Method and apparatus for monitoring greenhouse effect pollutants generated by fuel combustion

Abstract

A method and apparatus for monitoring greenhouse effect pollutants generated by fuel combustion, which method and apparatus belong to the technical field of flue gas monitoring. The monitoring method comprises: by using a diffusion position at a previous moment as a tracking point, determining a diffusion position at a current moment on the basis of an image recognition and tracking method, wherein the diffusion position is the position where a flue gas generated by fuel combustion diffuses in the ambient atmosphere (101); and performing flue gas collection at the diffusion position at the current moment, and performing monitoring, so as to obtain a monitoring result at the current moment (102). In the monitoring method, a flue gas diffusion position is tracked in a real atmospheric environment on the basis of the image recognition and tracking method, and monitoring is performed at the diffusion position, such that monitoring of fuel combustion emission in a real atmospheric environment is realized.

Description

A method and device for monitoring greenhouse effect pollutants from fuel combustion

[0001] This application claims priority to a Chinese patent application filed with the Patent Office of China on December 14, 2023, with application number 202311713169.7 and invention name “A method and device for monitoring greenhouse pollutants from fuel combustion”, the entire contents of which are incorporated herein by reference. Technical Field

[0002] The present invention relates to the technical field of flue gas monitoring, and in particular to a method and device for monitoring greenhouse effect pollutants from fuel combustion. Background Art

[0003] Addressing climate change is a major challenge facing human development. Emissions of greenhouse gas pollutants such as black carbon, CO2, and CH4 have a significant impact on global climate change. Anthropogenic greenhouse gas pollutants are primarily produced by the combustion of fossil fuels and biomass fuels, with emissions from residential fuel combustion in rural areas being a significant source. While some studies have observed atmospheric pollutant emissions from rural residential fuel combustion, these have primarily employed laboratory smog chamber combustion simulations. Observational studies of rural residential fuel combustion emissions in real atmospheric diffusion environments are limited, primarily due to limitations in observational methods.

[0004] Existing methods for observing black carbon from rural residential fuel combustion have the following problems: First, the reactant types, concentrations, and reaction conditions used in smoke chamber combustion simulations differ from the complex atmospheric environment. This simplification of atmospheric diffusion processes limits the accuracy of the results. Second, smoke chamber combustion simulations typically use a dilution channel sampling method, which involves forced dilution. This dilution factor is higher than what is actually achieved under natural conditions, potentially causing the measured components to enter the gaseous phase and maintain phase equilibrium, which can also lead to some deviations in the results. Summary of the Invention

[0005] Based on this, the purpose of the present invention is to provide a method and device for monitoring greenhouse pollutants from fuel combustion, so as to monitor fuel combustion emissions in a real atmospheric diffusion environment.

[0006] To achieve the above object, the present invention provides a method for monitoring greenhouse pollutants from fuel combustion, the method comprising the following steps:

[0007] The diffusion position at the previous moment is used as the tracking point, and the diffusion position at the current moment is determined based on image recognition and tracking methods; the diffusion position is the position where the smoke generated by the fuel combustion diffuses in the ambient atmosphere;

[0008] Smoke is collected at the diffusion position at the current moment, and the smoke collected at the current moment is monitored to determine the monitoring result at the current moment.

[0009] Optionally, the diffusion position at the previous moment is used as the tracking point, and the diffusion position at the current moment is determined based on an image recognition and tracking method, specifically including:

[0010] Taking the tracking point as the center, rotating the camera to obtain multiple images at different angles;

[0011] Perform smoke recognition on each image respectively and determine the image containing smoke as the target image;

[0012] Performing smoke segmentation on the target image to obtain the center position of the smoke contour;

[0013] Based on the angle of the target image, the parameters of the camera shooting the target image and the center position of the smoke outline in the target image, the actual position corresponding to the center position of the smoke outline is determined as the diffusion position at the current moment.

[0014] Optionally, smoke recognition is performed on each image separately to determine an image containing smoke as a target image, specifically including:

[0015] Inputting each of the images into a smoke recognition network model to determine whether each image contains smoke; the smoke recognition network model is obtained by training a convolutional neural network model;

[0016] Set the image containing the smoke as the target image.

[0017] Optionally, performing smoke segmentation on the target image to obtain the center position of the smoke contour specifically includes:

[0018] The target image is transmitted to a smoke segmentation model to obtain the smoke contour in the target image; the smoke segmentation model is obtained by training a YOLOv5s-seg network model;

[0019] Determine the center position of the smoke profile.

[0020] A fuel combustion greenhouse effect pollutant monitoring device, the device comprising: an unmanned aerial vehicle, and an image acquisition and transmission device, a microcontroller, a flue gas collection device, and a greenhouse effect pollutant monitoring device assembled on the unmanned aerial vehicle;

[0021] The image acquisition and transmission device includes an FPV camera and an image transmission module. The FPV camera is used to obtain images at multiple angles, and the image transmission module is used to transmit images at different angles back to the microcontroller.

[0022] The microcontroller is wirelessly connected to the drone control terminal, and is configured to determine a current diffusion position based on multiple images from different angles using an image recognition and tracking method, and send the current diffusion position to the drone control terminal; the diffusion position is a position where smoke generated by fuel combustion diffuses in the ambient atmosphere;

[0023] The microcontroller is also connected to the control end of the smoke collection device, and the microcontroller is also used to control the smoke collection device to collect smoke at the diffusion position at the current moment;

[0024] The greenhouse effect pollutant monitoring device is connected to the microcontroller, and is used to monitor the flue gas collected at the current moment, obtain monitoring results, and send the monitoring results to the microcontroller.

[0025] Optionally, the greenhouse pollutant monitoring equipment includes a black carbon monitor, a CO2 monitoring device and a CH4 monitoring device;

[0026] The smoke collection device includes: a carbon fiber tube, a black carbon sampling tube, a gas sampling tube and a gas observation chamber;

[0027] The carbon fiber tube is mounted on the UAV, the black carbon sampling tube is arranged inside the carbon fiber tube, and the gas sampling tube is fixed parallel to the outside of the carbon fiber tube;

[0028] The air inlet of the black carbon sampling tube is spaced apart from the drone by a preset distance, and the air outlet of the black carbon sampling tube is connected to the black carbon monitor;

[0029] The air inlet of the gas sampling pipe is spaced apart from the drone by a preset distance, the air outlet of the gas sampling pipe is connected to the gas observation room, and the CO2 monitoring device and the CH4 monitoring device are arranged in the gas observation room;

[0030] When flue gas is collected, the air inlet of the black carbon sampling tube and the air inlet of the gas sampling tube are both located at the diffusion position at the current moment.

[0031] Optionally, a particulate filter is provided at the front end of the gas inlet of the gas observation chamber, and the gas outlet of the gas observation chamber is connected to an air pump.

[0032] Optionally, in determining the diffusion position at the current moment based on an image recognition and tracking method according to images from multiple different angles, the microcontroller is specifically configured to:

[0033] Perform smoke recognition on each image respectively and determine the image containing smoke as the target image;

[0034] Performing smoke segmentation on the target image to obtain the center position of the smoke contour;

[0035] Based on the angle of the target image, the parameters of the camera shooting the target image and the center position of the smoke outline in the target image, the actual position corresponding to the center position of the smoke outline is determined as the diffusion position at the current moment.

[0036] Optionally, smoke recognition is performed on each image separately to determine an image containing smoke as a target image, specifically including:

[0037] Inputting each of the images into a smoke recognition network model to determine whether each image contains smoke; the smoke recognition network model is obtained by training a convolutional neural network model;

[0038] Set the image containing the smoke as the target image.

[0039] Optionally, performing smoke segmentation on the target image to obtain the center position of the smoke contour specifically includes:

[0040] Input the target image into a smoke segmentation model to obtain the smoke contour in the target image; the smoke segmentation model is obtained by training a YOLOv5s-seg network model;

[0041] Determine the center position of the smoke profile.

[0042] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0043] An embodiment of the present invention provides a method and device for monitoring greenhouse gas pollutants from fuel combustion. The method comprises: using the diffusion position at a previous moment as a tracking point, and determining the diffusion position at the current moment based on image recognition and tracking methods; the diffusion position is the location where smoke generated by fuel combustion diffuses in the ambient atmosphere; collecting smoke at the diffusion position at the current moment; and monitoring the smoke collected at the current moment to determine the monitoring result at the current moment. The present invention tracks the diffusion position of smoke in a real atmospheric environment based on image recognition and tracking methods, and monitors the diffusion position, thereby achieving monitoring of fuel combustion emissions in a real atmospheric environment.

[0044] Figures in the specification

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0046] FIG1 is a flow chart of a method for monitoring greenhouse gas pollutants from fuel combustion according to an embodiment of the present invention;

[0047] FIG2 is a schematic diagram of the structure of a fuel combustion greenhouse effect pollutant monitoring device provided by an embodiment of the present invention;

[0048] Explanation of the accompanying symbols: 1. UAV landing gear; 2. Microcontroller; 3. Gas observation chamber; 4. Particulate matter filter; 5. Black carbon monitor; 6. Battery compartment; 7. Image acquisition and transmission equipment; 8. Carbon fiber tube; 9. Particulate matter measurement chamber. DETAILED DESCRIPTION

[0049] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0050] The purpose of the present invention is to provide a method and device for monitoring greenhouse effect pollutants from fuel combustion, so as to realize monitoring of fuel combustion emissions in a real atmospheric diffusion environment.

[0051] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0052] Example 1

[0053] Embodiment 1 of the present invention provides a method for monitoring greenhouse gas pollutants from fuel combustion, as shown in FIG1 . The method comprises the following steps:

[0054] Step 101 , using the diffusion position at the previous moment as a tracking point, the diffusion position at the current moment is determined based on an image recognition and tracking method; the diffusion position is the position where the smoke generated by fuel combustion diffuses in the ambient atmosphere.

[0055] Step 101 uses the diffusion position at the previous moment as a tracking point and determines the diffusion position at the current moment based on image recognition and tracking methods; the diffusion position is the position where the smoke generated by fuel combustion diffuses in the ambient atmosphere, specifically including: rotating and shooting with the tracking point as the center to obtain images at multiple different angles.

[0056] Smoke recognition is performed on each image separately, and the image containing smoke is determined as the target image; smoke segmentation is performed on the target image to obtain the center position of the smoke outline; based on the angle of the target image, the parameters of the camera that captured the target image, and the center position of the smoke outline in the target image, the actual position corresponding to the center position of the smoke outline is determined as the diffusion position at the current moment.

[0057] For example, the present invention uses a smoke recognition network model for smoke recognition, which is trained using a convolutional neural network model. The present invention uses a smoke segmentation model for smoke segmentation, which is trained using a YOLOv5s-seg network model.

[0058] Among them, the specific steps for training the YOLOv5s-seg network model are:

[0059] A1. Construct a smoke image dataset: Take images of typical rural residential fuel combustion smoke in the wild, extract no less than 400 images from them, and manually annotate them to train the smoke segmentation model.

[0060] A2. Use the above dataset to train the model based on the YOLOv5s-seg network.

[0061] A3. Deploy the trained smoke segmentation model to a microcontroller for real-time smoke segmentation.

[0062] Step 102: Collect and monitor smoke at the current diffusion position to obtain the monitoring result at the current moment.

[0063] Example 2

[0064] Embodiment 2 of the present invention provides a fuel combustion greenhouse pollutant monitoring device, as shown in Figure 2, the device includes: an unmanned aerial vehicle, and a microcontroller 2 assembled on the unmanned aerial vehicle, a flue gas collection device, a greenhouse pollutant monitoring device, and an image acquisition and transmission device 7; the image acquisition and transmission device 7 includes an FPV camera and an image transmission module, the FPV camera is used to use the diffusion position at the previous moment as a tracking point, and rotate around the tracking point to capture multiple images at different angles, and the image transmission module is used to transmit the images at different angles back to the microcontroller 2; the microcontroller 2 is wirelessly connected to the unmanned aerial vehicle control terminal, the microcontroller 2 is used to determine the diffusion position at the current moment based on the images at the multiple different angles based on image recognition and tracking methods, and send the diffusion position at the current moment to the unmanned aerial vehicle control terminal; the diffusion position is the position where the flue gas generated by the fuel combustion diffuses in the ambient atmosphere; the greenhouse pollutant monitoring device is connected to the microcontroller 2, the greenhouse pollutant monitoring device is used to monitor the flue gas collected at the current moment, send the monitoring results to the microcontroller 2, and transmit them to the user end in real time via a WIFI base station.

[0065] The greenhouse effect pollutant monitoring equipment includes a black carbon monitor 5, a CO2 monitoring device and a CH4 monitoring device.

[0066] The flue gas collection device includes: a carbon fiber tube 8, a black carbon sampling tube, a gas sampling tube and a gas observation chamber 3; the carbon fiber tube 8 is mounted on a monitoring platform, the black carbon sampling tube is arranged inside the carbon fiber tube 8, and the gas sampling tube is fixed parallel to the outside of the carbon fiber tube 8.

[0067] The air inlet of the black carbon sampling tube is spaced a preset distance from the drone, and the air outlet of the black carbon sampling tube is connected to the black carbon monitor 5. The air inlet of the gas sampling tube is spaced a preset distance from the drone, and the air outlet of the gas sampling tube is connected to the gas observation chamber 3. The CO2 monitoring equipment and the CH4 monitoring equipment are arranged in the gas observation chamber 3. Exemplarily, a particulate matter filter 4 is provided at the front end of the air inlet of the gas observation chamber 3, and the air outlet of the gas observation chamber is connected to an air pump. Before the collected flue gas passes through the particulate matter filter 4 and enters the gas observation chamber 3, it is first monitored for particulate matter in the particulate matter measurement chamber 9.

[0068] In an embodiment of the present invention, a 1-meter-long carbon fiber tube 8 is fixed on a monitoring platform, and the black carbon sampling tube of the black carbon monitor 5 is passed through the interior of the carbon fiber tube 8. The air inlet is flush with the front end of the carbon fiber tube 8, and the air outlet is connected to the air inlet of the black carbon monitor 5. The CO2 and CH4 gas sampling tubes are fixed to the carbon fiber tube 8, and a vacuum pump is connected to collect the flue gas into the gas observation chamber 3. A particulate filter 4 is added in front of the gas observation chamber 3. The gas sampling tube and the black carbon sampling tube are both located at the bottom of the drone, and the air inlet is 1 meter away from the drone, which can effectively avoid the airflow interference caused by the operation of the drone rotor. It can achieve simultaneous monitoring of three greenhouse effect pollutants: black carbon, CO2, and CH4.

[0069] The black carbon sampling tube in the embodiment of the present invention is an antistatic silicone hose, and the carbon fiber tube 8 is a support tube with a length of 1 meter for supporting the black carbon sampling tube and the gas sampling tube. The microcontroller 2 used in the embodiment of the present invention is a Raspberry Pi.

[0070] The greenhouse pollutant monitoring equipment and monitoring data acquisition equipment in the embodiment of the present invention are integrated into a self-made monitoring platform, connected to the flue gas collection device, and fixed to the drone landing gear 1. To avoid impact when the drone lands, the monitoring platform is supported by a bracket and is 10 cm higher than the drone landing gear.

[0071] Exemplarily, the embodiment of the present invention is further provided with a temperature and humidity sensor and a communication module. The microcontroller in the embodiment of the present invention communicates with the user-end PC or mobile phone through a WIFI base station to send the monitoring results to the user in real time. The microcontroller also receives images taken by the FPV camera through the image transmission module and transmits the smoke identification results to the drone control terminal in real time.

[0072] The Raspberry Pi and the vacuum pump in the embodiment of the present invention are powered by a mobile power supply, which is arranged in the battery compartment 6.

[0073] The FPV camera is connected to the image transmission module, capturing smoke images and transmitting them via the image transmission module to a pre-set smoke recognition model in the Raspberry Pi for analysis. This embodiment of the present invention provides a deep learning-based smoke recognition method for identifying the location and contours of diffuse smoke in a field atmosphere. This method can be used to guide the flight path and position of a drone, ensuring that the sampling inlet remains within the smoke for monitoring.

[0074] In the embodiment of the present invention, a target image is first determined, and then the target image is segmented to obtain a smoke contour. Furthermore, in the embodiment of the present invention, a YOLOv5s-seg network is used to segment the smoke. The specific steps are as follows:

[0075] B1. Construct a smoke recognition dataset: Take images of typical rural residential fuel combustion smoke in the wild, extract no less than 400 images from them, and manually annotate them to train the smoke segmentation model.

[0076] B2. Use the above dataset to train the model based on the YOLOv5s-seg network.

[0077] B3. Deploy the trained smoke segmentation model to the Raspberry Pi. Use the camera image as input for the model to perform real-time smoke contour recognition and visualize the recognition results in a graph.

[0078] B4. During the monitoring process, the identification results are transmitted back to the drone control terminal. The user can refer to the smoke identification results to obtain the smoke diffusion position and control the drone to the accurate position to ensure that the air inlet of the sampling tube is always at the center of the smoke for collection and monitoring.

[0079] The drone in the embodiment of the present invention is a multi-rotor drone, and the microcontroller 2, the flue gas collection device, the greenhouse pollutant monitoring equipment and the image acquisition and transmission equipment are integrated into the monitoring platform and fixed on the drone landing gear 1, and can be disassembled, which is convenient and simple.

[0080] After rural residential fuels are ignited, smoke continues to be emitted. As the smoke diffuses, its color becomes lighter as it moves farther from the emission source. At this point, a camera captures the direction of the smoke diffusion and transmits the image back to the Raspberry Pi. Based on the smoke recognition results, the drone is guided to the smoke diffusion location and its position is adjusted in real time based on the recognition results, ensuring that the sampling tube nozzle is always at the center of the smoke for collection and monitoring. By monitoring the smoke at diffusion locations at different distances from the smoke emission outlet, data on black carbon, CO2, and CH4 with varying degrees of mixing with the ambient atmosphere can be obtained.

[0081] In summary, the implementation of the embodiments of the present invention has the following beneficial effects:

[0082] (1) Using drones for tracking observations solves the problem of ground-based fixed-point observations being unable to track flue gas monitoring. Drones can be flexibly controlled to track flue gas for observation, making the observation results more comprehensive. Furthermore, flue gas is collected and observed in a real atmospheric diffusion environment in the wild. Compared with smoke chamber combustion simulations, this is closer to the actual state of pollutants in the atmosphere, and the observed data is more realistic and accurate.

[0083] (2) Provide a smoke segmentation model based on YOLOv5s-seg, use FPV image transmission to identify the smoke diffusion position in real time during the monitoring process, guide the flight trajectory and position of the drone, and achieve more accurate pollutant collection and monitoring results.

[0084] (3) Design a flue gas collection device, including a 1-meter-long carbon fiber tube 8 fixed at the front end of the monitoring platform. Pass the black carbon instrument sampling tube inside the carbon fiber tube 8, and the air inlet is flush with the front end of the carbon fiber tube 8. Fix the CO2 and CH4 gas sampling tubes to the carbon fiber tube 8, connect the vacuum pump to collect the flue gas into the gas observation room 3, and add a particulate filter 4 in front of the gas observation room. The air inlet of the sampling tube is 1 meter away from the drone, which can effectively avoid the airflow interference caused by the operation of the drone rotor. It can achieve the simultaneous monitoring of three greenhouse effect pollutants: black carbon, CO2, and CH4.

[0085] (4) Use a high-power outdoor WIFI base station to communicate remotely with the microcontroller, so that users can view monitoring data in real time on a PC or mobile phone when the drone is operating in the air.

[0086] (5) Use a mobile power supply to power the Raspberry Pi and the air pump to reduce the power output of the drone and extend the flight time of the drone.

[0087] (6) The monitoring system of the present invention is highly operable and can be applied to most rotary-wing drones for monitoring pollutants from rural civilian fuel combustion. It effectively improves the shortcomings of previous observation methods, greatly improves the accuracy and representativeness of observation data, and is stable and reliable in data collection and transmission, which has broad application prospects.

[0088] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0089] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only intended to help understand the method and core concept of the present invention. At the same time, those skilled in the art will find that the specific implementation methods and application scopes may vary based on the concept of the present invention. In summary, the contents of this specification should not be construed as limiting the present invention.

Claims

1. A method for monitoring greenhouse pollutants from fuel combustion, characterized in that: The method comprises the following steps: The diffusion position at the previous moment is used as the tracking point, and the diffusion position at the current moment is determined based on the image recognition and tracking method; the diffusion position is the position where the smoke generated by the fuel combustion diffuses in the ambient atmosphere; Smoke is collected at the diffusion position at the current moment, and the smoke collected at the current moment is monitored to determine the monitoring result at the current moment.

2. The method for monitoring greenhouse pollutants from fuel combustion according to claim 1, characterized in that: The diffusion position at the previous moment is used as the tracking point, and the diffusion position at the current moment is determined based on the image recognition and tracking method, specifically including: Taking the tracking point as the center, rotating and shooting are performed to obtain images at multiple different angles; Perform smoke recognition on each image respectively, and determine the image containing smoke as the target image; Performing smoke segmentation on the target image to obtain the center position of the smoke contour; Based on the angle of the target image, the parameters of the camera shooting the target image and the center position of the smoke outline in the target image, the actual position corresponding to the center position of the smoke outline is determined as the diffusion position at the current moment.

3. The method for monitoring greenhouse pollutants from fuel combustion according to claim 2, characterized in that: Perform smoke recognition on each image respectively and determine the image containing smoke as the target image, specifically including: Input each of the images into a smoke recognition network model to determine whether each of the images contains smoke; the smoke recognition network model is obtained by training a convolutional neural network model; Set the image containing the smoke as the target image.

4. The method for monitoring greenhouse pollutants from fuel combustion according to claim 2, characterized in that: Performing smoke segmentation on the target image to obtain the center position of the smoke contour specifically includes: The target image is transmitted to a smoke segmentation model to obtain a smoke contour in the target image; the smoke segmentation model is obtained by training a YOLOv5s-seg network model; Determines the center position of the smoke profile.

5. A fuel combustion greenhouse effect pollutant monitoring device, characterized in that: The device includes: a drone, and an image acquisition and transmission device, a microcontroller, a smoke collection device and a greenhouse effect pollutant monitoring device assembled on the drone; The image acquisition and transmission device includes an FPV camera and an image transmission module, wherein the FPV camera is used to acquire images at multiple different angles, and the image transmission module is used to transmit images at different angles back to the microcontroller; The microcontroller is wirelessly connected to the drone control terminal, and the microcontroller is used to determine the diffusion position at the current moment based on the image recognition and tracking method according to the images at multiple different angles, and send the diffusion position at the current moment to the drone control terminal; the diffusion position is the position where the smoke generated by the fuel combustion diffuses in the ambient atmosphere; The microcontroller is also connected to the control end of the smoke collection device, and the microcontroller is also used to control the smoke collection device to collect smoke at the diffusion position at the current moment; The greenhouse effect pollutant monitoring device is connected to the microcontroller, and is used to monitor the flue gas collected at the current moment, obtain monitoring results, and send the monitoring results to the microcontroller.

6. The fuel combustion greenhouse effect pollutant monitoring device according to claim 5, characterized in that: The greenhouse effect pollutant monitoring equipment includes a black carbon monitor, a CO2 monitoring equipment and a CH4 monitoring equipment; The smoke collection device comprises: a carbon fiber tube, a black carbon sampling tube, a gas sampling tube and a gas observation chamber; The carbon fiber tube is mounted on the drone, the black carbon sampling tube is arranged inside the carbon fiber tube, and the gas sampling tube is fixed in parallel to the outside of the carbon fiber tube; The air inlet of the black carbon sampling tube is spaced apart from the drone by a preset distance, and the air outlet of the black carbon sampling tube is connected to the black carbon monitor; The air inlet of the gas sampling pipe is spaced apart from the drone by a preset distance, the air outlet of the gas sampling pipe is connected to the gas observation room, and the CO2 monitoring device and the CH4 monitoring device are arranged in the gas observation room; When smoke is collected, the air inlet of the black carbon sampling tube and the air inlet of the gas collection tube are both located at the diffusion position at the current moment.

7. The fuel combustion greenhouse effect pollutant monitoring device according to claim 6, characterized in that: A particle filter is arranged at the front end of the gas inlet of the gas observation chamber, and the gas outlet of the gas observation chamber is connected to the air pump.

8. The fuel combustion greenhouse effect pollutant monitoring device according to claim 5, characterized in that: In terms of determining the diffusion position at the current moment based on the image recognition and tracking method according to the images at multiple different angles, the microcontroller is specifically used to: Perform smoke recognition on each image respectively, and determine the image containing smoke as the target image; Performing smoke segmentation on the target image to obtain the center position of the smoke contour; Based on the angle of the target image, the parameters of the camera shooting the target image and the center position of the smoke outline in the target image, the actual position corresponding to the center position of the smoke outline is determined as the diffusion position at the current moment.

9. The fuel combustion greenhouse effect pollutant monitoring device according to claim 8, characterized in that: Perform smoke recognition on each image respectively and determine the image containing smoke as the target image, specifically including: Input each of the images into a smoke recognition network model to determine whether each of the images contains smoke; the smoke recognition network model is obtained by training a convolutional neural network model; Set the image containing the smoke as the target image.

10. The fuel combustion greenhouse pollutant monitoring device according to claim 8, characterized in that: Performing smoke segmentation on the target image to obtain the center position of the smoke contour specifically includes: The target image is input into a smoke segmentation model to obtain a smoke contour in the target image; the smoke segmentation model is obtained by training a YOLOv5s-seg network model; Determines the center position of the smoke profile.

Images