A multi-channel wide-spectrum gas leak monitoring system, method, apparatus, and medium

Abstract

The present application relates to a kind of multi-channel wide-spectrum gas leakage monitoring system, method, device and medium.The system includes: data acquisition module is configured to acquire the visible light image of monitoring area, obtains the spectral image of the respective corresponding characteristic absorption wave band of multiple leakage gases by multiple independent signal channels and visible light image and spectral image are fused to generate fusion video image;Data processing module is connected with data acquisition module, is configured to receive fusion video image and compare and analyze it with preset gas spectral image database, obtain leakage gas composition information and the spatial information related to gas leakage;Alarm module is connected with data processing module, is configured to respond to at least one of leakage gas composition information and spatial information and carry out visual positioning and presentation operation and trigger alarm.The present application can realize the real-time, online, multi-point multi-component synchronous monitoring and accurate positioning of leakage gas.

Description

Technical Field

[0001] This invention belongs to the field of industrial safety and environmental monitoring technology, and more specifically, relates to a multi-channel broadband gas leak monitoring system, method, device and medium. Background Technology

[0002] In modern industrial settings such as chemical industrial parks, oil storage and transportation facilities, and pharmaceutical workshops, monitoring the leakage of hazardous gases is a crucial step in ensuring production safety, preventing accidents, and mitigating environmental pollution. Traditional gas leak monitoring technologies generally suffer from limitations such as limited monitoring range, inability to simultaneously identify multiple components, difficulty in accurately locating leak sources, and high system deployment and maintenance costs. These limitations fail to meet the current industrial demands for real-time monitoring that is "large-area, comprehensive, multi-component, highly sensitive, and locatable."

[0003] Currently, common gas leak monitoring methods have significant limitations: (1) Point-type electrochemical or catalytic combustion sensors are usually deployed near potential leak points. The monitoring range is usually only centimeter-level, with monitoring blind spots. They are generally only for a single gas and are susceptible to poisoning and cross-interference. They have a high false alarm rate and short lifespan, and cannot achieve gas imaging and leak source location. (2) Open path Fourier transform infrared spectroscopy (OP-FTIR) has the advantages of long optical path, high sensitivity and multi-gas detection, but it is mostly single-point or single-path monitoring. Covering a large area requires multiple systems, which is costly and difficult to coordinate data. Its spectral range and gas identification ability are limited, and it cannot image in real time to intuitively present the leak location and diffusion trend. (3) Tunable diode laser absorption spectroscopy (TDLAS) has high sensitivity and fast response, but it is usually only for a single or a few gases. It is difficult to achieve wide coverage and multi-component synchronous monitoring. If multiple lasers or broadband tuning schemes are used, the cost and technical complexity will increase significantly. It also does not have imaging capabilities, which makes leak location difficult.

[0004] For example, Chinese patent CN221124311U discloses a methane gas infrared detector, which focuses on the design of the housing seal, moisture-proof structure, and installation adjustment mechanism, aiming to improve the environmental adaptability of single gas (methane) detection. However, it is still a point-based monitoring device and does not have imaging or multi-component identification capabilities. Chinese patent CN204855361U discloses an infrared dual-band combustible gas detector, which uses a dual-band infrared light source and an embedded signal processing system to improve detection stability and signal-to-noise ratio. However, it still only detects single or limited gas types and cannot achieve spatial location and visualization of the leak source.

[0005] Therefore, existing monitoring methods, due to their inherent limitations, are insufficient to meet the requirements of comprehensive coverage, real-time visualization, and precise source tracing in complex industrial scenarios. Thus, there is an urgent need to develop an integrated monitoring system capable of combining wide-spectrum detection, multi-channel acquisition, real-time imaging, intelligent identification, and positioning functions to systematically improve the completeness, accuracy, and response efficiency of gas leak monitoring. Summary of the Invention

[0006] To address the aforementioned problems in existing technologies, this invention aims to provide a multi-channel broadband gas leak monitoring system, method, device, and medium. This invention is applicable to real-time, online, multi-point, multi-component synchronous monitoring and precise location of leaks of various hazardous gases such as volatile organic compounds (VOCs), hydrogen sulfide, ammonia, chlorine, and methane in industries such as petroleum, chemical, coal, pharmaceutical, and environmental protection. Furthermore, the advantages of this invention lie in its wide coverage band and diverse gas identification types, enabling not only rapid location of leak sources and analysis of diffusion trends but also significant advantages in system cost and maintainability, providing an integrated solution for industrial safety monitoring.

[0007] To solve the above-mentioned technical problems or achieve the above-mentioned objectives, the present invention adopts the following technical solution:

[0008] According to a first aspect of the present invention, a multi-channel broadband gas leak monitoring system is provided, comprising:

[0009] The data acquisition module is configured to acquire visible light images of the monitoring area, obtain spectral images of the characteristic absorption bands of various leaked gases through multiple independent signal channels, and fuse and overlay the visible light images and spectral images to generate a fused video image of the monitoring area.

[0010] The data processing module is communicatively connected to the data acquisition module. The data processing module is configured to receive fused video images and compare and analyze the fused video images with a preset gas spectral image database to obtain information on the composition of the leaked gas and spatial information related to the gas leak.

[0011] An alarm module, which is communicatively connected to a data processing module, is configured to perform a visual location and presentation operation and trigger an alarm in response to at least one of the leaked gas composition information and spatial information.

[0012] The above-mentioned technical solution of the present invention realizes an integrated system architecture for multi-channel broadband gas leak monitoring. It can achieve real-time, online, multi-point, multi-component synchronous monitoring and accurate positioning of leaked gas by fusing visible light and multi-band spectral images. It solves the problems of limited monitoring range, inability to detect multiple gases simultaneously, inability to identify gas types, and difficulty in locating leak sources of traditional monitoring equipment.

[0013] In one embodiment of the present invention, the data acquisition module includes:

[0014] A visible light image acquisition module, configured to acquire visible light images of the monitored area;

[0015] At least two infrared signal acquisition modules constitute multiple independent signal channels. Each infrared signal acquisition module operates in a different band and is configured to acquire infrared spectral images of the characteristic absorption bands corresponding to the leaked gas within its respective operating band range.

[0016] In this technical solution, by configuring a visible light image acquisition module and at least two infrared signal acquisition modules with different working bands, the synchronous acquisition of multi-channel and multi-band spectral signals is realized, which enhances the coverage of the characteristic absorption spectra of various gases and improves the accuracy and efficiency of gas type identification and concentration inversion.

[0017] In one embodiment of the present invention, the data acquisition module further includes:

[0018] The laser module forms an independent signal channel, which can be combined with the infrared signal acquisition module and the visible light image acquisition module to form a multi-signal channel. The laser module is configured to emit modulated lasers into the monitoring area to identify leaked gases that cannot be identified or are difficult to identify by the infrared signal acquisition module.

[0019] In this technical solution, the present invention further expands the gas identification capability of the system by introducing a laser module as an additional independent signal channel, especially for gas types that are difficult to identify by infrared signals, thereby improving the detection sensitivity and applicability of the system.

[0020] In one embodiment of the present invention, the data acquisition module includes:

[0021] Visible light image acquisition module, which is configured to acquire visible light images of the monitored area;

[0022] A single infrared signal acquisition module is configured to acquire infrared spectral images of the characteristic absorption bands of the leaked gas within the corresponding operating band range.

[0023] The laser module constitutes an independent signal channel, which together with the single infrared signal acquisition module forms multiple independent signal channels. The laser module is configured to emit modulated lasers into the monitoring area to identify gases that cannot be identified or are difficult to identify by the single infrared signal acquisition module, thereby working in conjunction with the single infrared signal acquisition module to obtain spectral images of the characteristic absorption bands corresponding to various leaked gases.

[0024] In this technical solution, the present invention, based on configuring only a single infrared signal acquisition module, combines a laser module to form multiple independent signal channels, thereby simplifying the hardware structure while still possessing multi-gas identification capabilities, and reducing system cost and complexity.

[0025] In one embodiment of the present invention, the laser module is configured to perform the following operations:

[0026] The laser emission frequency is modulated to the characteristic absorption frequency of the target leaking gas, and the corresponding modulated laser beam is emitted into the monitoring area.

[0027] The intensity of the returned signal of the modulated laser beam is monitored by a receiver;

[0028] In response to a return signal strength lower than a preset signal strength threshold, the system determines that there is a target leaking gas in the monitoring area and forms a cross-shaped positioning mark.

[0029] In this technical solution, the laser module precisely modulates the laser frequency to the characteristic absorption spectrum of the target gas and detects its absorption attenuation signal, achieving highly selective and sensitive quantitative identification of specific leaked gases. This effectively solves the shortcomings of broadband infrared imaging in terms of gas identification specificity and quantitative accuracy. At the same time, when a leak is detected, the module actively generates and superimposes a cross-shaped positioning mark, transforming the abstract signal into intuitive spatial location information, significantly improving the entire monitoring system's ability to confirm leaked gases, quantitative accuracy, and intuitiveness of on-site positioning.

[0030] In one embodiment of the present invention, the system is configured to, after the laser module determines that there is a target leaking gas, retrieve a pre-made gas cloud image of the target leaking gas from the background through software, and overlay and fuse the gas cloud image with a cross-shaped positioning mark for visualization and alarm.

[0031] This technical solution transforms abstract detection signals into intuitive and concrete on-site situation maps by invoking preset gas cloud images and fusing them with laser positioning markers. This addresses the shortcomings of traditional gas alarms, which only provide data or simple icons and lack a clear visual representation of the scale and spread of leaks. It uses preset images to instantly simulate typical diffusion patterns, combined with precise crosshair positioning, allowing operators to grasp key information such as "where the leak is and how it might spread" without requiring professional interpretation. This significantly improves the intuitiveness of the alarm and the efficiency of decision support, achieving a technological leap from "detection alarm" to "situational awareness."

[0032] In one embodiment of the invention, the laser module includes a quantum cascade laser array.

[0033] In this technical solution, a quantum cascade laser array is used as the laser module, which has advantages such as a wide wavelength tuning range, high output power, and high modulation accuracy, and is suitable for trace gas detection and high-precision spectral analysis in complex environments.

[0034] In one embodiment of the present invention, at least two infrared signal acquisition modules include any of the following combinations:

[0035] (i) A combination of at least one long-wave infrared module and at least one mid-wave infrared module;

[0036] (ii) A combination of at least two long-wave infrared modules.

[0037] This technical solution utilizes a modular architecture that combines long-wave infrared modules with mid-wave infrared modules, or dual long-wave infrared modules, enabling flexible system configuration to adapt to the diverse types of leaked gases and monitoring needs in different industrial scenarios. The combination of long-wave and mid-wave modules achieves broad spectral coverage and band complementarity, effectively detecting various gases with strong absorption characteristics in different infrared bands. The dual long-wave combination, on the other hand, can finely identify specific gases or enhance anti-interference capabilities in complex environments by subdividing the long-wave spectrum. This design significantly improves the system's adaptability, detection reliability, and cost-effectiveness, achieving the engineering advantage of "one architecture, multiple configurations."

[0038] In one embodiment of the present invention, when at least two infrared signal acquisition modules are in combination (i), the at least two infrared signal acquisition modules include a long-wave infrared module with a working wavelength of 6-14μm and a mid-wave infrared module with a working wavelength of 3-5μm.

[0039] This technical solution employs a combination of a 6-14μm long-wave infrared module and a 3-5μm mid-wave infrared module, forming a complementary broadband detection system. This combination achieves full coverage detection of most common industrial leaked gases: the long-wave band (6-14μm) exhibits strong absorption characteristics for most volatile organic compounds (VOCs) and sulfides, while the mid-wave band (3-5μm) is particularly sensitive to small molecule gases such as methane and carbon monoxide, as well as high-temperature targets. This configuration allows the system to simultaneously and comprehensively monitor multiple gases with distinct characteristics in both bands, greatly expanding the range of detectable gases for a single system and significantly improving its comprehensive monitoring capabilities and reliability in complex industrial environments.

[0040] In one embodiment of the present invention, when at least two infrared signal acquisition modules are combined (i), the operating band of the long-wave infrared module is any sub-band within 6-14μm, and the operating band of the mid-wave infrared module is any sub-band within 3-5μm.

[0041] In this technical solution, the operating bands of the long-wave and mid-wave modules are further defined as sub-bands that can be flexibly selected within the ranges of 6-14μm and 3-5μm, giving the system a high degree of customizability for specific application scenarios and target gases. By precisely matching the strongest absorption peak of the target gas and avoiding interference spectral bands in the atmospheric window, this design can significantly improve the detection signal-to-noise ratio and selectivity. At the same time, it provides hardware configuration flexibility for optimizing the optical system, reducing costs, and adapting to detectors of different performance levels, realizing accurate and efficient monitoring for differentiated needs on a general platform.

[0042] In one embodiment of the present invention, when at least two infrared signal acquisition modules are in combination (ii), the combination of at least two long-wave infrared modules includes long-wave infrared modules with any number of non-overlapping sub-bands within the operating wavelength range of 6-14μm.

[0043] This technical solution employs multiple long-wave infrared sub-band modules with non-overlapping operating wavebands to achieve collaborative analysis and differential detection within the long-wave infrared spectrum (6-14μm). This design can effectively eliminate environmental background radiation interference by comparing and correlating multi-channel signals to target gases with multiple absorption characteristics or wide absorption bands within this spectrum. This significantly improves the accuracy of specific gas identification, concentration inversion precision, and anti-interference capability in complex scenarios, making it particularly suitable for precision leak monitoring and gas classification applications with extremely high detection specificity requirements.

[0044] In one embodiment of the present invention, at least two infrared signal acquisition modules include a long-wave infrared module with a working wavelength of 6-8μm and a long-wave infrared module with a working wavelength of 8-14μm.

[0045] This technical solution employs a combination of long-wave infrared modules operating in the 6-8μm and 8-14μm bands, achieving continuous, gapless coverage and synergistic enhancement of the long-wave infrared spectral band (6-14μm). This combination not only ensures comprehensive detection capabilities for various VOCs and industrial gases within this critical spectral band, but also improves the system's identification accuracy and concentration inversion reliability for gases with complex or broad-spectrum absorption characteristics through parallel acquisition and comparative analysis of two adjacent sub-bands. Simultaneously, it enhances stability and detail resolution under dynamic thermal background interference, providing a hardware foundation for high-precision, high-reliability gas leak imaging monitoring.

[0046] In one embodiment of the present invention, a single infrared signal acquisition module includes a long-wave infrared module with a working wavelength of 6-14μm.

[0047] In this technical solution, only a 6-14μm long-wave infrared module is configured in the simplified system architecture, which can still cover the absorption band of most common industrial gases. It is suitable for basic monitoring scenarios and has the advantages of low cost, simple structure and convenient maintenance.

[0048] In one embodiment of the present invention, the infrared signal acquisition module includes an infrared camera, and the detector of the infrared camera is an infrared focal plane detector.

[0049] In this technical solution, an infrared focal plane detector is used as the core component of the infrared camera. It has the advantages of fast response speed, high detection sensitivity and clear imaging, and is suitable for dynamic and large-scale gas leak monitoring.

[0050] In one embodiment of the present invention, the operating band of the infrared signal acquisition module is achieved by filter coating or infrared focal plane detector coating.

[0051] In this technical solution, the working band is locked by coating the filter or the detector. It has a simple structure, low cost, high stability, and is suitable for long-term stable operation in industrial environments.

[0052] In one embodiment of the present invention, the gas composition information includes gas type and gas concentration, and the spatial information includes gas leak location and gas leak diffusion direction.

[0053] In this technical solution, the system can not only identify the gas type and concentration, but also provide spatial information such as the leak location and diffusion direction, realizing the full-process monitoring capability from "detection" to "location" and then to "prediction".

[0054] In one embodiment of the present invention, the data processing module has a built-in gas spectral image database and is configured to identify gas types and invert gas concentrations based on fused video images and the gas spectral image database.

[0055] In this technical solution, by using a built-in gas spectral image database, the system can quickly compare and identify gas types, and combine image data to invert concentration, thereby improving the automation level of gas identification and quantitative analysis.

[0056] In one embodiment of the invention, the data processing module is configured to process the fused video image using a partial least squares algorithm and a neural network algorithm to retrieve the gas concentration.

[0057] In this technical solution, a hybrid algorithm combining partial least squares and neural networks is adopted, which not only ensures the computational efficiency of concentration inversion, but also improves the accuracy and robustness of concentration inversion by correcting environmental interference through neural networks.

[0058] In one embodiment of the present invention, the neural network algorithm is a deep learning model, which is trained to segment the gas leak plume region from the fused video image to improve the accuracy of leak source localization in a disturbed environment.

[0059] In this technical solution, a deep learning model is used to segment the gas leak plume region (i.e., the cloud-like region formed by the diffusion of leaked gas in the air), effectively eliminating background interference and improving the accuracy of leak source location in complex environments.

[0060] In one embodiment of the present invention, the data processing module adopts an architecture that coordinates edge computing units and cloud servers. The edge computing units are configured to process fused video images in real time to perform gas type identification and gas concentration inversion, and the cloud servers are configured to store historical data and train and optimize algorithm models.

[0061] This technical solution employs an edge computing and cloud-based collaborative architecture, which separates real-time data processing from model optimization, ensuring both response speed and support for continuous algorithm optimization and historical data analysis.

[0062] In one embodiment of the present invention, the data processing module is configured to estimate the location and leakage rate of the leakage source based on fused video images and combined with environmental meteorological parameters through a positioning algorithm.

[0063] In this technical solution, the leakage source is located and the rate is estimated by combining environmental meteorological parameters, which improves the environmental adaptability and prediction accuracy of the location algorithm and is suitable for outdoor scenarios with variable weather conditions.

[0064] In one embodiment of the present invention, the gas types include: methane, ethyl cyanoacrylate, methanol, ethanol, xylene, ethylphenylallyl fluoride, ethane, propane, allyl chloride, allyl bromide, furan, butane, butanone, benzene, toluene, pentane, hexane, hydrogen sulfide, ammonia, and chlorine.

[0065] In this technical solution, the system can identify a variety of industrial harmful gases, covering both common and special gases, and has a wide range of applications.

[0066] In one embodiment of the present invention, the alarm module is configured to fuse the composition information of the leaked gas with spatial information to generate and display a gas concentration cloud map, a leak location marker, and a diffusion direction indicator superimposed on a visible light image on an electronic map.

[0067] In this technical solution, by overlaying information such as concentration cloud map, leakage location, and diffusion direction onto electronic map and visible light image, a highly visualized display of monitoring results is achieved, which facilitates quick understanding and response by operators.

[0068] In one embodiment of the present invention, the alarm module is provided with multiple alarm thresholds and is configured to trigger a field linkage response corresponding to the level when the gas concentration value in the gas composition information reaches or exceeds any level alarm threshold.

[0069] This technical solution sets up multi-level alarm thresholds and combines them with an on-site linkage response mechanism to achieve a graded response from early warning to emergency handling, thereby improving the system's emergency response capabilities and security.

[0070] According to a second aspect of the present invention, a method for monitoring gas leaks using the multi-channel broadband gas leak monitoring system described above is provided, comprising the following steps:

[0071] S1. The visible light image of the monitoring area is acquired through the data acquisition module, and the spectral images of the characteristic absorption bands of various leaked gases are obtained by multiple independent signal channels. The visible light image and the spectral image are then fused and superimposed to generate a fused video image of the monitoring area.

[0072] S2. The data processing module receives the fused video image and compares and analyzes it with the preset gas spectral image database to obtain the composition information of the leaked gas and the spatial information related to the gas leak.

[0073] S3. The alarm module performs a visual location and presentation operation and triggers an alarm in response to at least one of the leaked gas composition information and spatial information.

[0074] This invention also provides a complete monitoring method based on a multi-channel broadband gas leak monitoring system, which realizes fully automated monitoring from data acquisition and processing to alarm, and is easy to operate and responds quickly.

[0075] According to a third aspect of the present invention, a multi-channel broadband gas leak monitoring device is provided, comprising:

[0076] A processor configured to execute computer-executable instructions;

[0077] A memory that stores one or more computer-executable instructions that, when executed by a processor, implement the steps of the method described above.

[0078] The present invention also provides an integrated monitoring device that realizes the hardware implementation of the method described above through the collaboration of the processor and memory, and is suitable for embedded, mobile or fixed monitoring scenarios.

[0079] According to a fourth aspect of the invention, a computer-readable storage medium is provided having a computer program stored thereon, which can be executed by a processor to implement the various steps of the method described above.

[0080] The present invention also provides a computer-readable storage medium that enables the method described above to be deployed in various computing devices in software form, facilitating the promotion, upgrading and cross-platform application of the system.

[0081] The technical solution provided by this invention has the following advantages compared with the prior art:

[0082] (1) In terms of monitoring mode, the present invention breaks through the physical limitations of traditional point sensors or single-path monitoring technology. Existing technologies can only provide gas concentration information at discrete points or on a single line, resulting in a large number of monitoring blind spots. However, the present invention achieves panoramic and area-based continuous monitoring of a radius of hundreds of meters through optical imaging and multi-channel fusion technology. A single system can cover the area that traditional solutions require the deployment of a large number of devices to monitor, fundamentally eliminating monitoring blind spots and greatly improving the integrity and reliability of monitoring.

[0083] (2) Regarding multi-component identification capabilities, this invention overcomes the shortcomings of traditional equipment, which can typically only detect a single or a few types of gases. By covering a wide spectral range and combining multiple characteristic absorption band channels, the system of this invention can simultaneously identify and quantitatively analyze more than a dozen or even dozens of industrial hazardous gases with different properties, such as methane, VOCs, hydrogen sulfide, and ammonia. This "one-stop" multi-gas simultaneous monitoring capability can comprehensively address the risk of mixed gas leaks in complex industrial environments and avoid monitoring loopholes that may arise due to the single function of the equipment.

[0084] (3) In terms of leakage situation awareness and emergency response support, this invention has achieved a qualitative leap. Most traditional technologies only provide concentration alarms and cannot inform the specific location and diffusion of the leak. This invention integrates gas concentration information with visible light background images and electronic maps in real time, which can intuitively generate and display gas concentration cloud maps, accurately mark the location of the leak source, and dynamically simulate the diffusion path and trend of the gas. This visualized leakage situation presentation transforms abstract data alarms into intuitive spatial decision-making information, enabling emergency personnel to quickly locate and accurately handle the situation, and significantly shortening the accident response time.

[0085] (4) In terms of monitoring performance and environmental adaptability, this invention significantly improves the robustness of the system through multi-source information fusion and intelligent algorithms. The system of this invention can comprehensively utilize broadband infrared imaging and active laser detection, combined with hybrid algorithms such as partial least squares and neural networks to process multi-channel data, effectively suppressing the influence of environmental noise such as temperature changes, humidity interference, and background radiation, and achieving a wide dynamic range and high sensitivity monitoring from ppm level (parts per million) to ppt level (parts per trillion). This not only reduces the false alarm rate and false negative rate, but also enables the system to work stably and reliably in more complex industrial environments.

[0086] (5) In terms of system economy and scalability, the present invention, through its highly integrated design, allows one system to replace multiple independent traditional monitoring devices, significantly reducing the overall cost in terms of equipment procurement, installation and wiring, and subsequent maintenance. At the same time, the modular hardware design and edge-cloud collaborative intelligent architecture adopted by the system of the present invention give it good flexibility, upgradeability and self-evolution capabilities, enabling flexible configuration of functions according to actual needs, and continuous optimization and performance improvement through algorithms, effectively protecting the user's long-term investment. Attached Figure Description

[0087] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0088] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0089] Figure 1 A block diagram of a multi-channel broadband gas leak monitoring system provided in an embodiment of the present invention is shown;

[0090] Figure 2 It shows Figure 1 A block diagram of a specific embodiment of the system;

[0091] Figure 3 It shows Figure 1 A block diagram of another specific embodiment of the system;

[0092] Figure 4 It shows Figure 1 A block diagram of yet another specific embodiment of the system;

[0093] Figure 5 It shows the use of Figure 1A flowchart illustrating the system's method for monitoring gas leaks.

[0094] Among them, 10 is the data acquisition module; 100 is the visible light image acquisition module; 101 is the infrared signal acquisition module; 102 is the laser module; 20 is the data processing module; and 30 is the alarm module. Detailed Implementation

[0095] To better understand the above-mentioned objectives, features, and advantages of the present invention, embodiments of the present invention will be further described below. It should be noted that, unless otherwise specified, embodiments of the present invention and features thereof can be combined with each other.

[0096] Numerous specific details are set forth in the following description in order to provide a full understanding of the invention, but the invention may also be practiced in other ways than those described herein; obviously, the embodiments in the specification are only some, and not all, of the embodiments of the invention.

[0097] This invention provides a multi-channel broadband gas leak monitoring system, method, apparatus, and medium, aiming to solve problems such as limited monitoring range, difficulty in identifying multiple components, inaccurate leak source location, and high system cost in existing technologies. The technical solution of this invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0098] Overall Architecture of Multi-channel Broad Spectrum Gas Leakage Monitoring System

[0099] Figure 1 A block diagram of a multi-channel broadband gas leak monitoring system provided in an embodiment of the present invention is shown. Figure 1 As shown, the multi-channel broadband gas leak monitoring system provided in this embodiment of the invention mainly includes a data acquisition module 10, a data processing module 20, and an alarm module 30.

[0100] The data acquisition module 10 is used to acquire visible light images of the monitoring area and spectral images of the characteristic absorption bands of various leaked gases. It acquires multidimensional spectral information through multiple independent signal channels (such as different infrared band channels and laser channels), and fuses the visible light images with the multi-band spectral images to generate a fused video image that includes spatial background and gas spectral information.

[0101] The data processing module 20 is communicatively connected to the data acquisition module 10. The data processing module 20 receives fused video images. This module has a built-in gas spectral image database and integrates intelligent analysis algorithms. Its core function is to compare and analyze the fused video images with the gas spectral image database to obtain information about the composition of the leaked gas (especially identifying the type of leaked gas and retrieving its concentration), and to combine this information with the image data to determine the spatial information such as the location and direction of gas leakage.

[0102] The alarm module 30 is communicatively connected to the data processing module 20. Based on the leaked gas composition information (type, concentration) and / or spatial information (location, diffusion direction) output by the data processing module 20, the alarm module 30 performs visual positioning and presentation on an electronic map or video screen (such as gas concentration cloud map, leak point marking, etc.), and triggers corresponding alarms and on-site linkage responses according to preset multi-level alarm thresholds.

[0103] Detailed Composition and Functions of Each Module in a Multi-channel Broadband Gas Leakage Monitoring System

[0104] (1) Data acquisition module 10

[0105] like Figure 2-4 As shown, the data acquisition module 10 has a variety of optional hardware configurations to adapt to different cost, accuracy, and application scenarios. Its core lies in constructing multiple independent spectral signal acquisition channels. The data acquisition module 10 includes submodule a and submodule b (the number can be multiple) and an optional submodule c. It must be ensured that the number of submodules in the data acquisition module 10 is greater than or equal to three.

[0106] Submodule a: Visible light image acquisition module 100. The visible light image acquisition module 100 is typically implemented using an industrial camera or industrial camera system. It is responsible for acquiring real-time visible light images of the monitored area, providing a background reference for the visualization, location, and overlay display of gas leaks. The industrial camera or industrial camera system must possess certain environmental adaptability, such as dustproof, waterproof, wide dynamic range, and electromagnetic interference resistance.

[0107] Submodule b: Infrared signal acquisition module 101. The infrared signal acquisition module 101 constitutes the most important spectral acquisition channel of the system. Its core component is an infrared camera, preferably an infrared focal plane detector, to achieve two-dimensional imaging. By configuring different wavelength filters for different cameras or directly coating the detector, their operating wavelengths can be locked, thus forming multiple independent infrared spectral channels. Common operating wavelength configurations include:

[0108] Long-wave infrared module (can be cooled or uncooled detector): The operating band is usually 6-14μm, suitable for detecting the radiation of objects at room temperature and the characteristic absorption of many VOCs, ammonia and other gases.

[0109] Mid-wave infrared module (can be cooled or uncooled detector): The operating wavelength is usually 3-5μm, and it has higher detection sensitivity for high-temperature gases or some gases that have absorption peaks in short-wave infrared (such as methane).

[0110] Each infrared signal acquisition module 101 is an infrared camera. The infrared signal acquisition module 101 may be a combination of at least one long-wave infrared module and at least one mid-wave infrared module, or a combination of at least two long-wave infrared modules.

[0111] For example, when the infrared signal acquisition module 101 is a combination of a long-wave infrared module and a mid-wave infrared module, the infrared signal acquisition module 101 may include a long-wave infrared module with an operating wavelength of 6-14μm and a mid-wave infrared module with an operating wavelength of 3-5μm. Alternatively, the operating wavelength of the long-wave infrared module may be any sub-band within 6-14μm (e.g., 6-8μm, 8-9μm, 9-10μm, 10-12μm, 12-14μm, etc.), and the operating wavelength of the mid-wave infrared module may be any sub-band within 3-5μm (e.g., 3-3.8μm and 3.8-5μm, or 3-3.4μm, 3.4-4.2μm and 4.2-5μm, etc.).

[0112] When the infrared signal acquisition module 101 is a combination of at least two long-wave infrared modules, the combination of at least two long-wave infrared modules includes any number (e.g., two) of long-wave infrared modules with operating wavelengths within the range of 6-14 μm. For example, a long-wave infrared module with an operating wavelength of 6-8 μm and a long-wave infrared module with an operating wavelength of 8-14 μm. Alternatively, the combination of long-wave infrared modules can also be, for example, a combination of four independent infrared modules with wavelengths of 6-7 μm, 7-8.5 μm, 8.5-10 μm, and 10.5-14 μm, or a combination of four independent infrared modules with wavelengths of 6.5-8 μm, 8-9.5 μm, 9.5-11.5 μm, and 11.5-13 μm, and so on.

[0113] Laser module 102 (optional): As a supplementary spectral channel, it constitutes an independent signal channel, forming a multi-signal channel together with infrared signal acquisition module 101 and visible light image acquisition module 100. Laser module 102 emits modulated laser light into the monitoring area to identify gases that the infrared signal acquisition module 101 cannot identify or has difficulty identifying. Laser module 102 emits modulated laser light into the monitoring area, utilizing the absorption of specific wavelength laser light by gas molecules to detect the gas. It is particularly suitable for detecting trace gases and specialty gases with weak or indistinguishable infrared signals. Laser module 102 can be a quantum cascade laser (QCL) array, which has a wide wavelength tuning range, high accuracy, and high power. Furthermore, the process of laser module 102 monitoring gas is as follows: the laser emission frequency is modulated to the characteristic absorption frequency of the target leaking gas, and a corresponding modulated laser beam is emitted into the monitoring area; the return signal intensity of the modulated laser beam is monitored by a receiver; in response to the return signal intensity being lower than a preset signal intensity threshold, it is determined that a target leaking gas exists in the monitoring area and a cross-shaped positioning mark is formed. For example, when monitoring methane gas, the laser module 102 modulates the laser emission frequency to the characteristic absorption frequency of methane gas, emitting a corresponding modulated laser beam towards the monitoring area. If the receiver does not receive a return signal, it determines that there is a leaking methane gas in that area and forms a cross-shaped location marker. Furthermore, after the laser module 102 determines the presence of a target leaking gas (e.g., methane), the entire system can retrieve a pre-made gas cloud image of the target leaking gas (e.g., methane) from the background via software, and overlay and fuse this methane gas cloud image with the cross-shaped location marker for visualization and alarm activation.

[0114] Reference Figure 2 As shown, in a preferred embodiment of this application, the data acquisition module 10 includes a visible light image acquisition module 100 and an infrared signal acquisition module 101. The infrared signal acquisition module 101 comprises at least two infrared signal acquisition modules, forming multiple independent signal channels. For example, each infrared signal acquisition module 101 is an infrared camera; therefore, the infrared signal acquisition module 101 includes multiple infrared cameras, such as infrared camera 1, infrared camera 2, ..., infrared camera n. Each infrared camera operates in a different wavelength band and is configured to acquire spectral images of the characteristic absorption bands corresponding to various leaked gases within its respective operating wavelength band. In this embodiment, the specific combination of the data acquisition module 10 is a visible light image acquisition module 100 + at least two infrared signal acquisition modules 101 with different wavelength bands.

[0115] Reference Figure 3As shown, in a preferred embodiment of this application, the data acquisition module 10 includes a visible light image acquisition module 100, an infrared signal acquisition module 101, and a laser module 102. The infrared signal acquisition module 101 comprises at least two infrared signal acquisition modules, forming multiple independent signal channels. For example, each infrared signal acquisition module 101 is an infrared camera; therefore, the infrared signal acquisition module 101 includes multiple infrared cameras, such as infrared camera 1, infrared camera 2, ..., infrared camera n. Each infrared camera operates in a different wavelength band and is configured to acquire spectral images of the characteristic absorption bands corresponding to various leaked gases within its respective operating wavelength band. The laser module 102 constitutes an independent signal channel, forming multiple signal channels together with the infrared signal acquisition module 101 and the visible light image acquisition module 100, and emits modulated laser light into the monitoring area to identify gases that the infrared signal acquisition module 101 cannot identify or finds difficult to identify. In this embodiment, the specific combination of the data acquisition module 10 is a visible light image acquisition module 100 + at least two infrared signal acquisition modules 101 with different wavelength bands + a laser module 102.

[0116] Reference Figure 4 As shown, in a preferred embodiment of this application, the data acquisition module 10 includes a visible light image acquisition module 100, an infrared signal acquisition module 101, and a laser module 102. The infrared signal acquisition module 101 is a single module, such as a single infrared camera. The single infrared camera acquires the spectral image of the characteristic absorption band corresponding to the leaked gas under the corresponding operating wavelength. The laser module 102 constitutes an independent signal channel, forming multiple independent signal channels together with the single infrared signal acquisition module. It emits modulated laser light into the monitoring area to identify gases that the single infrared signal acquisition module cannot identify or has difficulty identifying, thereby cooperating with the single infrared signal acquisition module to acquire spectral images of the characteristic absorption bands of different gases, i.e., acquiring spectral images of the characteristic absorption bands corresponding to various leaked gases. In this embodiment, the specific combination of the data acquisition module 10 is a visible light image acquisition module 100 + a single infrared signal acquisition module 101 + a laser module 102.

[0117] (2) Data processing module 20

[0118] This module is the system's "intelligent brain," responsible for information fusion, analysis, and decision-making. The data processing module 20 includes the following architecture:

[0119] Core Database: Built-in laboratory-calibrated gas spectral image database, storing standard spectral characteristics or "fingerprint" images of various target gases under different concentrations, temperatures, and humidity conditions.

[0120] Algorithm Engine: Employs hybrid intelligent algorithms for data processing.

[0121] Gas identification and concentration inversion: First, algorithms such as partial least squares (PLS) are used to quickly process multi-channel spectral data to initially invert gas concentration. Then, neural network algorithms (such as shallow networks or deep learning models) are used to correct the results to compensate for errors caused by environmental interference (temperature, humidity, background radiation), significantly improving the accuracy and robustness of concentration inversion.

[0122] Leakage source localization and plume segmentation: In complex backgrounds, a trained deep learning model (such as one based on ResNet or DETR architecture) is used to intelligently segment the gas leak plume region (i.e., the cloud-like region formed by the diffusion of leaked gas in the air) in the fused video images, eliminating background interference from equipment, buildings, etc., thereby more accurately determining the leak range and locating the leak source. Combined with environmental meteorological parameters (wind speed, wind direction), the leak rate can be further estimated and the diffusion path predicted.

[0123] Computing Architecture: An edge-cloud collaborative architecture is adopted. Edge computing units (such as embedded devices based on FPGAs and GPUs) are responsible for the rapid processing of real-time data, processing fused video images from data acquisition module 10 in real time for gas type identification and gas concentration inversion. The cloud server is responsible for storing massive amounts of historical data, training and optimizing deep learning models, and performing macro-level data analysis and report generation.

[0124] (3) Alarm module 30

[0125] This module enables intuitive presentation of monitoring results and emergency response coordination.

[0126] Visualization: The gas type and concentration information output by the data processing module 20 is displayed in real time on an electronic map and a visible light background video in the form of a pseudo-color "concentration cloud map". At the same time, the location of the leak source is clearly marked, and the estimated direction of gas diffusion is indicated by arrows and other means.

[0127] Tiered alarm and linkage: Preset multi-level alarm thresholds (such as early warning, alarm, and emergency alarm). When the gas concentration reaches different levels, the system triggers the corresponding level of response: from software pop-ups and on-site audible and visual alarms to automatic linkage with the emergency system (such as starting the fan, closing the valve, and pushing information to the management terminal), realizing an automated closed loop from early warning to emergency response.

[0128] Workflows of three different system architectures

[0129] (1) Visible light + multi-infrared multi-channel gas leak monitoring system

[0130] Refer again Figure 2 As shown, the Figure 2The basic multi-channel broadband gas leak monitoring system architecture, combining visible light and multiple infrared sensors, was demonstrated. The core system comprises three main modules, with the following hierarchy and components:

[0131] Data acquisition module 10: As the front-end sensing core, it includes sub-module a (visible light image acquisition module 100): composed of an industrial camera, responsible for acquiring visible light background images of the monitoring area; sub-module b (infrared signal acquisition module 101): there are at least two of them, that is, each infrared signal acquisition module 101 is an independent infrared camera, composed of at least two infrared cameras (infrared camera 1, infrared camera 2... infrared camera n), each infrared camera is equipped with different filter or detector coating, and has different working wavelengths (such as mid-wave 3-5μm, long-wave 6-14μm, etc.), forming multiple independent spectral channels, specifically capturing the characteristic absorption band spectral images of different gases.

[0132] Data processing module 20: As the "intelligent hub" of the system, it receives image data transmitted by data acquisition module 10, has a built-in gas spectral image database, and integrates the core processing flow of "visible light infrared image enhancement → gas target segmentation → gas type identification". Through algorithms, it realizes gas type identification and concentration inversion, and determines the gas leakage location and diffusion direction.

[0133] Alarm module 30: As a terminal response unit, based on the data processing results, it provides at least four types of alarm functions, namely concentration alarm, location alarm, type alarm, diffusion direction alarm, etc., to realize multi-dimensional warning of leakage information.

[0134] In this system architecture, the industrial camera of the data acquisition module 10 and multiple infrared cameras are aligned with the same monitoring area, simultaneously capturing visible light background images and multiple infrared spectral images of different bands (reflecting gas absorption intensity). The data acquisition module 10 fuses the multi-source images, overlaying the multi-band infrared spectral images with the visible light background image to generate a fused video image that combines intuitive on-site scene information with multispectral gas information. The fused video image is transmitted to the data processing module 20 in real time. The data processing module 20 calls its built-in gas spectral image database to compare the multi-band spectral features in the fused image, identifies the gas type and inverts the gas concentration using algorithms such as partial least squares (PLS), and determines the location and diffusion direction of the gas leak using image processing technology. The alarm module 30 receives the analysis results, overlays the gas concentration as a pseudo-color "concentration cloud map" onto the electronic map and the visible light image, accurately marks the leak point, and triggers the corresponding level of audible and visual alarm or on-site linkage response.

[0135] (2) Visible light + multiple infrared + laser multi-channel gas leak monitoring system

[0136] Refer again Figure 3As shown, the Figure 3 The system architecture, which integrates visible light, multiple infrared sensors, and laser, was demonstrated. It integrates all core sensing channels to achieve wide coverage, high sensitivity, and high precision monitoring. The architecture consists of the following components:

[0137] Data acquisition module 10 includes sub-module a (visible light image acquisition module 100): an industrial camera that provides a clear background; sub-module b (infrared signal acquisition module 101): two or more infrared modules in different wavelength bands (such as mid-wave 3-5μm + long-wave 6-14μm) to form multiple independent signal channels; and sub-module c (laser module 102): a quantum cascade laser array that also forms an independent signal channel, combining high-sensitivity detection and precise positioning functions.

[0138] Data processing module 20: As the system's "intelligent hub", it receives image data transmitted by the data acquisition module, has a built-in gas spectral image database, and integrates the core processing flow of "visible light infrared image enhancement → gas target segmentation → gas type identification". Through algorithms, it realizes gas type identification and concentration inversion, and determines the gas leakage location and diffusion direction.

[0139] Alarm module 30: As a terminal response unit, based on the data processing results, it provides at least four types of alarm functions, namely concentration alarm, location alarm, type alarm, diffusion direction alarm, etc., to realize multi-dimensional warning of leakage information.

[0140] In this system, the visible light image acquisition module 100, the multi-band infrared signal acquisition module 101, and the laser module 102 work together to acquire visible light field images, multi-band infrared spectral images, laser absorption detection signals, laser crosshair positioning markers, etc. The data acquisition module 10 performs complex fusion of multi-source data, superimposing wide-range gas information from multi-band infrared and laser spatial calibration information onto the visible light background to generate a fused video image with extremely high information dimensions. The fused video image is transmitted to the data processing module 20 in real time. The data processing module 20 calls the built-in gas spectral image database to compare the multi-band spectral features in the fused image, identifies the gas type and inverts the gas concentration using algorithms such as partial least squares (PLS), and determines the location and diffusion direction of the gas leak by combining image processing technology. The alarm module 30 receives the analysis results, superimposes the gas concentration as a pseudo-color "concentration cloud map" onto the electronic map and the visible light image, accurately marks the leak point, and triggers the corresponding level of audible and visual alarm or on-site linkage response.

[0141] (3) Visible light + single infrared + laser multi-channel gas leak monitoring system

[0142] Refer again Figure 4 As shown, the Figure 4The system architecture, which combines visible light, single infrared, and laser, was demonstrated to enhance the detection capabilities of specialty and trace gases. The architecture consists of the following components:

[0143] Data acquisition module 10 includes sub-module a (visible light image acquisition module 100): an industrial camera that acquires visible light background images; sub-module b (infrared signal acquisition module 101): a single general-purpose infrared camera (e.g., long-wavelength 6-14μm) that covers the characteristic absorption bands of most common industrial gases; and sub-module c (laser module 102): composed of a laser (preferably a quantum cascade laser array) that serves as an independent signal channel, capable of emitting modulated lasers or laser beams and forming positioning markers (e.g., cross-shaped) to identify gases that the infrared signal acquisition module 101 cannot or has difficulty identifying.

[0144] Data processing module 20: As the system's "intelligent hub", it receives image data transmitted by the data acquisition module, has a built-in gas spectral image database, and integrates the core processing flow of "visible light infrared image enhancement → gas target segmentation → gas type identification". Through algorithms, it realizes gas type identification and concentration inversion, and determines the gas leakage location and diffusion direction.

[0145] Alarm module 30: As a terminal response unit, based on the data processing results, it provides at least four types of alarm functions, namely concentration alarm, location alarm, type alarm, diffusion direction alarm, etc., to realize multi-dimensional warning of leakage information.

[0146] In this system, the visible light image acquisition module 100 (industrial camera) and a single infrared signal acquisition module 101 (single infrared camera) work synchronously to acquire visible light background images and single-band infrared spectral images. The laser module 102 emits a modulated laser beam of a specific wavelength towards the monitoring area, forming a cross-shaped positioning marker for spatial calibration. The data acquisition module 10 fuses the visible light image, the single-band infrared spectral image, and the laser spatial calibration information to generate a fused video image with richer information dimensions, compensating for the shortcomings of a single infrared channel in gas type coverage. The fused video image is transmitted to the data processing module 20 in real time. The data processing module 20 calls upon its built-in gas spectral image database to compare the multi-band (infrared + laser) spectral features in the fused image, identifies the gas type and retrieves the gas concentration using algorithms such as partial least squares (PLS), and simultaneously determines the location and diffusion direction of gas leaks using image processing techniques. The alarm module 30 receives the analysis results, overlays the gas concentration in the form of a pseudo-color "concentration cloud map" onto the electronic map and visible light image, accurately marks the leak point, and triggers the corresponding level of audible and visual alarm or on-site linkage response.

[0147] In the above embodiments of the present invention, the gas types include: methane, ethyl cyanoacrylate, methanol, ethanol, xylene, ethylphenyl allyl fluoride, ethane, propane, allyl chloride, allyl bromide, furan, butane, butanone, benzene, toluene, pentane, hexane, hydrogen sulfide, ammonia, and chlorine. The gas types mentioned are not limited to the specific types described above.

[0148] Multi-channel broadband gas leak monitoring method

[0149] like Figure 5 As shown, Figure 5 This paper demonstrates a complete methodology for gas monitoring using a multi-channel broadband gas leak detection system, comprising the following three steps:

[0150] S1. The visible light image of the monitoring area is acquired through the data acquisition module, and the spectral images of the characteristic absorption bands of various leaked gases are obtained by relying on multiple independent signal channels. The visible light image and the spectral image are then fused and superimposed to generate a fused video image of the monitoring area.

[0151] In step S1, the data acquisition module 10 adopts... Figure 2 The architecture fuses multi-infrared spectral images with visible light images. The data acquisition module adopts... Figure 3 The architecture fuses multi-infrared spectral images, laser signal spectral images, and visible light images. The data acquisition module adopts... Figure 4 The architecture fuses individual infrared spectral images, laser signal spectral images, and visible light images.

[0152] S2. The data processing module receives the fused video image and compares and analyzes it with a preset gas spectral image database to obtain information on the composition of the leaked gas (such as gas type and concentration) and spatial information related to the gas leak (such as leak location and diffusion direction).

[0153] In step S2, a partial least squares algorithm is used for rapid concentration inversion. Neural network algorithms (such as deep learning models) are combined for interference compensation and plume segmentation. Environmental meteorological parameters (wind speed, wind direction) can be integrated for leak source location and diffusion prediction.

[0154] S3. The alarm module performs a visualization and display operation and triggers an alarm in response to at least one of the leaked gas composition information and the spatial information.

[0155] In step S3, a gas concentration cloud map is overlaid on the electronic map, the location of the leak source is marked, and the direction of gas diffusion is indicated. Preset multi-level alarm thresholds trigger corresponding levels of on-site linkage responses (such as audible and visual alarms, starting fans, and closing valves).

[0156] In addition, embodiments of the present invention also relate to a multi-channel broadband gas leak monitoring device, comprising: a processor configured to execute computer-executable instructions; and a memory storing one or more computer-executable instructions, which, when executed by the processor, implement steps S1-S3 of the method described above.

[0157] Furthermore, embodiments of the present invention also relate to a computer-readable storage medium having a computer program stored thereon, the computer program being executable by a processor to implement steps S1-S3 of the method described above.

[0158] Example 1 – General-purpose monitoring system for conventional industrial parks

[0159] This embodiment 1 is designed for large-area conventional monitoring scenarios such as petroleum and chemical industries. It adopts a three-module fusion scheme without laser modules, with cost controllability and system stability as the core, to meet the daily leakage monitoring needs of industrial parks and is suitable for medium and low concentration gas leakage detection scenarios.

[0160] 1. System composition and parameter configuration

[0161] The data acquisition module 10 adopts a three-module fusion architecture consisting of a long-wave (6-14μm) uncooled infrared module, a mid-wave (3-5μm) cooled infrared module, and a visible light module. The visible light image acquisition module 100 uses an industrial high-definition camera with a resolution of 1920×1080, a frame rate of 25fps, an 8-50mm zoom lens, and IP65 protection (dustproof and waterproof) and electromagnetic interference resistance, used to acquire real-time scene images of the monitored area. The infrared signal acquisition module 101 uses two infrared cameras: the long-wave uncooled infrared module uses a 640×512 resolution infrared focal plane detector with a pixel pitch of 17μm, a noise equivalent temperature difference ≤50mK, requires no cooling device, and is suitable for continuous all-weather operation; the mid-wave cooled infrared module uses a Stirling cooler, a detector resolution of 1280×1024, a pixel pitch of 10μm, a noise equivalent temperature difference ≤20mK, and can capture weak spectral signals from low-concentration gases. Both modules lock in their operating wavelengths through custom filter coatings to avoid spectral cross-interference.

[0162] The data processing module 20 is based on an FPGA architecture for integrated processing. It has a built-in spectral database covering more than 20 common industrial hazardous gases, including methane, methanol, ethanol, benzene, toluene, hydrogen sulfide, and ammonia. It also establishes a model of the correspondence between image grayscale and gas concentration based on laboratory calibration. The algorithm adopts a hybrid method combining PLS (partial least squares) and shallow neural networks. First, PLS is used to quickly invert the gas concentration, and then the neural network is used to correct errors caused by environmental interferences such as temperature and humidity, so as to achieve accurate concentration calculation.

[0163] The alarm module integrates the park's electronic map, supports GIS positioning, and converts gas concentration data into a pseudo-color concentration cloud map, which is then overlaid on a visible light image to visualize the leak area. It sets three alarm thresholds (low concentration warning, medium concentration alarm, and high concentration emergency alarm). A pop-up notification appears when the concentration is low; medium concentration triggers an on-site audible and visual alarm; and high concentration triggers the park's emergency system, pushing the leak location, gas type, and diffusion direction to the management terminal and automatically marking emergency facilities within a 50-meter radius of the leak source.

[0164] 2. Workflow

[0165] After system startup, the three modules of data acquisition module 10 operate synchronously. The visible light image acquisition module acquires scene images, while the dual-band infrared module acquires spectral signals in the 6-14μm and 3-5μm bands respectively. The data is transmitted to the FPGA processing unit at a rate of 100Mbps. The processing unit of data processing module 20 performs preprocessing on the infrared signals, including noise reduction, filtering, and band alignment. It compares the dual-band spectral data with the built-in database to determine the gas type and uses a hybrid algorithm to invert the gas concentration. It further combines the visible light image to achieve spatial positioning of the leak area and generates a two-dimensional imaging map containing concentration information. The alarm module 30 determines in real time whether the concentration exceeds the threshold. If it does, it triggers an alarm, marks the GPS coordinates of the leak point on the electronic map, generates a diffusion trend curve based on historical meteorological data of the park, and executes corresponding linkages according to the alarm level. The processing unit generates a monitoring log every hour, including the average concentration of each area, abnormal records, etc., and synchronizes it to the cloud for archiving.

[0166] 3. Adapted scenarios and performance indicators

[0167] This embodiment is suitable for open areas such as large chemical industrial parks and oil reserves, with a monitoring radius of up to 500 meters and no blind spots. The system can simultaneously identify 17 target gases, with detection limits down to the ppm level (e.g., methane 0.5 ppm, hydrogen sulfide 1 ppm), a positioning error of ≤3 meters, and a data processing delay of ≤200ms. Compared to multiple independent OP-FTIR systems, the deployment cost is reduced by approximately 40%, and with no easily damaged sensors, the annual maintenance cost is only about 15% of that of existing point sensor systems.

[0168] Example 2 – High-Precision Special Scene Monitoring System

[0169] This embodiment is designed for scenarios requiring high precision and multi-component identification, such as pharmaceutical plants and fine chemical workshops. It integrates quantum cascade laser modules and deep learning algorithms, making it suitable for trace gas leak detection and precise positioning in complex environments.

[0170] 1. System composition and parameter configuration

[0171] The data acquisition module 10 adopts a four-module fusion architecture consisting of a long-wavelength (8-14μm) uncooled infrared module, a long-wavelength (6-8μm) uncooled infrared module, a visible light module, and a quantum cascade laser array module. The visible light image acquisition module 100 uses a 2K high-definition industrial camera (2560×1440 resolution), with a frame rate of 30fps, equipped with an anti-fog lens and supporting low-light enhancement. The infrared signal acquisition module 101 consists of two long-wavelength uncooled infrared modules, achieving band differentiation (6-8μm and 8-14μm) through customized coating. Both have a resolution of 1280×720, a pixel pitch of 12μm, a noise equivalent temperature difference ≤30mK, and support for multi-frame stacking noise reduction. The laser module 102 adopts a quantum cascade laser array module, which uses a 4-channel quantum cascade laser array with a laser power of ≥50mW per channel and a wavelength tuning accuracy of 0.01μm. The laser module 102 integrates a photoacoustic resonant cavity (volume 0.5ml), which amplifies weak signals through light-heat-sound conversion to improve the sensitivity of trace gas detection.

[0172] The data processing module 20 adopts an "edge computing + cloud collaboration" architecture. Real-time processing is achieved at the edge using GPUs, while the cloud is used for model training and data storage. It includes an expanded gas spectral database, adding spectral samples of special gases such as ethyl cyanoacrylate, allyl chloride, and furan. An analysis model is constructed by combining multi-dimensional parameters such as temperature, humidity, and wind speed. The algorithm incorporates a DETR deep learning model based on the ResNet-50 feature extractor. It first segments the leak plume region, then inverts the concentration using multi-channel spectral data, and optimizes the localization bounding box using the Hungarian algorithm.

[0173] The alarm module 30 features high-precision visualization, supporting dynamic refreshing of the concentration cloud map (refresh rate 10Hz), and can magnify to display leakage details in an area of ​​0.1 square meters. The alarm module 30 is linked to the workshop process control system, automatically triggering equipment shutdown and ventilation system activation in the event of a high-concentration leak. It integrates a leakage diffusion prediction model, combining real-time meteorological data to simulate the gas diffusion range over the next 10 minutes and generate emergency evacuation route suggestions.

[0174] 2. Workflow

[0175] After system startup, the four modules of data acquisition module 10 simultaneously acquire data. A dual-band infrared camera acquires the basic spectral signal, and a quantum cascade laser array emits modulated laser light, which is converted into an electrical signal by a photoacoustic resonator and transmitted along with the visible light image to the edge GPU processing unit of data processing module 20. The edge unit performs spatiotemporal alignment and preprocessing, uses the DETR model to segment the leak plume and eliminate background interference, and combines the quantum cascade laser-enhanced signal with a deep learning algorithm to invert trace gas concentrations and locate the leak source coordinates. Real-time data is uploaded to the cloud, where historical data and meteorological parameters are combined to optimize the diffusion prediction model and update the concentration cloud map and diffusion path on the electronic map. Alarm module 30 triggers an alarm based on the concentration threshold. Even if a ppt-level concentration of special gases (such as ethyl cyanoacrylate and allyl bromide) is detected, an immediate warning is issued, and a visualized monitoring report (including information such as gas type, concentration change curve, and leak source location accuracy) is generated.

[0176] 3. Adapted scenarios and performance indicators

[0177] This embodiment is applicable to scenarios such as pharmaceutical plants, fine chemical workshops, and specialty gas storage facilities. It can penetrate equipment obstructions to achieve monitoring, with a monitoring radius of up to 300 meters. It can simultaneously identify more than 30 gases, with detection limits as low as ppt (e.g., ethyl cyanoacrylate 50 ppt, allyl chloride 100 ppt), positioning error ≤ 1 meter, and data processing delay ≤ 300 ms.

[0178] Example 3 – Single Long-Wave Infrared and Laser Combined Monitoring System

[0179] This embodiment 3 is applicable to semi-open scenarios where hydrocarbons are the primary monitoring target, such as oil tank areas, natural gas pressure regulating stations, and LNG loading and unloading areas. The data acquisition module 10 in this system adopts a three-module architecture: a single long-wave infrared module, a visible light module, and a laser module. The laser module identifies gases that cannot be detected or are difficult to detect by a single long-wave infrared module, and laser positioning markers are used to quickly pinpoint leak sources.

[0180] 1. System composition and parameter configuration

[0181] The data acquisition module 10 consists of three sub-modules: a visible light image acquisition module 100, employing a 2-megapixel low-light industrial camera with a frame rate of 25 fps, featuring strong light suppression and fog penetration capabilities, providing clear background images at night and in adverse weather conditions; an infrared signal acquisition module 101, a single long-wavelength uncooled infrared module with a working wavelength of 6-14μm, a detector resolution of 640×512, a noise equivalent temperature difference ≤40 mK, and a pixel pitch of 17 μm; and a laser module 102, employing a tunable semiconductor laser with wavelengths covering the 1.5-2.0 μm or 3-5 μm range, enhancing the detection of gases such as methane or hydrogen sulfide, and identifying gases that a single long-wavelength infrared module cannot detect or finds difficult to detect. The laser output, after collimation and beam expansion, forms a fan-shaped scanning beam, which can be superimposed on the image to generate a cross-shaped positioning mark for assisting in spatial alignment and leak point calibration.

[0182] The data processing module 20 is based on an embedded AI computing platform and has a built-in characteristic absorption spectrum library for more than 10 gases, including methane, ethane, propane, and benzene series compounds. It uses wavelength modulation spectroscopy to process laser signals, achieving ppt-level detection sensitivity; it performs real-time non-uniformity correction and temperature radiation calibration on infrared images, and retrieves gas concentrations through a gray-scale-concentration mapping model. Algorithm-wise, it employs Kalman filtering to fuse laser point scan concentration and infrared array imaging data, improving the accuracy of leak source location.

[0183] The alarm module 30 supports real-time overlay of gas concentration contour lines on the video feed and marks the core leak point locked by laser with a flashing crosshair. The alarm module 30 has alarm levels, triggering corresponding audible and visual alarms based on concentration thresholds, and can push location information to the handheld terminal of the inspection personnel.

[0184] 2. Workflow

[0185] Simultaneously, the visible light image acquisition module 100, infrared signal acquisition module 101, and laser module 102 of the data acquisition module 10 are activated. The visible light image acquisition module 100 provides a clear background video; the infrared signal acquisition module 101 acquires the spatial distribution and preliminary concentration of gas over a large area through gas absorption characteristics; the laser module 102 performs high-sensitivity point scanning on specific gases (such as methane and hydrogen sulfide) and uses crosshairs to assist in locating the leak point. The data processing module 20 simultaneously receives the infrared image and laser scanning signal. The infrared image is corrected and calibrated, and the gas concentration is inverted using a grayscale-concentration model; the laser signal is analyzed with high sensitivity using wavelength modulation spectroscopy. Subsequently, the two-dimensional concentration distribution of the infrared imaging and the precise concentration data of the laser point are fused using a Kalman filter algorithm to improve the accuracy and reliability of leak source location. The processing results are transmitted to the alarm module 30 in real time. Gas concentration contour lines are superimposed on the visible light video screen, and the core location of the leak locked by the laser is marked with a flashing cross. At the same time, the software retrieves the gas cloud image of the specific gas monitored by the laser module 102 from the background. The gas cloud image is fused with the cross mark of the laser signal for visualization. The alarm level is automatically determined according to the concentration threshold, and the corresponding audible and visual alarm is triggered. At the same time, the information such as the leak location and concentration is pushed to the handheld terminal of the inspection personnel, completing the closed loop from monitoring to early warning.

[0186] 3. Adapted scenarios and performance indicators

[0187] This system is suitable for scenarios such as oil and gas storage tank areas, gas transmission stations, and chemical loading and unloading platforms, with a monitoring radius of approximately 200 meters. It can simultaneously detect more than 10 common hydrocarbons and hydrogen sulfide gases, with leak source location accuracy better than 3 meters. The system has a compact structure and can be installed on a pan-tilt unit or fixed bracket. It supports solar power supply and is suitable for areas without mains power. Compared to traditional systems, the cost is reduced by approximately 30%, making it particularly suitable for cost-effective monitoring and rapid emergency location of major hazardous gases such as methane.

[0188] Example 4 – Broadband Fusion Infrared Gas Monitoring System

[0189] This embodiment provides an infrared gas identification system with the ability to coordinate mid-wave and long-wave broadband monitoring. It is particularly suitable for complex industrial scenarios that require comprehensive and three-dimensional monitoring of various gases such as light hydrocarbons, toxic VOCs and acidic gases, such as large-scale integrated chemical industrial parks, national energy reserve bases or key environmental risk control areas.

[0190] 1. System composition and parameter configuration

[0191] This system adopts a modular and scalable architecture design. The core of the data acquisition module 10 includes:

[0192] Visible light image acquisition module 100: It adopts a 4K high-definition dual-spectrum PTZ camera, supports H.265+ encoding, and has functions such as fog penetration, strong light suppression and wide dynamic range, providing high-definition background images and all-weather scene perception capabilities.

[0193] Infrared signal acquisition module 101: This part is composed of multiple core multi-band fusion infrared modules working together to form a wide spectral coverage and high spectral resolution capability. The specific configuration is as follows: Mid-wave infrared multispectral module: Its core is a high-performance cooled mid-wave infrared focal plane detector (resolution 1024×768). Through high-speed filter wheel switching, the working wavelength coverage is 3-5μm, and it can be subdivided into multiple sub-bands for acquisition, such as: 3-3.4μm, 3.4-4.2μm, 4.2-5μm. These modules exhibit high sensitivity to light gases such as methane and carbon monoxide, as well as high-temperature targets. The long-wave infrared multispectral module, with its core being a high-sensitivity cooled long-wave infrared focal plane detector (resolution 1024×768), also employs a filter wheel switching mechanism. Its operating band covers the key gas absorption region of 6-14 μm and can be further subdivided into multiple sub-bands, such as 6-8 μm, 8-9.5 μm, 9.5-11.5 μm, and 11.5-14 μm. These modules demonstrate excellent detection performance for most VOCs, sulfides, ammonia, and other gases with strong absorption characteristics. The infrared signal acquisition module 101, through a combination of at least one mid-wave module and at least one long-wave module, or a combination of multiple long-wave sub-band modules, constructs multiple independent and complementary spectral signal channels, achieving broadband, multi-sub-band synchronous imaging detection from 3 μm to 14 μm.

[0194] Data Processing Module 20: Employs a high-real-time edge computing server equipped with a multi-core CPU and a high-performance GPU. Its core is a wide-band joint spectral processing engine and a unified gas spectral feature database. This database integrates high-precision absorption spectra and corresponding standard spectral images of dozens of gases from mid-wave to long-wave. The engine uses a hierarchical spectral unmixing algorithm: First, it performs independent preprocessing (non-uniformity correction, radiometric calibration) and preliminary gas identification on the data streams of each sub-band of mid-wave and long-wave; then, it performs cross-band joint analysis, utilizing the unique combination of absorption characteristics of different gases in the mid-wave and long-wave bands to solve the problem of distinguishing spectrally overlapping gases within a single spectral band, and simultaneously reconstructs the concentration distribution maps of all target gases.

[0195] Alarm Module 30: Features intelligent correlation alarm and risk situation fusion capabilities. It can not only display multiple gas concentration cloud maps from different modules on the same geographic information base map, layered and overlaid with different colors and transparency to achieve a comprehensive view of the entire area; it can also automatically calculate and dynamically display a comprehensive regional risk heat map based on the physicochemical properties of gases (toxicity, flammability, reactivity) and real-time concentrations. It supports correlation analysis of multiple gas leak events, comprehensive estimation of emission flux, and simulation and early warning of multi-pollutant coupled diffusion based on real-time meteorological data.

[0196] 2. Workflow

[0197] After system deployment, the modules work collaboratively, following this process: First, synchronous data acquisition is performed. The visible light module provides real-time panoramic video, while multiple infrared modules operate independently in parallel, each driving its own filter wheel to switch at high speed, acquiring infrared images in different sub-bands at a cycle of approximately 0.5 seconds, forming a synchronous but complementary "spatiotemporal-spectral" data cube. Then, hierarchical data processing is performed. Data from the mid-wave and long-wave modules enter the preprocessing channel for non-uniformity correction, bad pixel removal, and radiometric calibration. The sub-databases of their respective spectral bands are then used for the first round of spectral demixing, generating preliminary gas identification results and concentration distribution maps. The broadband joint calculation engine receives the preliminary results from both channels, performs high-precision spatiotemporal registration, and then uses the complete broadband database for joint inversion. The algorithm accurately removes cross-interference by comparing the absorption feature combinations of the same spatial location point in multiple sub-bands of mid-wave and long-wave, greatly improving the identification accuracy and quantitative reliability of components in complex gas mixtures. It ultimately generates a high-confidence concentration distribution dataset containing all target gases. This dataset is then deeply fused with high-resolution visible light images and geographic information to provide structured data for the alarm and visualization modules. Finally, intelligent alarm and situation presentation are implemented. The alarm module 30 makes multi-dimensional decisions based on the fused results. In addition to concentration threshold alarms for single gases, it can define and trigger composite event alarms (such as "methane leakage accompanied by abnormally high benzene concentration") and automatically initiate preset emergency analysis processes (such as diffusion simulation and source tracing analysis). All data, alarm events, and analysis results are stored in a database, supporting historical review, compliance report generation, and multi-dimensional analysis of emission trends.

[0198] 3. Adapted scenarios and performance indicators

[0199] This system is specifically designed for ultra-industrial scenarios with diverse gas types, intertwined risks, and extremely high regulatory requirements. Typical applications include: large-scale integrated petrochemical parks and refining and chemical integration bases, comprehensively monitoring gas leakage risks throughout the entire process from raw materials (natural gas, light hydrocarbons), intermediates (olefins) to products (various organic chemicals); national-level strategic energy reserves (LNG, crude oil) and surrounding environmentally sensitive areas, monitoring potential fugitive VOC emissions while preventing flammable and explosive gas leaks; and large-scale urban waste treatment centers (including incineration, landfill, and leachate treatment), simultaneously monitoring landfill gas (mainly methane) and malodorous, acidic gases and characteristic pollutants generated during the incineration process.

[0200] The effective monitoring distance for typical gases in this embodiment is: 300-1000 meters for medium-wave light gases and 150-500 meters for long-wave toxic / VOCs gases. This embodiment can simultaneously perform online quantitative analysis of more than 20 gas types, covering methane, ethane, benzene compounds, hydrogen sulfide, ammonia, and various VOCs. Through the fusion of medium-wave and long-wave broadband bands and multi-sub-band collaborative analysis, this embodiment improves the system's accuracy in identifying mixed gases with severe spectral interference and difficulty in distinguishing them in a single band by more than 15% compared to traditional single-band systems, achieving an overall identification accuracy of >93%. This embodiment provides a complete solution from real-time wide-area monitoring, intelligent multi-component identification, and precise leak location to risk situation fusion, coordinated early warning, and in-depth analysis. It is a cutting-edge technology and equipment for achieving coordinated control of organized and fugitive emissions in industrial sites and improving environmental safety early warning capabilities.

[0201] As can be seen from Embodiments 1-4 above, the multi-channel broadband gas leak monitoring system proposed in this application, through the deep integration of a modular, configurable data acquisition architecture and intelligent information processing technology, constructs an innovative monitoring system integrating "imaging-spectrum-location". The core of this system lies in the flexible construction of multiple independent spectral signal channels: it can either form a basic multispectral detection network by deploying at least two infrared modules with different operating bands (e.g., a combination of long-wave and mid-wave infrared, or two different long-wave infrared combinations); or it can adopt a compact architecture of "a single infrared module + laser module", utilizing the laser as an independent high-sensitivity, high-selectivity channel to effectively expand the detection range for specific or trace gases and identify gases that the infrared module cannot or is difficult to identify. Furthermore, this application can also achieve lightweight hyperspectral detection capabilities while maintaining the real-time performance of area array imaging through multiple sub-band detection sub-channels within a broadband band, significantly improving the accuracy of type identification and discrimination of complex mixed gases.

[0202] Furthermore, at the data processing level, the system of this invention generates fused video images rich in spatial and spectral information by deeply fusing visible light, multi-band infrared, and laser data. Relying on a built-in gas spectral image database and employing a hybrid intelligent algorithm combining partial least squares and neural networks, along with a deep learning plume segmentation model, the system achieves end-to-end intelligent analysis from gas type identification and concentration inversion to precise leak source location and diffusion trend prediction. Moreover, the system adopts an "edge-cloud" collaborative computing architecture, ensuring real-time response at the edge and supporting continuous algorithm optimization and deep data mining in the cloud. Combined with multi-level alarm thresholds and highly visual displays (such as overlaying concentration cloud maps, leak point markers, and diffusion directions on an electronic map), it forms a complete safety closed loop from real-time perception, intelligent diagnosis, precise location, to tiered early warning and coordinated response.

[0203] In summary, the system of this invention represents a fundamental shift from traditional point and line monitoring to wide-area surface monitoring. A single system can perform continuous, blind-spot-free panoramic imaging monitoring of an area with a radius of hundreds of meters, intuitively presenting the spatial distribution and movement of gas clouds, greatly improving the completeness and intuitiveness of monitoring. Simultaneously, the system possesses synchronous, rapid, and accurate multi-component gas identification capabilities, capable of simultaneously identifying and quantitatively analyzing more than ten or even dozens of industrial hazardous gases, effectively solving the problem of spectral interference in mixed gases. Furthermore, the system successfully unifies high-sensitivity detection with high spatial positioning accuracy, providing precise data for efficient emergency response. Moreover, through multi-source information fusion and intelligent environmental compensation algorithms, this invention demonstrates excellent adaptability to complex environments and anti-interference capabilities, ensuring stable and reliable operation in real industrial scenarios. Finally, the modular design and integrated architecture of this system, along with the evolvability capabilities brought by the "edge-cloud" architecture, provide users with a cost-effective, easy-to-maintain, and long-term upgrade potential overall solution, whose comprehensive cost-effectiveness far surpasses that of deploying multiple single-function traditional monitoring systems. Therefore, this application effectively solves the core pain points of "invisible, unrecognizable, inaccurate, and slow response" in industrial gas safety monitoring, and provides advanced and reliable technical means for intelligent and precise safety and environmental management.

[0204] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to the process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0205] The above descriptions are merely embodiments of this application, which enable those skilled in the art to understand and implement this application. Various modifications to the embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments described herein, but is to be accorded the widest scope consistent with the principles and features disclosed herein.

Claims

1. A multi-channel broadband gas leak monitoring system, characterized in that, include: The data acquisition module is configured to acquire visible light images of the monitoring area, obtain spectral images of the characteristic absorption bands of various leaked gases through multiple independent signal channels, and fuse and superimpose the visible light images and the spectral images to generate a fused video image of the monitoring area; wherein the data acquisition module includes at least one infrared signal acquisition module and a laser module, the laser module constitutes an independent signal channel, and together with the infrared signal acquisition module forms multiple independent signal channels, and the laser module is configured to emit a modulated laser beam into the monitoring area to identify leaked gases that the infrared signal acquisition module cannot identify or has difficulty identifying; The data processing module is communicatively connected to the data acquisition module. The data processing module is configured to receive the fused video image and compare it with a preset gas spectral image database to obtain information about the composition of the leaking gas and spatial information related to the gas leak. After the laser module determines the presence of a target leaking gas, the module retrieves a pre-made gas cloud image of the target leaking gas from the background via software and overlays and fuses the gas cloud image with a positioning marker for visualization. This assists in the comparison and analysis between the fused video image and the gas spectral image database to distinguish mixed gases with spectral interference within the monitoring area. An alarm module is communicatively connected to the data processing module. The alarm module is configured to perform a visual location and presentation operation and trigger an alarm in response to at least one of the leaked gas composition information and the spatial information.

2. The multi-channel broadband gas leak monitoring system according to claim 1, characterized in that, The data acquisition module includes: A visible light image acquisition module, configured to acquire visible light images of the monitored area; At least two infrared signal acquisition modules constitute multiple independent signal channels. Each infrared signal acquisition module operates in a different band and is configured to acquire infrared spectral images of the characteristic absorption bands corresponding to the leaked gas within its respective operating band range. The laser module, together with the at least two infrared signal acquisition modules and the visible light image acquisition module, can form a multi-signal channel.

3. The multi-channel broadband gas leak monitoring system according to claim 1, characterized in that, The data acquisition module includes: A visible light image acquisition module, configured to acquire visible light images of the monitored area; A single infrared signal acquisition module is configured to acquire an infrared spectral image of the characteristic absorption band corresponding to the leaked gas within a corresponding operating band range; The laser module works in conjunction with the single infrared signal acquisition module to acquire spectral images of the characteristic absorption bands corresponding to various leaked gases.

4. The multi-channel broadband gas leak monitoring system according to claim 2 or 3, characterized in that, The laser module is configured to perform the following operations: The laser emission frequency is modulated to the characteristic absorption frequency of the target leaking gas, and the corresponding modulated laser beam is emitted into the monitoring area. The intensity of the returned signal of the modulated laser beam is monitored by a receiver; In response to the return signal strength being lower than a preset signal strength threshold, it is determined that there is a target leaking gas in the monitoring area and a cross-shaped positioning mark is formed.

5. The multi-channel broadband gas leak monitoring system according to claim 4, characterized in that, The system is configured to overlay and fuse atmospheric cloud images with cross-shaped positioning markers for visualization and alarm purposes.

6. The multi-channel broadband gas leak monitoring system according to claim 4, characterized in that, The laser module includes a quantum cascaded laser array.

7. The multi-channel broadband gas leak monitoring system according to claim 2, characterized in that, The at least two infrared signal acquisition modules include any of the following combinations: (i) A combination of at least one long-wave infrared module and at least one mid-wave infrared module; (ii) A combination of at least two long-wave infrared modules.

8. The multi-channel broadband gas leak monitoring system according to claim 7, characterized in that, When the at least two infrared signal acquisition modules are in combination (i), the at least two infrared signal acquisition modules include a long-wave infrared module with a working wavelength of 6-14μm and a mid-wave infrared module with a working wavelength of 3-5μm.

9. The multi-channel broadband gas leak monitoring system according to claim 7, characterized in that, When the at least two infrared signal acquisition modules are in combination (i), the operating band of the long-wave infrared module is any sub-band within 6-14μm, and the operating band of the mid-wave infrared module is any sub-band within 3-5μm.

10. The multi-channel broadband gas leak monitoring system according to claim 7, characterized in that, When the at least two infrared signal acquisition modules are in combination (ii), the combination of the at least two long-wave infrared modules includes any number of non-overlapping sub-bands of long-wave infrared modules with a working wavelength of 6-14μm.

11. The multi-channel broadband gas leak monitoring system according to claim 10, characterized in that, The at least two infrared signal acquisition modules include a long-wave infrared module with a working wavelength of 6-8μm and a long-wave infrared module with a working wavelength of 8-14μm.

12. The multi-channel broadband gas leak monitoring system according to claim 3, characterized in that, The single infrared signal acquisition module includes a long-wave infrared module with a working wavelength of 6-14μm.

13. The multi-channel broadband gas leak monitoring system according to claim 2 or 3, characterized in that, The infrared signal acquisition module includes an infrared camera, and the detector of the infrared camera is an infrared focal plane detector.

14. The multi-channel broadband gas leak monitoring system according to claim 13, characterized in that, The operating band of the infrared signal acquisition module is achieved through filter coating or infrared focal plane detector coating.

15. The multi-channel broadband gas leak monitoring system according to claim 1, characterized in that, The leaked gas composition information includes gas type and gas concentration, and the spatial information includes gas leak location and gas leak diffusion direction.

16. The multi-channel broadband gas leak monitoring system according to claim 15, characterized in that, The data processing module has a built-in gas spectral image database and is configured to identify gas types and retrieve gas concentrations based on the fused video images and the gas spectral image database.

17. The multi-channel broadband gas leak monitoring system according to claim 16, characterized in that, The data processing module is configured to use a partial least squares algorithm and a neural network algorithm to process the fused video image to retrieve the gas concentration.

18. The multi-channel broadband gas leak monitoring system according to claim 17, characterized in that, The neural network algorithm is a deep learning model, which is trained to segment the gas leak plume region from the fused video image to improve the accuracy of leak source location in interference environments.

19. The multi-channel broadband gas leak monitoring system according to claim 15, characterized in that, The data processing module adopts an architecture that combines edge computing units and cloud servers. The edge computing units are configured to process the fused video images in real time to perform gas type identification and gas concentration inversion, while the cloud servers are configured to store historical data and train and optimize algorithm models.

20. The multi-channel broadband gas leak monitoring system according to claim 15, characterized in that, The data processing module is configured to estimate the location and leakage rate of the leakage source based on the fused video image and combined with environmental meteorological parameters, using a positioning algorithm.

21. The multi-channel broadband gas leak monitoring system according to claim 15, characterized in that, The gas types include: methane, ethyl cyanoacrylate, methanol, ethanol, xylene, ethylbenzene, allyl fluorine, ethane, propane, allyl chloride, allyl bromide, furan, butane, butanone, benzene, toluene, pentane, hexane, hydrogen sulfide, ammonia, and chlorine.

22. The multi-channel broadband gas leak monitoring system according to claim 1, characterized in that, The alarm module is configured to fuse the composition information of the leaked gas with the spatial information, and generate and display a gas concentration cloud map, a leak location marker, and a diffusion direction indicator superimposed on the visible light image on an electronic map.

23. The multi-channel broadband gas leak monitoring system according to claim 1, characterized in that, The alarm module is equipped with multiple alarm thresholds and is configured to trigger a corresponding on-site linkage response when the gas concentration value in the leaked gas composition information reaches or exceeds any level of alarm threshold.

24. A method for monitoring gas leaks using the multi-channel broadband gas leak monitoring system according to any one of claims 1-23, characterized in that, Includes the following steps: S1. The visible light image of the monitoring area is acquired through the data acquisition module, and the spectral images of the characteristic absorption bands of various leaked gases are obtained by relying on multiple independent signal channels. The visible light image and the spectral image are then fused and superimposed to generate a fused video image of the monitoring area. The data acquisition module includes at least one infrared signal acquisition module and a laser module. The laser module constitutes an independent signal channel and together with the infrared signal acquisition module, forms multiple independent signal channels. The laser module is configured to emit a modulated laser beam into the monitoring area to identify leaked gases that the infrared signal acquisition module cannot identify or finds difficult to identify. S2. The data processing module receives the fused video image and compares and analyzes it with a preset gas spectral image database to obtain information on the composition of the leaking gas and spatial information related to the gas leak. After the laser module determines that there is a target leaking gas, the software retrieves a pre-made gas cloud image of the target leaking gas from the background and overlays and fuses the gas cloud image with the positioning marker for visualization. This can assist in the comparison and analysis between the fused video image and the gas spectral image database to distinguish mixed gases with overlapping spectra in the monitoring area. S3. The alarm module performs a visualization and display operation and triggers an alarm in response to at least one of the leaked gas composition information and the spatial information.

25. A multi-channel broadband gas leak monitoring device, characterized in that, include: A processor configured to execute computer-executable instructions; A memory that stores one or more computer-executable instructions that, when executed by the processor, implement the steps of the method of claim 24.

26. A computer-readable storage medium, characterized in that, It contains a computer program that can be executed by a processor to implement the steps of the method of claim 24.

Images