High-resolution seawater distributed temperature monitoring method

Abstract

The invention relates to a high-resolution seawater distributed temperature monitoring method. The method comprises the following steps: S1, forming a two-dimensional temperature matrix by a plurality of groups of distributed Raman signals acquired in a plurality of continuous time frames; s2, noise reduction is carried out on the two-dimensional temperature matrix through a non-local mean algorithm, the overall temperature trend of a seawater area to be subjected to temperature measurement in the two-dimensional temperature matrix is reserved, and the two-dimensional temperature matrix after primary noise reduction is obtained; s3, noise reduction is carried out on the two-dimensional temperature matrix after primary noise reduction through a total variation denoising algorithm, details of a temperature change area in the two-dimensional temperature matrix after primary noise reduction are reserved, and a two-dimensional temperature matrix after secondary noise reduction is obtained; and S4, adding the collected distributed Raman signals as a new line to the last line of the two-dimensional temperature matrix after secondary noise reduction to form a next time frame two-dimensional temperature matrix, reducing noise through the non-local mean algorithm, and taking the last line as the current temperature of the seawater area to be subjected to temperature measurement.

Description

Technical Field

[0001] The invention relates to the field of seawater temperature monitoring, and in particular to a high-resolution distributed seawater temperature monitoring method. Background Art

[0002] Sea water temperature is an important environmental factor for the occurrence of red tides. 20-30℃ is the suitable temperature range for the occurrence of red tides. A sudden increase in water temperature of more than 2℃ within a week is a precursor to the occurrence of red tides.

[0003] The distributed optical fiber temperature measurement system is a sensing system developed in recent years for real-time measurement of spatial temperature field distribution. The system uses light waves as information carriers. During the propagation of light wave signals in optical fibers, their intensity, polarization state, phase and other characteristic quantities will change due to the influence of external environmental factors such as temperature, electromagnetics, and vibration. By monitoring these changes, it is possible to monitor the environment and measure the temperature field of the optical fiber in real time. The optical time domain reflectometry (OTDR) technology can be used to accurately locate the measurement point. In the prior art, "Novel transient-state analysis approach for distributed temperature sensing based on spontaneous Raman scattering[J]" (Opto-Electronics Review, 2024) proposed a temperature sensing technology based on Raman anti-Stokes scattering. This technology uses the OTDR algorithm to perform differential analysis on the integrated Raman back-anti-Stokes scattering signal, reconstruct the temperature distribution along the measured optical fiber, and achieve a temperature resolution of 2°C and a spatial resolution of 1 meter. In actual seawater temperature measurements used for red tide warnings, seawater temperature changes are affected by ocean currents, climate, day and night temperature changes, etc. At the same time, the characteristics of seawater temperature changes such as gentle changes and noise in the fiber optic temperature measurement system will also affect the detection results. If the noise reduction method provided by existing technologies is used, it is difficult to capture the changing trends and details of seawater, making it difficult for existing distributed seawater temperature measurement methods to achieve high resolution.

[0004] The purpose of the present invention is to design a high-resolution distributed seawater temperature monitoring method in view of the above-mentioned problems in the prior art. Summary of the invention

[0005] In view of the problems existing in the above-mentioned prior art, the present invention provides a high-resolution distributed seawater temperature monitoring method, which can effectively solve at least one problem existing in the above-mentioned prior art.

[0006] The technical solution of the present invention is: A high-resolution distributed seawater temperature monitoring method comprises the following steps: S1, collecting distributed Raman signals from the seawater area to be measured in a plurality of consecutive time frames, and forming a two-dimensional temperature matrix with a plurality of groups of distributed Raman signals obtained in a plurality of consecutive time frames, wherein the rows of the two-dimensional temperature matrix represent the changes of the Raman signals at different distance positions, and the columns of the two-dimensional temperature matrix represent the changes of the Raman signals at different time frames; S2, denoising the two-dimensional temperature matrix by a non-local means algorithm, retaining the overall temperature trend of the seawater area to be measured in the two-dimensional temperature matrix, and obtaining a two-dimensional temperature matrix after one denoising; S3, denoising the two-dimensional temperature matrix after the primary denoising by using a total variation denoising algorithm, retaining the temperature change area details in the two-dimensional temperature matrix after the primary denoising, and obtaining the two-dimensional temperature matrix after the secondary denoising; S4, in the next time frame, distributed Raman signal acquisition is performed on the seawater area to be measured, and the acquired distributed Raman signal is added as a new row to the last row of the two-dimensional temperature matrix after the secondary denoising to form the two-dimensional temperature matrix of the next time frame, and the two-dimensional temperature matrix of the next time frame is denoised by the non-local mean algorithm, and the last row in the denoising result is used as the current temperature of the seawater area to be measured.

[0007] Furthermore, in step S1, each Raman signal is sampled multiple times and averaged to obtain a distributed Raman signal.

[0008] Furthermore, step S2 is specifically as follows: Adjust the smoothing parameters of the non-local mean algorithm, perform denoising on the two-dimensional temperature matrix using the non-local mean algorithm corresponding to each smoothing parameter, calculate the signal-to-noise ratio of each denoising result, and obtain the smoothing parameter with the highest signal-to-noise ratio; Dynamically adjust the smoothing parameter according to the ocean current position of the seawater area to be measured and the temperature fluctuation of the seawater area to be measured; The two-dimensional temperature matrix is ​​denoised using the non-local means algorithm corresponding to the dynamically adjusted smoothing parameter.

[0009] Furthermore, dynamically adjusting the smoothing parameter according to the ocean current position of the seawater area to be measured and the temperature fluctuation of the seawater area to be measured includes: If the ocean current position where the seawater area to be measured is located is in a local end flow area, then the smoothing parameter is reduced, otherwise the smoothing parameter is increased; If the temperature fluctuation around the seawater area to be measured is greater than a preset threshold, the smoothing parameter is reduced; otherwise, the smoothing parameter is increased.

[0010] Further, adjusting the smoothing parameters of the non-local means algorithm includes: The smoothing parameter is increased with a step size of 0.1, an initial value of 0.5, and a maximum value of 10.

[0011] Furthermore, the similarity window size of the non-local means algorithm is defined as , where N is a positive integer; Before denoising the two-dimensional temperature matrix after the primary denoising by the total variation denoising algorithm, it includes: Divide into blocks.

[0012] Further, step S3 includes: Adjust the regularization parameters of the total variation denoising algorithm, denoise the two-dimensional temperature matrix after the primary denoising with the total variation denoising algorithm corresponding to each regularization parameter, calculate the signal-to-noise ratio of each denoising result, and obtain the regularization parameter with the highest signal-to-noise ratio; Dynamically adjust the regularization parameter according to the ocean current position of the seawater area to be measured and the temperature fluctuation of the seawater area to be measured; The two-dimensional temperature matrix after the primary denoising is denoised by using the total variation denoising algorithm corresponding to the dynamically adjusted regularization parameter.

[0013] Furthermore, dynamically adjusting the regularization parameter according to the ocean current position of the seawater area to be measured and the temperature fluctuation of the seawater area to be measured includes: If the ocean current position where the temperature-measured seawater area is located is in a local end flow area, then the regularization parameter is reduced; otherwise, the regularization parameter is increased; If the ambient temperature fluctuation of the seawater area to be measured is greater than a preset threshold, the regularization parameter is reduced; otherwise, the regularization parameter is increased.

[0014] Furthermore, adjusting the regularization parameter of the total variation denoising algorithm includes: The regularization parameter is increased with a step size of 0.1, 0.1 as the initial value of the regularization parameter, and 10 as the maximum value of the regularization parameter.

[0015] Therefore, the present invention provides the following effects and / or advantages: The present application collects one-dimensional Raman signals to form a two-dimensional temperature matrix, and then retains the overall temperature trend of the seawater area to be measured in the two-dimensional temperature matrix through a non-local mean algorithm, and retains the temperature change area details in the two-dimensional temperature matrix after the first denoising through a total variation denoising algorithm, so as to obtain a temperature field including the historical trend and details of the seawater temperature, and performs new denoising in combination with the newly acquired Raman signal in the next time frame, so as to obtain the current temperature of the seawater area to be measured.

[0016] The method provided in this application improves the resolution of distributed seawater temperature detection to ±0.30°C, which can achieve high-resolution monitoring of ocean temperature for red tide warning, etc. At the same time, the method provided in this application has a short response time, which meets the real-time requirements.

[0017] The method provided in the present application dynamically adjusts the parameters in the non-local mean algorithm and the total variation denoising algorithm according to the ocean current position of the seawater area to be measured, the temperature fluctuation of the seawater area to be measured, etc., so as to adapt to the short-term changes in seawater temperature, retain the changing trend of seawater temperature, and filter out noise.

[0018] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description and the drawings.

[0019] It is to be understood that both the foregoing general description and the following detailed description of the present invention are exemplary and explanatory and are intended to provide further explanation of the invention as claimed. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A flowchart diagram of one embodiment of the present invention is provided.

[0021] Figure 2 Schematic diagram of the raw data of distributed Raman signals.

[0022] Figure 3 This is a curve obtained after the Raman signal is denoised using the method provided by the present application at 40°C.

[0023] Figure 4 This is a curve obtained after the Raman signal is denoised using the method provided by the present application at 60°C.

[0024] Figure 5 This is a curve obtained after the Raman signal is denoised using the method provided by the present application at 80°C.

[0025] Figure 6 Schematic diagram of the noise reduction results of various algorithms. DETAILED DESCRIPTION

[0026] In order to facilitate understanding by those skilled in the art, the present invention is now described in further detail by way of examples: refer to Figure 1 , a high-resolution distributed seawater temperature monitoring method, comprising the following steps: S1, collecting distributed Raman signals from the seawater area to be measured in a plurality of consecutive time frames, and forming a two-dimensional temperature matrix with a plurality of groups of distributed Raman signals obtained in a plurality of consecutive time frames, wherein the rows of the two-dimensional temperature matrix represent the changes of the Raman signals at different distance positions, and the columns of the two-dimensional temperature matrix represent the changes of the Raman signals at different time frames; Furthermore, in step S1, each Raman signal is sampled multiple times and averaged to obtain a distributed Raman signal.

[0027] In this step, the temperature of the seawater area to be measured is obtained by using a Raman distributed fiber optic temperature sensor. A plurality of gratings are arranged on a fiber optic temperature sensor, and each grating can obtain a temperature signal, thereby obtaining temperature signals corresponding to multiple positions on a fiber optic temperature sensor. In this embodiment, the construction method of the two-dimensional temperature matrix is ​​to continuously obtain Raman signals within 20 groups of time frames at the same test temperature, and each Raman signal is obtained by 5000 cumulative averages. These signals are then stacked in time series to form a two-dimensional temperature matrix, in which the rows of the two-dimensional temperature matrix represent the changes of the Raman signals at different distances, and the columns correspond to different time frames of the Raman signals. Specifically, sampling is performed on a sensing optical fiber of about 2.3 km, and the system sampling interval is 0.4 m. The number of sampling points for each Raman signal is 5635, the total number of signal sequences is 20, and the size of the combined two-dimensional temperature matrix is ​​20×5635. The cross-sectional view of the two-dimensional temperature matrix is ​​shown in FIG. Figure 2 As shown, Figure 2 The original two-dimensional temperature matrix obtained in this way is shown. These traces contain 20 frames of original Raman signals, and 14m of the optical fiber is placed in a water environment with a fixed temperature. Figure 2 The 14m long data signal is magnified and displayed.

[0028] S2, denoising the two-dimensional temperature matrix by a non-local means algorithm, retaining the overall temperature trend of the seawater area to be measured in the two-dimensional temperature matrix, and obtaining a two-dimensional temperature matrix after one denoising; The non-local means algorithm reduces noise by considering pixels in similar areas in a two-dimensional temperature matrix. It mainly relies on the similarity between pixels rather than the pixel values ​​in the local area. It can effectively remove noise in the two-dimensional temperature matrix and retain details, thereby retaining the overall temperature trend of the seawater area to be measured in the two-dimensional temperature matrix.

[0029] S3, denoising the two-dimensional temperature matrix after the primary denoising by using a total variation denoising algorithm, retaining the temperature change area details in the two-dimensional temperature matrix after the primary denoising, and obtaining the two-dimensional temperature matrix after the secondary denoising; The total variation denoising algorithm denoises by minimizing the total variation of the image (i.e., the sum of the pixel value gradients). It can effectively reduce the noise in the two-dimensional temperature matrix after the primary denoising and is excellent in maintaining the edges or textures corresponding to the seawater temperature changes, thereby retaining the temperature change area details in the two-dimensional temperature matrix after the primary denoising.

[0030] Furthermore, this embodiment can preliminarily reduce the fiber noise in the two-dimensional temperature matrix by first using the non-local mean algorithm for denoising, thereby improving the quality of the temperature data in the two-dimensional temperature matrix and making the subsequent total variation denoising algorithm more stable and efficient. At the same time, first performing the non-local mean algorithm denoising and then performing the total variation denoising algorithm for denoising can prevent the total variation denoising algorithm from smoothing out the details in the two-dimensional temperature matrix, especially the high-frequency part. When the non-local mean algorithm is then used for denoising, the effect of the non-local mean algorithm may be reduced because the two-dimensional temperature matrix has lost some detail information. Therefore, performing the total variation denoising first may cause the image to lose detail retention, making the subsequent non-local mean algorithm denoising effect less than expected.

[0031] S4, in the next time frame, distributed Raman signal acquisition is performed on the seawater area to be measured, and the acquired distributed Raman signal is added as a new row to the last row of the two-dimensional temperature matrix after the secondary denoising to form the two-dimensional temperature matrix of the next time frame, and the two-dimensional temperature matrix of the next time frame is denoised by the non-local mean algorithm, and the last row in the denoising result is used as the current temperature of the seawater area to be measured.

[0032] Specifically, on the basis of the original two-dimensional temperature matrix, a new Raman signal needs to be added to the two-dimensional Raman data matrix denoised by the non-local mean algorithm, so that the size of the new data matrix is ​​equal to 21× 5635. The first 20 data are the two-dimensional temperature matrix after secondary denoising obtained in step S3, and the 21st added data is the original Raman signal, which is then denoised by the non-local mean algorithm. After denoising, the 21st data is the current temperature distribution of the sensing fiber.

[0033] Since the newly acquired Raman signal is added as a new row to the two-dimensional matrix, the time span of the matrix is ​​expanded. The non-local means algorithm can more effectively suppress transient noise (such as sensor random errors) by searching for similar patterns across time frames (such as periodic temperature fluctuations), while retaining the overall trend of temperature evolution, thereby maximizing the use of information on changes in seawater temperature in the time dimension, retaining the spatiotemporal coherence of the temperature field, inheriting historical noise reduction results, and ensuring that the newly added data is consistent with historical trends.

[0034] Furthermore, step S2 is specifically as follows: Adjust the smoothing parameters of the non-local mean algorithm, perform denoising on the two-dimensional temperature matrix using the non-local mean algorithm corresponding to each smoothing parameter, calculate the signal-to-noise ratio of each denoising result, and obtain the smoothing parameter with the highest signal-to-noise ratio; Dynamically adjust the smoothing parameter according to the ocean current position of the seawater area to be measured and the temperature fluctuation of the seawater area to be measured; The two-dimensional temperature matrix is ​​denoised using the non-local means algorithm corresponding to the dynamically adjusted smoothing parameter.

[0035] In this step, the non-local mean algorithm is a prior art, and the smoothing parameter is a Gaussian kernel attenuation parameter in the weight calculation of the non-local mean algorithm, which determines the sensitivity of the weights between similar blocks. The larger the smoothing parameter, the stronger the smoothing effect but the details may be blurred, otherwise more details can be retained but the noise reduction ability may be insufficient. Since in the beginning, how the smoothing parameter of the non-local mean algorithm is set is positional, all smoothing parameters are traversed within a certain range to adjust the noise reduction effect of the non-local mean algorithm on the two-dimensional temperature matrix, and the smoothing parameter corresponding to the best signal-to-noise ratio is used as the basic parameter. Then, since the seawater temperature usually changes more slowly and is affected by factors such as seasonality, depth, and ocean flow, the seawater temperature changes rapidly when the seasons alternate, the deeper the seawater, the more the seawater temperature changes, and the seawater temperature is prone to more short-term changes when the ocean current is in the local end flow area. These situations will increase the appearance of noise and produce a stronger temperature change trend. Therefore, it is necessary to dynamically adjust the smoothing parameter so that the two-dimensional temperature matrix after noise reduction can retain the trend and details of the seawater temperature at multiple levels.

[0036] Furthermore, dynamically adjusting the smoothing parameter according to the ocean current position of the seawater area to be measured and the temperature fluctuation of the seawater area to be measured includes: If the ocean current position where the seawater area to be measured is located is in a local end flow area, then the smoothing parameter is reduced, otherwise the smoothing parameter is increased; If the temperature fluctuation around the seawater area to be measured is greater than a preset threshold, the smoothing parameter is reduced; otherwise, the smoothing parameter is increased.

[0037] Specifically, when the ocean current position of the temperature seawater area to be measured is in a local end flow area, the seawater temperature is prone to more short-term changes, or when the temperature around the seawater to be measured changes greatly, such as in seasons with alternating temperatures, during the time period when the sun rises or sets, when the seawater to be measured is in a shallow sea, etc., the temperature around the seawater to be measured changes greatly, which can easily cause the temperature of the seawater to be measured to change in the short term. At this time, a smaller smoothing parameter is used to retain these local short-term changes.

[0038] Specifically, if the ocean current position where the seawater area to be measured is located is in a local end flow area, the smoothing parameter is reduced by 20%, otherwise the smoothing parameter is increased by 20%; If the temperature fluctuation around the seawater area to be measured is greater than a preset threshold, the smoothing parameter is reduced by 20%, otherwise the smoothing parameter is increased by 20%.

[0039] Further, adjusting the smoothing parameters of the non-local means algorithm includes: The smoothing parameter is increased with a step size of 0.1, an initial value of 0.5, and a maximum value of 10.

[0040] This method can gradually adjust the filtering parameters through small steps and observe the impact of different smoothing parameters on the denoising effect.

[0041] Furthermore, the similarity window size of the non-local means algorithm is defined as , where N is a positive integer; Before denoising the two-dimensional temperature matrix after the primary denoising by the total variation denoising algorithm, it includes: Divide into blocks.

[0042] In this step, the similarity window size of the non-local mean algorithm can be set manually, and in this embodiment, N is 3. Since the two-dimensional temperature matrix obtained in S1 is large in scale, directly applying TV denoising to the global matrix will lead to extremely high computational complexity and memory requirements. After dividing the two-dimensional temperature matrix into smaller sub-blocks, each sub-block can be denoised independently by the total variation denoising algorithm, and multi-core CPU or GPU parallel computing is used to significantly improve the processing speed.

[0043] Further, step S3 includes: Adjust the regularization parameters of the total variation denoising algorithm, denoise the two-dimensional temperature matrix after the primary denoising with the total variation denoising algorithm corresponding to each regularization parameter, calculate the signal-to-noise ratio of each denoising result, and obtain the regularization parameter with the highest signal-to-noise ratio; Dynamically adjust the regularization parameter according to the ocean current position of the seawater area to be measured and the temperature fluctuation of the seawater area to be measured; The two-dimensional temperature matrix after the primary denoising is denoised by using the total variation denoising algorithm corresponding to the dynamically adjusted regularization parameter.

[0044] Furthermore, dynamically adjusting the regularization parameter according to the ocean current position of the seawater area to be measured and the temperature fluctuation of the seawater area to be measured includes: If the ocean current position where the temperature-measured seawater area is located is in a local end flow area, then the regularization parameter is reduced; otherwise, the regularization parameter is increased; If the ambient temperature fluctuation of the seawater area to be measured is greater than a preset threshold, the regularization parameter is reduced; otherwise, the regularization parameter is increased.

[0045] Furthermore, adjusting the regularization parameter of the total variation denoising algorithm includes: The regularization parameter is increased with a step size of 0.1, 0.1 as the initial value of the regularization parameter, and 10 as the maximum value of the regularization parameter.

[0046] In the above steps, the adjustment principle of the specific regularization parameter can refer to the adjustment principle of the smoothing parameter of the non-local means algorithm.

[0047] Experimental data When the pulse width is 40 ns, the temperature of the optical fiber is set to 40 ℃, 60 ℃, and 80 ℃ respectively, and the effective temperature measurement length of the optical fiber is 14 m, Figures 3-5 As shown in the figure, the temperature curve obtained by denoising the Raman signal using the method provided by the present application is smoother than the unfiltered temperature curve, and the measured temperature is close to the temperature of the optical fiber. It can be seen from the figure that after denoising by the method provided by the present application, the temperature accuracy of the Raman signal is improved from ±2.82 ℃ to ±0.30 ℃, effectively removing the noise in the Raman signal. And, it shows that the method provided by the present application still has a certain filtering ability and robustness at different temperatures.

[0048] refer to Figure 6 , the method provided by this application is compared with the wavelet threshold denoising algorithm, two-dimensional non-local mean algorithm, and bilateral filtering algorithm in the prior art. The temperature of the optical fiber is 60 ℃. It can be seen that the filtering result of the algorithm provided by this application is closer to the temperature of the optical fiber. And, as shown in Table 1, the temperature calculated by the algorithm provided by this application has the highest accuracy, and the corresponding time is within an acceptable range, meeting the needs of high-precision temperature measurement. Table 1. Comparison of denoising performance of different filters on raw Raman signals

[0049] In summary, the algorithm provided in this application achieves a good balance between temperature accuracy and system real-time performance. It can not only provide higher temperature measurement accuracy but also maintain a faster response time, basically meeting the real-time requirements of the RDTS system, and is more suitable for application scenarios that require both temperature measurement accuracy and response time.

[0050] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0051] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0052] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0053] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0054] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms should not be understood as necessarily being directed to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they are contradictory.

Claims

1. A high-resolution distributed seawater temperature monitoring method, characterized in that: The following steps are involved: S1, collecting distributed Raman signals from the seawater area to be measured in a plurality of consecutive time frames, and forming a two-dimensional temperature matrix with a plurality of groups of distributed Raman signals obtained in a plurality of consecutive time frames, wherein the rows of the two-dimensional temperature matrix represent the changes of the Raman signals at different distance positions, and the columns of the two-dimensional temperature matrix represent the changes of the Raman signals at different time frames; S2, denoising the two-dimensional temperature matrix by a non-local means algorithm, retaining the overall temperature trend of the seawater area to be measured in the two-dimensional temperature matrix, and obtaining a two-dimensional temperature matrix after one denoising; S3, denoising the two-dimensional temperature matrix after the primary denoising by using a total variation denoising algorithm, retaining the temperature change area details in the two-dimensional temperature matrix after the primary denoising, and obtaining the two-dimensional temperature matrix after the secondary denoising; S4, in the next time frame, distributed Raman signal acquisition is performed on the seawater area to be measured, and the acquired distributed Raman signal is added as a new row to the last row of the two-dimensional temperature matrix after the secondary denoising to form the two-dimensional temperature matrix of the next time frame, and the two-dimensional temperature matrix of the next time frame is denoised by the non-local mean algorithm, and the last row in the denoising result is used as the current temperature of the seawater area to be measured.

2. A high-resolution distributed seawater temperature monitoring method according to claim 1, characterized in that: In step S1, each Raman signal is sampled multiple times and averaged to obtain a distributed Raman signal.

3. A high-resolution distributed seawater temperature monitoring method according to claim 1, characterized in that: Step S2 is specifically as follows: Adjust the smoothing parameters of the non-local mean algorithm, perform denoising on the two-dimensional temperature matrix using the non-local mean algorithm corresponding to each smoothing parameter, calculate the signal-to-noise ratio of each denoising result, and obtain the smoothing parameter with the highest signal-to-noise ratio; Dynamically adjust the smoothing parameter according to the ocean current position of the seawater area to be measured and the temperature fluctuation of the seawater area to be measured; The two-dimensional temperature matrix is ​​denoised using the non-local means algorithm corresponding to the dynamically adjusted smoothing parameter.

4. A high-resolution distributed seawater temperature monitoring method according to claim 3, characterized in that: According to the ocean current position of the seawater area to be measured and the temperature fluctuation of the seawater area to be measured, dynamically adjusting the smoothing parameter includes: If the ocean current position where the seawater area to be measured is located is in a local end flow area, then the smoothing parameter is reduced, otherwise the smoothing parameter is increased; If the temperature fluctuation around the seawater area to be measured is greater than a preset threshold, the smoothing parameter is reduced; otherwise, the smoothing parameter is increased.

5. A high-resolution distributed seawater temperature monitoring method according to claim 4, characterized in that: Adjusting the smoothing parameters of the non-local means algorithm includes: The smoothing parameter is increased with a step size of 0.1, an initial value of 0.5, and a maximum value of 10.

6. A high-resolution distributed seawater temperature monitoring method according to claim 1, characterized in that: The similarity window size of the non-local means algorithm is defined as , where N is a positive integer; Before denoising the two-dimensional temperature matrix after the primary denoising by the total variation denoising algorithm, it includes: Divide into blocks.

7. A high-resolution distributed seawater temperature monitoring method according to claim 6, characterized in that: Step S3 includes: Adjust the regularization parameters of the total variation denoising algorithm, denoise the two-dimensional temperature matrix after the primary denoising with the total variation denoising algorithm corresponding to each regularization parameter, calculate the signal-to-noise ratio of each denoising result, and obtain the regularization parameter with the highest signal-to-noise ratio; Dynamically adjust the regularization parameter according to the ocean current position of the seawater area to be measured and the temperature fluctuation of the seawater area to be measured; The two-dimensional temperature matrix after the primary denoising is denoised by using the total variation denoising algorithm corresponding to the dynamically adjusted regularization parameter.

8. A high-resolution distributed seawater temperature monitoring method according to claim 7, characterized in that: According to the ocean current position of the seawater area to be measured and the temperature fluctuation of the seawater area to be measured, dynamically adjusting the regularization parameter includes: If the ocean current position where the temperature-measured seawater area is located is in a local end flow area, then the regularization parameter is reduced; otherwise, the regularization parameter is increased; If the ambient temperature fluctuation of the seawater area to be measured is greater than a preset threshold, the regularization parameter is reduced; otherwise, the regularization parameter is increased.

9. A high-resolution distributed seawater temperature monitoring method according to claim 7, characterized in that: The regularization parameters for adjusting the total variation denoising algorithm include: The regularization parameter is increased with a step size of 0.1, 0.1 as the initial value of the regularization parameter, and 10 as the maximum value of the regularization parameter.

Images