and a secondary obstacle avoidance instruction, enabling the robot to more precisely avoid the obstacles. Based on the first-time obstacle avoidance instruction and the secondary obstacle avoidance instruction, the turning of the robot is optimized.Combine the path data and optimize the steering angle and path smoothness of the robot through a PID controller or a fuzzy control algorithm. Ensure the accurate driving direction of the robot. While avoiding obstacles, prevent large - scale turning that may cause inefficiency. Optimize the movement of the robot based on the primary obstacle - avoidance path data and the secondary obstacle - avoidance path data. Adjust the movement speed and acceleration of the robot to ensure smoothness and efficiency during obstacle avoidance. Use a speed - position - time curve to optimize the motion performance of the robot near obstacles, avoid sudden acceleration and deceleration, and generate robot steering optimization data and robot movement optimization data to provide guidance for the next step of execution. According to the optimized robot steering optimization data and robot movement optimization data, optimize the pose dynamic instructions of the robot execution data. Through a dynamic optimization algorithm, adjust the current position and pose (including roll, pitch, and yaw angles) of the robot in real - time to ensure the precise implementation of explosion - proof operations. Based on the real - time position feedback and target position of the robot, adjust the robot pose to complete complex operation actions, generate pose dynamic instruction optimization data, and ensure that the robot can execute tasks efficiently and safely during explosion - proof operations.

[0143] Preferably, step S41 includes the following steps:

[0144] Step S411: Configure the gateway hardware parameters of the Internet of Things gateway: The input range of the gateway power supply is 12V ± 5% DC, set the operating temperature range of the gateway to be from - 20°C to + 60°C, configure the communication interface of the gateway to be a dual - band of 2.4GHz and 5GHz, and set the signal transmission rate to be 150Mbps or 433Mbps;

[0145] Step S412: Configure the communication protocol of the Internet of Things gateway: Set the data transmission protocol to the MQTT protocol and set the message quality service level to QoS1;

[0146] Step S413: Configure the network connection and bandwidth management of the Internet of Things gateway: Configure the static IP address to be 192.168.1.10 and limit the bandwidth usage of the gateway to 10Mbps.

[0147] Therefore, from any perspective, the embodiments should be regarded as exemplary and non - restrictive. The scope of the present invention is defined by the appended claims rather than the above description. Thus, all changes falling within the meaning and scope of the equivalent elements of the application documents are intended to be encompassed within the present invention.

[0148] The above are only specific embodiments of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but rather to the broadest scope consistent with the principles and novel features invented herein.

Claims

1. A method for processing distributed multi-source heterogeneous sensor data, characterized in that, Applied to the emergency scenario of toxic gas leakage in chemical industrial parks, deployed on fixed multi-gas detectors, airborne infrared spectral imaging devices of unmanned aerial vehicles, and intelligent explosion-proof robot terminals, including the following steps: Step S1: Real-time collect the environmental parameters of the leakage area through a distributed sensor array; unify the coordinate system and align the timestamps of the environmental parameters of the leakage area to generate a standardized leakage situation dataset; Step S2: Extract the gas diffusion characteristics and thermal imaging characteristics of the standardized leakage situation dataset to calculate the gas cloud diffusion path probability; generate a risk decision instruction based on a preset leakage level classification neural network for the gas cloud diffusion path probability to obtain a risk decision instruction; Step S3: Invoke a matching emergency plan based on the risk decision instruction; analyze the implementation steps of the emergency plan and perform instruction conversion to generate an emergency plan instruction; sort the priorities of the emergency plan instructions to generate a sequential execution instruction set; Step S4: Use the Internet of Things gateway to distribute and dynamically optimize the sequential execution instruction set to the intelligent explosion-proof robot terminal to perform explosion-proof operations.

2. The distributed multi-source heterogeneous sensor data processing method according to claim 1, characterized in that Step S1 includes the following steps: Step S11: Use a fixed multi-gas detector to synchronously obtain the concentration gradient data of hydrogen sulfide, chlorine, and benzene series; scan the thermal radiation distribution of the leakage source through an airborne infrared spectral imaging device of an unmanned aerial vehicle to obtain thermal radiation distribution data; perform heterogeneous data integration on the concentration gradient data and thermal radiation distribution data through a distributed sensor array to obtain the environmental parameters of the leakage area; Step S12: Perform coordinate system unification processing on the environmental parameters of the leakage area to generate unified leakage environment data; Step S13: Align the timestamps of the unified leakage environment data to generate aligned leakage environment data; Step S14: Perform standardized conversion on the aligned leakage environment data to generate a standardized leakage situation dataset.

3. The distributed multi-source heterogeneous sensor data processing method according to claim 1, wherein The extraction of the gas diffusion characteristics and thermal imaging characteristics of the standardized leakage situation dataset in Step S2 to calculate the gas cloud diffusion path probability includes: Perform gas concentration gradient analysis on the standardized leakage situation dataset to generate gas diffusion gradient data; Perform thermal imaging spectral analysis on the standardized leakage situation dataset, extract the thermal distribution characteristics, and generate thermal imaging characteristic data; Perform spatio-temporal correlation analysis based on the gas diffusion gradient data and thermal imaging characteristic data to generate gas cloud diffusion trend data; Perform probability modeling on the gas cloud diffusion trend data to calculate the gas cloud diffusion path probability under different wind speeds and environmental conditions, and generate gas diffusion path probability data.

4. The distributed multi-source heterogeneous sensor data processing method according to claim 3, wherein The thermal imaging spectral analysis of the standardized leakage situation dataset includes: Import the standardized leakage situation dataset into FLIR ResearchIR for spectral analysis parameter settings: set the spectral range to 8μm - 14μm, the frame rate to 60Hz, the temperature range to -20°C - 650°C, the emissivity is set to 0.95, and the Beijing temperature compensation is set to the average temperature of the on-site environment to obtain spectral analysis setting data; Perform non-uniformity correction on the standardized leakage situation dataset through the spectral analysis setting data, and select Gaussian 3×3 filtering to remove high-frequency noise to obtain spectral analysis filtering data; Performing hotspot analysis on the spectral analysis filtered data based on a preset high-temperature region threshold to generate thermal imaging feature data.

5. The distributed multi-source heterogeneous sensor data processing method according to claim 3, characterized in that The spatio-temporal correlation analysis based on the gas diffusion gradient data and the thermal imaging feature data includes: Performing spatial interpolation analysis on the gas diffusion gradient data to generate continuous gas concentration field data; Calculating the temperature gradient of the thermal imaging feature data to obtain thermal convection influence data; Performing multi-dimensional spatio-temporal matching on the continuous gas concentration field data and the thermal convection influence data to generate a gas diffusion-thermal imaging correlation matrix; Performing trend fitting based on the gas diffusion-thermal imaging correlation matrix to generate gas cloud diffusion trend data.

6. The distributed multi-source heterogeneous sensor data processing method according to claim 1, characterized in that, The construction process of the preset leakage level classification neural network in step S2 includes: Obtaining meteorological data; Extracting the leakage source location coordinates of the standardized leakage situation dataset to obtain leakage source location data; Merging the meteorological data, the leakage source location data, and the gas cloud diffusion path probability to generate neural network input data; Performing feedforward calculation on the neural network input data, and calculating through a multi-layer perceptron to generate the probability distribution of the leakage risk level, and taking the calculation result as the output of the model to obtain the leakage risk level risk probability data; Performing post-processing on the leakage risk level risk probability data to generate the final leakage risk level decision data; Constructing a leakage level classification neural network according to the neural network input data, the leakage risk level risk probability data, and the final leakage risk level decision data to obtain the preset leakage level classification neural network.

7. The distributed multi-source heterogeneous sensor data processing method according to claim 1, wherein The invoking of the matching emergency plan based on the risk decision instruction in step S3 includes: Analyzing the leakage risk level in the risk decision instruction to obtain leakage risk level data; Extracting the emergency response requirements from the risk decision instruction according to the leakage risk level data to obtain emergency response requirement data; Performing a query operation on the leakage risk level data, and matching the corresponding emergency plan from the preset emergency plan library to generate emergency plan template matching data; Performing risk assessment on the selected emergency plan template, and adjusting the plan steps of the emergency plan template matching data according to the risk assessment result to obtain the matching emergency plan.

8. The distributed multi-source heterogeneous sensor data processing method according to claim 1, characterized in that Step S4 includes the following steps: Step S41: Initializing and configuring the IoT gateway; Step S42: Sending a timing execution instruction set to the intelligent explosion-proof robot terminal through the configured IoT gateway to activate the explosion-proof mode, and collecting the execution process data of the intelligent explosion-proof robot to obtain robot execution data; Step S43: Extracting the acoustic wave vibration characteristics from the robot execution data to obtain acoustic wave vibration characteristic data; constructing a three-dimensional obstacle map based on the acoustic wave vibration characteristic data; Step S44: Dynamically optimizing the robot execution data according to the three-dimensional obstacle map to perform explosion-proof operations.

9. The distributed multi-source heterogeneous sensor data processing method according to claim 8, characterized in that, Step S41 includes the following steps: Step S411: Configure the gateway hardware parameters of the Internet of Things gateway: the gateway power input range is 12V±5%DC, set the working temperature range of the gateway to be -20°C to +60°C, configure the communication interface of the gateway to be a dual-band of 2.4GHz and 5GHz, and set the signal transmission rate to 150Mbps or 433Mbps; Step S412: Configure the communication protocol of the Internet of Things gateway: set the data transmission protocol to the MQTT protocol and set the message quality service level to QoS1; Step S413: Configure the network connection and bandwidth management of the Internet of Things gateway: configure the static IP address to be 192.168.1.10 and limit the bandwidth usage of the gateway to 10Mbps.

10. A distributed multi-source heterogeneous sensor data processing system, characterized in that For implementing the distributed multi-source heterogeneous sensor data processing method as described in claim 1, the distributed multi-source heterogeneous sensor data processing system includes: A data acquisition module, which is used to collect the environmental parameters of the leakage area in real time through a distributed sensor array; unify the coordinate system and align the timestamps of the environmental parameters of the leakage area to generate a standardized leakage situation data set; A risk decision module, which is used to extract the gas diffusion characteristics and thermal imaging characteristics of the standardized leakage situation data set to calculate the probability of the gas cloud diffusion path; generate a risk decision instruction based on the preset leakage level classification neural network for the probability of the gas cloud diffusion path to obtain a risk decision instruction; An instruction generation module, which is used to call the matching emergency plan based on the risk decision instruction; analyze the implementation steps of the emergency plan and perform instruction conversion to generate an emergency plan instruction; sort the priorities of the emergency plan instructions to generate a sequential execution instruction set; An instruction distribution module, which is used to distribute and dynamically optimize the sequential execution instruction set to the intelligent explosion-proof robot terminal by using the Internet of Things gateway to perform explosion-proof operations.

Images