Gas concentration field reconstruction method based on TDLAS (Tunable Diode Laser Absorption Spectroscopy) double-node adaptive scanning

Abstract

The invention relates to a gas concentration field reconstruction method based on TDLAS (tunable diode laser absorption spectroscopy) double-node self-adaptive scanning. The method comprises the following steps: arranging an annular slide rail on the section of a flue, arranging two TDLAS sensors on the annular slide rail on the inner wall of the flue, independently measuring by the two TDLAS sensors, and defining a world coordinate system and a sensor coordinate system; concentration field reconstruction grid division is carried out, and for a laser light path emitted by a TDLAS sensor, a solving method of a weight coefficient of the laser light path in an effective grid is determined; global coarse scanning is carried out, column concentration measurement is completed, and an initial concentration field of each sensor is reconstructed; identifying a high-concentration area, and performing fine scanning on the high-concentration area; and S5, fusing the initial concentration field obtained by coarse scanning and the region Rp concentration field obtained by fine scanning to form final concentration field distribution.

Description

Technical Field

[0001] This invention belongs to the field of gas concentration monitoring, specifically relating to a gas concentration field reconstruction method based on TDLAS dual-node adaptive scanning. Background Technology

[0002] Currently, chimney gas concentration monitoring mainly employs single-point sampling or single-path laser measurement methods. Single-point sampling methods cannot reflect the concentration distribution across a cross-section, resulting in insufficient spatial representativeness. Single-path tunable diode laser absorption spectroscopy (TDLAS) technology can only obtain path integral concentration and cannot achieve concentration field visualization.

[0003] Existing optical tomography techniques typically require the deployment of numerous fixed sensors or intensive rotating scans to achieve reconstruction. The former results in complex and costly systems, while the latter leads to long data acquisition times, making it difficult to meet the real-time monitoring needs of industrial sites. Furthermore, existing methods have shortcomings in terms of measurement path optimization and the adaptability of reconstruction algorithms.

[0004] Therefore, there is an urgent need for a technical solution that can achieve efficient and accurate concentration field reconstruction with limited hardware costs. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a gas concentration field reconstruction method based on TDLAS dual-node adaptive scanning. The technical solution is as follows: A gas concentration field reconstruction method based on TDLAS dual-node adaptive scanning includes the following steps: S1, set an annular slide rail on the cross section of the flue, place two TDLAS sensors on the annular slide rail on the inner wall of the flue, measure independently with the two TDLAS sensors, define the world coordinate system and the sensor coordinate system, and give the parametric equation of the laser emitted by the TDLAS sensor in the world coordinate system. S2, perform concentration field reconstruction mesh generation, and determine the solution method for the weight coefficient of the laser light path emitted by the TDLAS sensor in the effective mesh; S3, perform a global coarse scan to complete column concentration measurement and reconstruct the initial concentration field of each sensor; S4. Identify high-concentration areas and perform detailed scanning on these areas, as follows: S41, Set the concentration threshold C t This is used to distinguish high-density meshes from the background; S42, iterate through all the meshes, if the mesh density... Then the grid is marked as a high-concentration grid, forming a high-concentration grid set. ; S43, for sets Connectivity component analysis was performed on the grid to group spatially adjacent high-concentration grids into the same high-concentration region, ultimately identifying the high-concentration region R. p ; S44, Identify high-concentration areas R p Principal axis direction: computational region R p The centroid, obtaining region R p Find the center coordinates of all grid cells within the region; calculate its geometric centroid and covariance matrix; calculate the eigenvalues ​​and eigenvectors of the covariance matrix; the eigenvector corresponding to the largest eigenvalue is the principal axis direction and azimuth angle of the principal axis direction for that region. α 1; S46, Based on the spindle direction, determine the initial positions of the two TDLAS sensors for fine scanning on the annular slide rail, wherein the initial measurement position of one sensor is the azimuth angle of the spindle direction; determine the other measurement positions and the fine scanning step size; S47, determine the scanning angle range for each sensor's fine scan measurement position; S48, place the two TDLAS sensors at the initial measurement position of the annular slide rail respectively, and both sensors perform a synchronous scan of the column concentration within their respective scanning range with a fine scan step size; after the scan is completed, rotate the two TDLAS sensors synchronously to the next measurement position, and both sensors perform a synchronous scan of the column concentration within their respective scanning range again with a fine scan step size, until all measurement positions are scanned synchronously. S49, based on the concentration field reconstruction method in S3, for the high concentration region R p Concentration field reconstruction was performed to obtain region R. p The detailed scan results for each grid concentration yielded the region R. p Concentration field ; S5, the initial concentration field obtained from the coarse scan and the region R obtained by fine scanning p Concentration field The mixture is then fused to form the final concentration field distribution. The method is as follows: For each grid j, calculate the total weight of its coverage by the fine-scanning laser path; set a coverage threshold and define the fusion weight of each grid; for each grid j, calculate the final concentration value, and then calculate the final concentration field distribution. Concentration field visualization processing is performed to obtain a concentration field distribution map.

[0006] Furthermore, in S1, the world coordinate system and the sensor coordinate system are defined as follows: the origin O of the world coordinate system is the center of the flue. w With the TDLAS sensor as the origin O sEstablish a sensor coordinate system pointing to the origin O of the world coordinate system. w Two sensor coordinate systems are established with the positive Y-axis as the direction.

[0007] Furthermore, the method for S2 is as follows: Construct the circumscribed square of the circular slide rail, centered at the origin of the world coordinate system; determine the boundary coordinates of the circumscribed square, divide the square region into initial grids in the X and Y directions respectively, and determine the boundary coordinate formulas of the grid in row a and column b; For each initial grid cell, determine its center coordinates; Calculate the Euclidean distance from the center of each initial grid to the origin of the world coordinate system. Set a distance threshold. If the Euclidean distance is less than the distance threshold, the initial grid is retained and called a valid grid; otherwise, it is discarded. Assign a continuous one-dimensional index j, j = 1,2,…,M, to all retained grids, and establish a mapping relationship between the retained grids and the two-dimensional grid coordinates (a,b); Let the laser beam path emitted by the TDLAS sensor be i, and its current azimuth angle be α. i The square angle of the laser beam path i emitted by the TDLAS sensor in the sensor coordinate system is β. i The parametric equations for the i-th laser path are as follows: (5) Solve for the intersection parameter t of the i-th laser path and the j-th grid boundary. ij1 and t ij2 Then the weight coefficient of the i-th laser path in the j-th grid is: (6) Furthermore, the S3 method is as follows: S31, Perform a global coarse scan to complete column concentration measurement: Determine the initial positions of the two TDLAS sensors on the annular slide rail, the measurement positions, the coarse scan step size, and the scanning angle range in one measurement; At each measurement position, both sensors perform a synchronous scan of the column concentration within the scanning angle range with the same coarse scan step size; After each scan, the two sensors rotate synchronously along the annular slide rail and move to the next measurement position, sequentially completing the scan of each measurement position to complete the coarse scan of column concentration measurement; S32, Reconstruct the initial concentration field: Initialize the concentration of each grid, and continuously update the concentration of each grid using an algebraic iteration method until the convergence condition is met, thereby obtaining the coarse scan result of the concentration of each grid and obtaining the initial concentration field.

[0008] Furthermore, in S31, suppose that one of the sensors obtains N laser light paths after a coarse scan, and the column concentration corresponding to the i-th laser light path is IC. i The column concentration of the region to be measured through the i-th laser beam path is expressed as: (7) In the formula, C j This represents the average concentration in the j-th grid. l ij Let be the weight coefficient of the i-th laser path in the j-th grid, where j = 1, 2, ..., M.

[0009] Furthermore, the S32 method is as follows: S321, initialize the concentration of each grid cell, C 1 j = 0 ppm; S322 uses an algebraic iteration method to continuously update the concentration of each grid cell until the convergence condition is met. The iteration formula is as follows: (8) In the formula, k is the iteration number; i is the laser path number; j is the grid number; It is a relaxation factor; S323, Determine the stopping condition for iteration: The stopping condition is that the number of iterations reaches a specified value, or the root mean square error... ,in, The value was obtained experimentally. S324, after the iteration is completed, the coarse scan result of the concentration of each grid is obtained, and the initial concentration field is obtained.

[0010] Furthermore, a curve is plotted with the number of iterations on the x-axis and the RMSE value on the y-axis. The inflection point of the curve is then used to determine... The value of .

[0011] Furthermore, S41, C is calculated based on the statistical properties of the initial concentration field. t : (9) In the formula, The average concentration of the initial concentration field; denoted as the standard deviation of the initial concentration field.

[0012] Furthermore, the method for step S47 is as follows: [The text abruptly shifts to a different topic] ...the target region R... p The world coordinates of all grid vertices are transformed to sensor coordinates; in the sensor coordinates, the azimuth angle of each grid vertex is calculated, and the minimum angle at each sensor measurement position is found. and maximum angle The angle is extended to ensure complete coverage of area R. p Finally, the scanning angle range of each sensor's fine scanning measurement position was determined. ) = ( ),in, This is for the extended angle value.

[0013] Furthermore, the S5 method is as follows: S51, For each grid j, calculate the total weight of its coverage by the fine-scan laser path: (12) In the formula, This represents the total number of laser optical paths used for fine scanning. This represents the weighting coefficient of the i-th laser path in the j-th grid during fine scanning. S52, Set coverage threshold Define the fusion weights for each grid. : (13) in, This indicates that the grid fully trusts the results of the detailed scan; This indicates complete trust in the coarse scan results; S53, for each grid j, the final concentration value is obtained by the following formula: (14) S54. Based on the final grid concentration obtained, perform concentration field visualization processing to obtain a concentration field distribution map.

[0014] Compared with the prior art, the beneficial effects of the technical solution of the present invention are: 1) High scanning efficiency: coarse scanning enables rapid positioning, while fine scanning concentrates resources, avoiding the time-consuming problem of full-circumference high-density scanning.

[0015] 2) Adjustable resolution: By linearly moving the slide rail, spatial resolution exceeding that of a fixed sensor array can be easily achieved in the region of interest.

[0016] 3) Flexible path configuration: The slide rail provides the ability to measure at any angle, and the two TDLAS sensors can be controlled independently to form an asymmetric measurement path.

[0017] 4) Redundancy and fault tolerance: If one TDLAS sensor fails, another device can still independently complete the degraded mode scanning measurement. Attached Figure Description

[0018] Figure 1 is a schematic diagram of image coordinate matching; Figure 2 Divide the area to be measured into a grid; Figure 3 This is a schematic diagram of the coarse scanning laser path distribution; Figure 4 A two-dimensional schematic diagram of the algebraic iteration method; Figure 5 This is a schematic diagram of the laser optical path for fine scanning. Detailed Implementation

[0019] The present invention will be further described below with reference to specific embodiments.

[0020] The present invention provides a gas concentration field reconstruction method based on TDLAS dual-node adaptive scanning, comprising the following steps: Step 1, as follows Figure 1 As shown, define and initialize the coordinate system; In this embodiment, the inner diameter of the circular chimney is 1 m, and the center of the chimney is taken as the origin O of the world coordinate system. w The positive X-axis direction is horizontal to the right, and the positive Y-axis direction is vertically upward. Set the angle of the positive X-axis direction to 0°, and increase the angle clockwise.

[0021] Two TDLAS sensors are placed on an annular slide rail on the inner wall of the chimney, with a radius r = 0.5 m. The origin O is taken as the sensor. s Establish a sensor coordinate system pointing to the origin O of the world coordinate system. w The direction is the positive Y-axis. Rotate 90° counterclockwise to get the positive X-axis. Set the angle of the positive X-axis to 0° and increase the angle clockwise.

[0022] Let the azimuth angle of a certain TDLAS sensor be α, then its coordinates in the world coordinate system are (0.5cosα, 0.5sinα). Let the azimuth angle of the laser emitted by the TDLAS sensor in the sensor coordinate system be β, then its parametric equation in the world coordinate system is: (1) Step two: Based on the coordinate system established in step one, perform mesh generation for concentration field reconstruction and calculate weight coefficients.

[0023] In this embodiment, the concentration field reconstruction mesh is generated using the following method: Figure 2 As shown, construct the circumscribed square of the circular slide rail, with the origin O of the world coordinate system as the center. w Centered on a point with sides of length 1 m, the boundary coordinates of this square are X... min = -0.5 m, X max = 0.5 m, Y min = -0.5 m, Ymax = 0.5 m. Divide the square region into m = 50 equal parts in both the X and Y directions, forming an initial grid of 50×50, with a grid side length of... = 1 / 50 = 0.02 m. The formula for the boundary coordinates of the grid in row a and column b is: (2) For each initial grid cell, calculate its center coordinates: (3) Calculate the center of each initial grid to World coordinate system Euclidean distance from the origin: (4) like If the condition is met, the grid is retained; otherwise, it is discarded. All retained grids are assigned a continuous one-dimensional index j (j = 1,2,…,M), and a mapping relationship is established between them and the two-dimensional grid coordinates (a,b).

[0024] According to formula (1), the parametric equation of the i-th laser path is obtained as follows: (5) The intersection parameter t of the i-th laser path and the j-th grid boundary is determined according to formulas (2) and (5). ij1 and t ij2 Then the weight coefficient of the i-th laser path in the j-th grid is: (6) Step 3: After completing the mesh generation for concentration field reconstruction and calculating the weight coefficients, perform a global coarse scan.

[0025] In this embodiment, the two TDLAS sensors are initially positioned at 0° and 45° on the annular slide rail. Figure 3 As shown, the measurement process includes four measurement positions. At each measurement position, both sensors measure their respective positions in 5° increments. sensor The column concentration within the coordinate system, ranging from 60° to 120°, is scanned synchronously. After each scan, the two sensors synchronously rotate 90° clockwise along the circular slide rail and move to the next measurement position, until they sequentially cover four sets of positions: (0°, 45°), (90°, 135°), (180°, 225°), and (270°, 315°). A total of N laser paths are obtained, with the column concentration corresponding to the i-th laser path being IC. i After completing the column concentration measurement, the initial concentration field is reconstructed.

[0026] Step two completed the creation of the concentration field reconstruction grid and the weighting coefficients for each grid. Calculations, such as Figure 4 Therefore, the column concentration of the region to be measured through the i-th laser beam path can be expressed as: (7) In the formula, C j This represents the average concentration of the j-th grid.

[0027] Step 1: Initialize the concentration for each grid cell, C 1 j = 0 ppm.

[0028] Step 2 involves continuously updating the concentration of each grid cell using an algebraic iteration method until the convergence condition is met. The iteration formula is as follows: (8) In the formula, k is the iteration number; i is the laser path number; j is the grid number; The relaxation factor is 0.3 in this embodiment.

[0029] Step 3: Determine the iteration stopping condition. The stopping condition is that the number of iterations reaches a specified value, or the root mean square error reaches a certain threshold. ,in, The value was obtained experimentally: a curve was plotted with the number of iterations on the x-axis and the RMSE value on the y-axis. Initially, the RMSE value decreased sharply, then plateaued. The inflection point of the curve was used to determine the RMSE value. The value of .

[0030] Step 4: After the iteration is complete, the coarse scan results of the concentration for each grid cell are obtained. Obtain the initial concentration field.

[0031] Step four: After reconstructing the initial concentration field through a global coarse scan, identify the high-concentration region R. p And perform detailed scanning of high-concentration areas, such as Figure 5 As shown, assume the blue area in the diagram is R. p .

[0032] Step 1: Set a concentration threshold C t It is used to distinguish high-concentration grids from the background, and C is calculated based on the statistical properties of the initial concentration field. t : (9) In the formula, The average concentration of the initial concentration field; denoted as the standard deviation of the initial concentration field.

[0033] Step 2: Traverse all the grids, if the grid density... Then the grid is marked as a high-concentration grid, forming a high-concentration grid set. .

[0034] Step 3, for the set Connectivity component analysis was performed on the grid to group spatially adjacent high-concentration grids into the same high-concentration region, ultimately identifying the high-concentration region R. p .

[0035] Step 4: Identify the high-concentration region R p Principal axis direction: computational region R p The centroid, obtaining region R p Center coordinates of all grids Calculate its geometric centroid as ( ) = ( After obtaining the geometric centroid, calculate the covariance matrix: (10) in, , , .

[0036] Calculate the eigenvalues ​​of the covariance matrix and eigenvectors Solve the following equation: (11) in, For eigenvalues, These are the eigenvectors.

[0037] Among them, the eigenvector corresponding to the largest eigenvalue ( The direction of the principal axis of the region is , and its azimuth angle is . .

[0038] Step 6: Determine the position angles of the four sensors for fine scanning based on the spindle direction. , , , .

[0039] Step 7: Calculate the angular range scanned by the sensor at each location, and define the target area R. p The world coordinates of all grid vertices are transformed to sensor coordinates. In the sensor coordinates, the azimuth angle of each grid vertex is calculated, and the minimum angle at each sensor location is found. and maximum angle And the angle is extended to ensure complete coverage of area R. p The final scanning angle range for each sensor position is determined as follows: ( ) = ( ).in, .

[0040] Step 8: Place the two TDLAS sensors on the circular slide rail respectively. and At the designated position, both sensors perform a synchronous scan of the column concentration within their respective scan ranges, with a step size of 1°. After the scan is completed, the two TDLAS sensors are synchronously rotated 180° clockwise to... and Position: Both sensors simultaneously scan the column concentration within their respective scanning ranges again, with a step size of 1°.

[0041] Step 9: Based on the concentration field reconstruction method in Step 3, perform high-concentration region R... p Concentration field reconstruction was performed to obtain region R. p The detailed scan results for each grid concentration yielded the region R. p Concentration field .

[0042] Step 5: Apply the initial concentration field obtained from the coarse scan. and the region R obtained by fine scanning p Concentration field The mixture is then fused to form the final concentration field distribution. .

[0043] Step 1: For each grid j, calculate the total weight of its coverage by the fine-scan laser path: (12) In the formula, This represents the total number of laser optical paths used for fine scanning. This represents the weighting coefficient of the i-th laser path in the j-th grid during fine scanning.

[0044] Step 2: Set the coverage threshold Define the fusion weights for each grid. : (13) in, This indicates that the grid fully trusts the results of the detailed scan; This indicates complete trust in the coarse scan results. In this embodiment, .

[0045] Step 3: For each grid j, the final concentration value is obtained by the following formula: (14) Step 4: Based on the final grid concentration obtained, perform concentration field visualization processing to obtain a concentration field distribution map.

Claims

1. A gas concentration field reconstruction method based on TDLAS dual-node adaptive scanning, comprising the following steps: S1, set an annular slide rail on the cross section of the flue, place two TDLAS sensors on the annular slide rail on the inner wall of the flue, measure independently with the two TDLAS sensors, define the world coordinate system and the sensor coordinate system, and give the parametric equation of the laser emitted by the TDLAS sensor in the world coordinate system. S2, perform concentration field reconstruction mesh generation, and determine the solution method for the weight coefficient of the laser light path emitted by the TDLAS sensor in the effective mesh; S3, perform a global coarse scan to complete column concentration measurement and reconstruct the initial concentration field of each sensor; S4. Identify high-concentration areas and perform detailed scanning on these areas, as follows: S41, Set the concentration threshold C t This is used to distinguish high-density meshes from the background. S42, traverse all meshes, if the mesh density... Then the grid is marked as a high-concentration grid, forming a high-concentration grid set. ; S43, for sets Connectivity component analysis was performed on the grid to group spatially adjacent high-concentration grids into the same high-concentration region, ultimately identifying the high-concentration region R. p ; S44, Identify high-concentration areas R p Principal axis direction: computational region R p The centroid, obtaining region R p The center coordinates of all grid cells within the grid; Calculate its geometric centroid and covariance matrix, then calculate the eigenvalues ​​and eigenvectors of the covariance matrix. The eigenvector corresponding to the largest eigenvalue is the principal axis direction and azimuth angle of the principal axis direction for that region. α 1; S46, Based on the spindle direction, determine the initial positions of the two TDLAS sensors for fine scanning on the annular slide rail, wherein the initial measurement position of one sensor is the azimuth angle of the spindle direction; determine the other measurement positions and the fine scanning step size; S47, determine the scanning angle range for each sensor's fine scan measurement position; S48, place the two TDLAS sensors at the initial measurement position of the annular slide rail respectively, and both sensors perform a synchronous scan of the column concentration within their respective scanning range with a fine scan step size; after the scan is completed, rotate the two TDLAS sensors synchronously to the next measurement position, and both sensors perform a synchronous scan of the column concentration within their respective scanning range again with a fine scan step size, until all measurement positions are scanned synchronously. S49, based on the concentration field reconstruction method in S3, for the high concentration region R p Concentration field reconstruction was performed to obtain region R. p The detailed scan results for each grid concentration yielded the region R. p Concentration field ; S5, the initial concentration field obtained from the coarse scan and the region R obtained by fine scanning p Concentration field The mixture is then fused to form the final concentration field distribution. The method is as follows: For each grid j, calculate the total weight of its coverage by the fine-scanning laser path; set a coverage threshold and define the fusion weight of each grid; for each grid j, calculate the final concentration value, and then calculate the final concentration field distribution. Concentration field visualization processing is performed to obtain a concentration field distribution map.

2. The gas concentration field reconstruction method based on TDLAS dual-node adaptive scanning according to claim 1, characterized in that, In S1, the world coordinate system and the sensor coordinate system are defined as follows: the origin O of the world coordinate system is the center of the flue. w With the TDLAS sensor as the origin O s Establish a sensor coordinate system pointing to the origin O of the world coordinate system. w Two sensor coordinate systems are established with the positive Y-axis as the direction.

3. The gas concentration field reconstruction method based on TDLAS dual-node adaptive scanning according to claim 1, characterized in that, The method for S2 is as follows: Construct the circumscribed square of the circular slide rail, centered at the origin of the world coordinate system; determine the boundary coordinates of the circumscribed square, divide the square region into initial grids in the X and Y directions respectively, and determine the boundary coordinate formulas of the grid in row a and column b; For each initial grid cell, determine its center coordinates; Calculate the Euclidean distance from the center of each initial grid to the origin of the world coordinate system. Set a distance threshold. If the Euclidean distance is less than the distance threshold, retain the initial grid and call it a valid grid; otherwise, discard it. Assign a continuous one-dimensional index j, j = 1,2,…,M, to all retained grids, and establish a mapping relationship between the retained grids and the two-dimensional grid coordinates (a,b); Let the laser beam path emitted by the TDLAS sensor be i, and its current azimuth angle be α. i The square angle of the laser beam path i emitted by the TDLAS sensor in the sensor coordinate system is β. i The parametric equations for the i-th laser path are as follows: (5) Solve for the intersection parameter t of the i-th laser path and the j-th grid boundary. ij1 and t ij2 Then the weight coefficient of the i-th laser path in the j-th grid is: (6)。 4. The gas concentration field reconstruction method based on TDLAS dual-node adaptive scanning according to claim 1, characterized in that, The method for S3 is as follows: S31, Perform a global coarse scan to complete column concentration measurement: Determine the initial position of the two TDLAS sensors on the annular slide rail, the measurement position, the coarse scan step size, and the scanning angle range in one measurement; At each measurement position, both sensors perform a synchronous scan of the column concentration within the scanning angle range with the same coarse scan step size; After each scan, the two sensors rotate synchronously along the annular slide rail and move to the next measurement position, sequentially completing the scan of each measurement position to complete the coarse scan of column concentration measurement; S32, Reconstruct the initial concentration field: Initialize the concentration of each grid, and continuously update the concentration of each grid using an algebraic iteration method until the convergence condition is met, thereby obtaining the coarse scan result of the concentration of each grid and obtaining the initial concentration field.

5. The gas concentration field reconstruction method based on TDLAS dual-node adaptive scanning according to claim 4, characterized in that, In S31, suppose one of the sensors obtains N laser paths after a coarse scan, and the column concentration corresponding to the i-th laser path is IC. i The column concentration of the region to be measured through the i-th laser beam path is expressed as: （7） In the formula, C j This represents the average concentration in the j-th grid. l ij Let be the weight coefficient of the i-th laser path in the j-th grid, where j = 1, 2, ..., M.

6. The gas concentration field reconstruction method based on TDLAS dual-node adaptive scanning according to claim 5, characterized in that, The S32 method is as follows: S321, initialize the concentration of each grid cell, C 1 j = 0 ppm; S322 uses an algebraic iteration method to continuously update the concentration of each grid cell until the convergence condition is met. The iteration formula is as follows: （8） In the formula, k is the iteration number; i is the laser path number; j is the grid number; It is a relaxation factor; S323, Determine the stopping condition for iteration: The stopping condition is that the number of iterations reaches a specified value, or the root mean square error... ,in, The value was obtained experimentally. S324, after iteration, obtain the coarse scan results of the concentration of each grid and obtain the initial concentration field.

7. The gas concentration field reconstruction method based on TDLAS dual-node adaptive scanning according to claim 6, characterized in that, Plot a curve with the number of iterations on the x-axis and the RMSE value on the y-axis, and determine the inflection point of the curve. The value of .

8. The gas concentration field reconstruction method based on TDLAS dual-node adaptive scanning according to claim 1, characterized in that, S41, Calculate C based on the statistical properties of the initial concentration field. t : （9） In the formula, The average concentration of the initial concentration field; denoted as the standard deviation of the initial concentration field.

9. The gas concentration field reconstruction method based on TDLAS dual-node adaptive scanning according to claim 1, characterized in that, The method for step S47 is as follows: [The text abruptly ends here, likely due to an incomplete sentence or a formatting error.] p The world coordinates of all grid vertices are transformed to sensor coordinates; in the sensor coordinates, the azimuth angle of each grid vertex is calculated, and the minimum angle at each sensor measurement position is found. and maximum angle The angle is extended to ensure complete coverage of area R. p Finally, the scanning angle range of each sensor's fine scanning measurement position was determined. ) = ( ),in, This is for the extended angle value.

10. The gas concentration field reconstruction method based on TDLAS dual-node adaptive scanning according to claim 1, characterized in that, The specific method for S5 is as follows: S51, For each grid j, calculate the total weight of its coverage by the fine-scan laser path: （12） In the formula, This represents the total number of laser optical paths used for fine scanning. This represents the weighting coefficient of the i-th laser path in the j-th grid during fine scanning. S52, Set coverage threshold Define the fusion weights for each grid. : （13） in, This indicates that the grid fully trusts the results of the detailed scan; This indicates complete trust in the coarse scan results; S53, for each grid j, the final concentration value is obtained by the following formula: （14） S54. Based on the final grid concentration obtained, perform concentration field visualization processing to obtain a concentration field distribution map.

Images