Optical fiber air pressure sensor and manufacturing method thereof
By connecting the reference cavity and the sensing cavity in parallel, combining the spectral analyzer and artificial intelligence signal processing technology, the spectrogram is analyzed to improve the accuracy of the fiber air pressure sensor, solving the problem of the reduction in the accuracy of the traditional fiber air pressure sensor in extreme environments, achieving high accuracy Air pressure measurement.
Patent Information
- Application Number
- CN202410738060.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-07
- Publication Date
- 2024-11-08
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Under the influence of environmental factors and temperature changes, traditional fiber optic air pressure sensors have reduced measurement accuracy, making it difficult to provide reliable air pressure data in extreme environments.
Using parallel reference cavity and sensor cavity, combined with spectral analyzer and artificial intelligence-based signal processing technology, the spectral map is analyzed to determine the barometric measurement value through spectral feature enhancement, local feature extraction, multi-scale aggregation and global feature aggregation representation modules. .
It improves the performance of fiber optic air pressure sensor, realizes high-precision measurement of air pressure changes, enhances the ability to identify and analyze the sensor output signals, and reduces interference from external factors.
Smart Images

Figure CN118913514A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the field of optical fiber air pressure sensors, and in particular to an optical fiber air pressure sensor and a manufacturing method thereof. Background Art
[0002] Fiber optic pressure sensor is a sensor that uses fiber optic technology to measure air pressure. It converts air pressure information by detecting changes in optical signals transmitted in optical fibers. Due to the advantages of optical fiber such as anti-electromagnetic interference, corrosion resistance, light weight and small size, fiber optic pressure sensor has a wide range of applications in environmental monitoring, meteorology, aerospace, deep sea exploration and other fields.
[0003] However, traditional fiber optic pressure sensors usually use the interference principle to determine the air pressure by measuring the phase or intensity changes of the optical signal. However, due to environmental factors, temperature changes, equipment aging, etc., the output signal of the sensor may be disturbed, resulting in a decrease in measurement accuracy.
[0004] Therefore, an optimized fiber optic air pressure sensor is desired. Summary of the invention
[0005] The present disclosure is made in view of the above problems. One object of the present disclosure is to provide a fiber optic air pressure sensor and a method for manufacturing the same.
[0006] An embodiment of the present disclosure provides an optical fiber air pressure sensor, comprising:
[0007] A reference cavity and a sensing cavity connected in parallel, wherein the reference cavity and the sensing cavity each have an optical signal input end;
[0008] A spectrum analyzer, wherein the optical signal output end of the spectrum analyzer is connected to the optical signal input ends of the reference cavity and the sensor cavity respectively;
[0009] The spectrum analyzer is used to analyze the spectrum graphs generated by the reference cavity and the sensing cavity to obtain the air pressure measurement value.
[0010] For example, in the optical fiber air pressure sensor according to an embodiment of the present disclosure, the optical spectrum analyzer comprises:
[0011] A spectral feature enhancement module, used for performing contrast normalization on the spectrum to obtain a detail-enhanced spectrum;
[0012] A spectral local region feature extraction module, used for performing local spectral region feature extraction on the detail enhanced spectral image to obtain a sequence of local spectral feature images;
[0013] A spectral local feature multi-scale aggregation module, used for respectively passing each local spectral feature graph in the sequence of local spectral feature graphs through the feature multi-scale aggregation module to obtain a sequence of local spectral feature focus representation vectors;
[0014] A global spectral feature information transfer aggregation representation module is used to obtain a global spectral feature transfer aggregation representation vector by transmitting the sequence of the local spectral feature focus representation vectors through a node message propagation fusion network of fusion node importance;
[0015] The air pressure measurement module is used to determine the air pressure measurement value based on the global spectral feature transfer aggregation representation vector.
[0016] For example, in the optical fiber air pressure sensor according to an embodiment of the present disclosure, the spectral local area feature extraction module includes:
[0017] A spectral image segmentation unit, used for performing image segmentation on the detail enhanced spectral image to obtain a sequence of local spectral images;
[0018] The local spectral feature extraction unit is used to pass each local spectral graph in the sequence of local spectral graphs through a spectral feature extractor based on a hole convolutional neural network model to obtain the sequence of local spectral feature graphs.
[0019] For example, in the optical fiber air pressure sensor according to an embodiment of the present disclosure, the spectral local feature multi-scale aggregation module includes:
[0020] A spectral global semantic feature extraction unit, used for extracting a spectral global semantic feature vector from the local spectral feature map;
[0021] A spectral local semantic feature extraction unit, used for extracting a spectral local semantic feature vector from the local spectral feature map;
[0022] The spectral multi-scale semantic focus representation unit is used to fuse the spectral global semantic feature vector and the spectral local semantic feature vector to obtain the local spectral feature focus representation vector.
[0023] For example, in the optical fiber air pressure sensor according to an embodiment of the present disclosure, the spectral global semantic feature extraction unit is further used to:
[0024] Upsampling the local spectral feature map to obtain an upsampled local spectral feature map;
[0025] Performing global mean pooling processing on each feature matrix along the channel dimension in the upsampled local spectral feature map to obtain an upsampled spectral semantic channel feature vector;
[0026] Point convolution encoding is performed on the upsampled spectral semantic channel feature vector to obtain the spectral global semantic feature vector.
[0027] For example, in the optical fiber air pressure sensor according to an embodiment of the present disclosure, the spectral local semantic feature extraction unit is further used to:
[0028] Upsampling the local spectral feature map to obtain an upsampled local spectral feature map;
[0029] Performing point convolution encoding on the upsampled local spectral feature map to obtain a channel-modulated upsampled local spectral feature map;
[0030] Two-dimensional convolution encoding is performed on the channel-modulated upsampled local spectral feature map to obtain the spectral local semantic feature vector.
[0031] For example, in the optical fiber air pressure sensor according to an embodiment of the present disclosure, the global spectral feature information transmission aggregation representation module includes:
[0032] A neighbor node semantic fusion unit, used for calculating the positional sum of the first (s-1) local spectral feature focused representation vectors in the sequence of the local spectral feature focused representation vectors to obtain a neighbor node semantic fusion feature vector;
[0033] A weighted optimization unit, configured to perform weighted processing on each position feature value in the s-th local spectral feature focus representation vector by using the sum of a trainable preset hyperparameter and one as a weighting coefficient to obtain a weighted optimized local spectral feature focus representation vector;
[0034] A multi-layer perception unit, used for adding the neighbor node semantic fusion feature vector and the weighted optimized local spectral feature focus representation vector by position and then processing them through a multi-layer perceptron to obtain a spectral feature information transfer aggregation feature vector corresponding to the sth local spectral feature focus representation vector;
[0035] The aggregation unit is used to calculate the positional sum between the spectral feature information transfer aggregation feature vectors corresponding to each local spectral feature focus representation vector in the sequence of the local spectral feature focus representation vector to obtain the global spectral feature transfer aggregation representation vector.
[0036] For example, according to the optical fiber air pressure sensor of an embodiment of the present disclosure, the air pressure measurement module is used to: pass the global spectral feature aggregation representation vector through a decoder-based air pressure parser to obtain the air pressure measurement value.
[0037] The embodiment of the present disclosure also provides a method for manufacturing an optical fiber air pressure sensor, which comprises:
[0038] Selecting a first capillary tube and a first single-mode optical fiber of predetermined length and flatness;
[0039] Removing the coating layer at the end of the first single-mode optical fiber, and cutting and flattening the end face of the first single-mode optical fiber to obtain a first single-mode optical fiber after end face cutting;
[0040] Filling one end of the first capillary with ultraviolet glue until the first capillary is completely filled with the colloid to obtain the first capillary filled with ultraviolet glue;
[0041] Inserting the first single-mode optical fiber after end-face cutting into the first capillary filled with UV glue, and inserting the other end of the first capillary filled with UV glue into the first single-mode optical fiber after end-face cutting, so that the space between the two end faces of the first single-mode optical fiber after end-face cutting is filled with UV glue;
[0042] Using an ultraviolet curing lamp to irradiate the ultraviolet glue-filled first capillary tube into which the first single-mode optical fiber after end-face cutting is inserted, so that the ultraviolet glue is cured under the influence of ultraviolet rays to obtain a reference cavity;
[0043] Selecting a second capillary tube and a second single-mode optical fiber of predetermined length and flatness;
[0044] Removing the coating layer of the end of the second single-mode optical fiber, and cutting and flattening the end face of the second single-mode optical fiber to obtain a second single-mode optical fiber after end face cutting;
[0045] Inserting the second single-mode optical fiber after end-face cutting into one end of the second capillary, dripping ultraviolet glue into it and curing it to obtain a cured second capillary;
[0046] The other end of the cured second capillary is inserted into the second single-mode optical fiber after the end face is cut to form an air cavity, and ultraviolet glue is dripped into it for curing to obtain a sensing cavity.
[0047] According to the optical fiber air pressure sensor and its manufacturing method of the embodiments of the present disclosure, it uses a spectrum analyzer and introduces artificial intelligence-based signal processing technology at the back end, which can analyze the spectrum generated by the reference cavity and the sensing cavity to measure the air pressure value. This can further improve the recognition and analysis capabilities of the sensor output signal, thereby improving the performance of the optical fiber air pressure sensor and achieving high-precision measurement of air pressure changes. BRIEF DESCRIPTION OF THE DRAWINGS
[0048] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the following briefly introduces the drawings of the embodiments of the present disclosure. Obviously, the drawings described below only relate to some embodiments of the present disclosure, and are not intended to limit the present disclosure.
[0049] Figure 1A schematic diagram of the structure of the optical fiber air pressure sensor in an embodiment of the present disclosure is shown;
[0050] Figure 2 A schematic diagram of the reference cavity structure in an embodiment of the present disclosure is shown;
[0051] Figure 3 A schematic diagram of the structure of the sensing cavity in an embodiment of the present disclosure is shown;
[0052] Figure 4 A schematic diagram of the structure of a spectrum analyzer of an optical fiber air pressure sensor in an embodiment of the present disclosure is shown;
[0053] Figure 5 A schematic diagram of the structure of a spectral local area feature extraction module of an optical fiber air pressure sensor in an embodiment of the present disclosure is shown;
[0054] Figure 6 A schematic diagram of the structure of a spectral local feature multi-scale aggregation module of an optical fiber pressure sensor in an embodiment of the present disclosure is shown;
[0055] Figure 7 A schematic diagram showing the structure of a global spectral feature information transmission and aggregation representation module of an optical fiber air pressure sensor in an embodiment of the present disclosure is shown; and
[0056] Figure 8 A flow chart of a method for manufacturing an optical fiber air pressure sensor in an embodiment of the present disclosure is shown. DETAILED DESCRIPTION
[0057] The technical solutions in the embodiments of the present disclosure will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present disclosure, rather than all embodiments. Based on the embodiments of the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work also fall within the scope of protection of the present disclosure.
[0058] The terms used in this specification are those general terms currently widely used in the art in consideration of the functions of the present disclosure, but these terms may vary according to the intention of a person of ordinary skill in the art, precedents, or new technologies in the art. In addition, specific terms may be selected by the applicant, and in this case, their detailed meanings will be described in the detailed description of the present disclosure. Therefore, the terms used in the specification should not be understood as simple names, but rather as an overall description based on the meaning of the terms and the present disclosure.
[0059] Although the present disclosure makes various references to certain modules in the system according to embodiments of the present disclosure, any number of different modules can be used and run on a user terminal and / or server. The modules are illustrative only, and different aspects of the systems and methods can use different modules.
[0060] Flowcharts are used in this disclosure to illustrate the operations performed by the system according to the embodiments of the present disclosure. It should be understood that the preceding or following operations are not necessarily performed precisely in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. At the same time, other operations may also be added to these processes, or one or more operations may be removed from these processes.
[0061] In addition, the application architecture diagram in the embodiment of the present disclosure is intended to more clearly illustrate the technical solution in the embodiment of the present disclosure, and does not constitute a limitation on the technical solution provided in the embodiment of the present disclosure. Of course, for other application architectures and business applications, the technical solution provided in the embodiment of the present disclosure is also applicable to similar problems.
[0062] The following is a non-restrictive description of the fiber optic air pressure sensor provided according to at least one embodiment of the present disclosure through several examples or embodiments. As described below, different features in these specific examples or embodiments can be combined with each other without conflicting with each other to obtain new examples or embodiments, and these new examples or embodiments also fall within the scope of protection of the present disclosure.
[0063] The fiber optic pressure sensor is an advanced measurement tool that uses fiber optic technology to accurately capture changes in air pressure. The sensor works by monitoring tiny changes in the light signal transmitted in the optical fiber, thereby converting air pressure information into readable data. The excellent properties of optical fiber, such as high resistance to electromagnetic interference, resistance to corrosion, light weight and compact size, make fiber optic pressure sensors widely used in many fields, including environmental monitoring, meteorological research, aerospace, and deep-sea exploration.
[0064] Although fiber optic pressure sensors have many advantages, they usually work based on the principle of interferometry, which involves the precise measurement of the phase or intensity changes of light signals to determine the air pressure. However, this measurement method may be affected by many factors, such as fluctuations in environmental conditions, unstable temperature, and aging of equipment, which may interfere with the output signal of the sensor and affect the accuracy of the measurement.
[0065] This sensor needs to have higher stability and accuracy to reduce the negative impact of external factors on the measurement results. The optimized sensor will be able to provide more reliable air pressure data, thus playing a greater role in key application scenarios, such as scientific research in extreme environments or in industrial processes that require high-precision data.
[0066] In view of the above technical problems, in the technical solution of this application, an optical fiber air pressure sensor is proposed, such as Figures 1 to 3As shown, it includes: a reference cavity 100 and a sensing cavity 200 connected in parallel, wherein the reference cavity 100 and the sensing cavity 200 respectively have an optical signal input end; a spectrum analyzer 800, wherein the optical signal output end of the spectrum analyzer 800 is connected to the optical signal input end of the reference cavity 100 and the sensing cavity 200 respectively; wherein the spectrum analyzer 800 is used to analyze the spectrum graphs generated by the reference cavity 100 and the sensing cavity 200 to obtain the air pressure measurement value. Through the application of the spectrum analyzer and the introduction of artificial intelligence-based signal processing technology at the back end, the spectrum graphs generated by the reference cavity and the sensing cavity can be analyzed to measure the air pressure value, which can further improve the recognition and analysis capabilities of the sensor output signal, thereby improving the performance of the optical fiber air pressure sensor and achieving high-precision measurement of air pressure changes.
[0067] More specifically, the optical fiber pressure sensor includes a reference cavity 100 and a sensing cavity 200 connected in parallel. The reference cavity 100 includes a first capillary 110. First single-mode optical fibers 120 are inserted into both ends of the first capillary 110. The first capillary 110 is filled with ultraviolet glue between the end faces of the first single-mode optical fibers 120 at both ends to form a glue cavity. The sensing cavity 200 includes a second capillary 210. Second single-mode optical fibers 220 are inserted into both ends of the second capillary 210. Air is filled into the second capillary 210 between the end faces of the second single-mode optical fibers 220 at both ends to form an air cavity. The splicing interface between the second capillary 210 and the second single-mode optical fiber 220 is sealed by ultraviolet glue.
[0068] Among them, one end of the first single-mode optical fiber 120 inserted into the first capillary 110 and one end of the second single-mode optical fiber 220 inserted into the second capillary 21 have no coating layer and the end surface is flat. The depth of the first single-mode optical fiber 120 inserted into the first capillary 110 is 4-6mm; the depth of the second single-mode optical fiber 220 inserted into the second capillary 210 is 4-6mm. The first capillary 210 and the second capillary 210 are made of quartz capillaries. The joint between the second single-mode optical fiber 220 and the second capillary 210 is sealed by 0.03-0.04ml ultraviolet glue.
[0069] It should be understood that the capillary is made of quartz material, which is a material with high transparency, low thermal expansion coefficient and good chemical stability, and is suitable for optical applications. In one example, the single-mode optical fiber model is Corning SNF-28 (inner diameter is 8.2μm, cladding diameter is 125μm), the selected capillary specifications are inner diameter 150μm, outer diameter 365μm, and the length of the capillary is between 10mm and 15mm. The specific length can be adjusted according to needs. The length of the glue cavity in the reference cavity 100 is similar to but not equal to the length of the air cavity in the sensing cavity 200.
[0070] The reference cavity 100 and the sensing cavity 200 are placed in the pressure gauge 500 . The output ends of the reference cavity 100 and the sensing cavity 200 are respectively connected to the signal input end of the spectrum analyzer 800 through the coupler 300 , and the signal output end of the spectrum analyzer 800 is connected to the computer 600 .
[0071] Furthermore, if Figure 4 As shown, the spectrum analyzer 800 includes: a spectral feature enhancement module 810, which is used to perform contrast normalization on the spectrum to obtain a detail enhanced spectrum; a spectral local area feature extraction module 820, which is used to perform local spectral area feature extraction on the detail enhanced spectrum to obtain a sequence of local spectral feature maps; a spectral local feature multi-scale aggregation module 830, which is used to pass each local spectral feature map in the sequence of local spectral feature maps through a feature multi-scale aggregation module to obtain a sequence of local spectral feature focus representation vectors; a global spectral feature information transfer aggregation representation module 840, which is used to pass the sequence of local spectral feature focus representation vectors through a node message propagation fusion network that fused node importance to obtain a global spectral feature transfer aggregation representation vector; and an air pressure measurement module 850, which is used to determine the air pressure measurement value based on the global spectral feature transfer aggregation representation vector.
[0072] Specifically, in the spectrometer, first of all, considering that the contrast of the spectrogram may be uneven due to the influence of ambient light during the acquisition of the spectrogram, and the local area may be brighter or darker, therefore, in the technical solution of the present application, it is necessary to perform contrast normalization on the spectrogram to obtain a detail-enhanced spectrogram. It should be understood that contrast normalization is an image processing technology that can extract local contrast features of an image, adjust the contrast of the image to reduce the impact of global contrast changes in the spectrogram, and highlight the details and texture information of the spectrogram, so that the details of the image are more clearly visible. In the spectrogram, this helps to highlight those pressure change signals that may not be obvious due to noise or environmental interference. Specifically, the contrast normalization operation of the spectrogram can be achieved by subtracting the local mean from each pixel value of the spectrogram and dividing it by the local standard deviation. By enhancing the details in the spectrogram through the contrast normalization operation, different spectral features can be more easily identified and distinguished, which is crucial for subsequent spectral feature extraction and air pressure measurement.
[0073] Then, since the spectrum of the optical fiber pressure sensor contains multiple local areas, each area may contain specific information related to the change in air pressure. Therefore, in order to be able to decompose the complex spectrum into smaller parts, so as to analyze the semantics of the spectrum more carefully and comprehensively to identify and parse the characteristics of air pressure changes, in the technical solution of the present application, the detail enhanced spectrum is further segmented to obtain a sequence of local spectrums, and each local spectrum in the sequence of local spectrums is respectively passed through a spectral feature extractor based on a hole convolutional neural network model to obtain a sequence of local spectral feature maps. It should be understood that the hole convolutional neural network is a deep learning model for processing image data, which is characterized by being able to increase the effective receptive field of the network while maintaining the size of the receptive field, thereby capturing a wider range of contextual information, which helps to extract more accurate and comprehensive spectral local area features in the local spectrum, identify subtle changes and patterns in the air pressure in the spectrum, and thus achieve more accurate perception and analysis of air pressure changes.
[0074] Accordingly, if Figure 5 As shown, the spectral local area feature extraction module 820 includes: a spectral image segmentation unit 821, used to perform image segmentation on the detail enhanced spectrogram to obtain a sequence of local spectrograms; a local spectral feature extraction unit 822, used to pass each local spectrogram in the sequence of local spectrograms through a spectral feature extractor based on a void convolutional neural network model to obtain a sequence of local spectral feature maps.
[0075] Furthermore, considering that the pressure changes in the spectrogram may be manifested at different scales of the spectrogram, in order to extract richer and more distinctive feature representations from the local spectrogram, it is helpful to improve the accuracy of subsequent pressure analysis. In the technical solution of the present application, each local spectral feature map in the sequence of the local spectral feature map is further processed by a feature multi-scale aggregation module to obtain a sequence of local spectral feature focus representation vectors. Through the processing of the feature multi-scale aggregation module, the feature information of different scales in the local area of the spectrogram can be integrated, that is, the local semantic feature information of the spectrum at different scales can be integrated and focused through the feature multi-scale aggregation module, so that the final feature representation is more representative and discriminative. Such a processing method helps to highlight the key features in the spectral semantics and improve the sensitivity and accuracy to the pressure changes. At the same time, the processing method of multi-scale feature aggregation helps to enhance the stability of the local semantic features of the spectrum, and can maintain a good feature expression even when the spectrogram is interfered by noise or some information is missing.
[0076] Accordingly, if Figure 6As shown, the spectral local feature multi-scale aggregation module 830 includes: a spectral global semantic feature extraction unit 831, used to extract a spectral global semantic feature vector from the local spectral feature map; a spectral local semantic feature extraction unit 832, used to extract a spectral local semantic feature vector from the local spectral feature map; a spectral multi-scale semantic focus representation unit 833, used to fuse the spectral global semantic feature vector and the spectral local semantic feature vector to obtain the local spectral feature focus representation vector.
[0077] Among them, the spectral global semantic feature extraction unit 831 is further used to: upsample the local spectral feature map to obtain an upsampled local spectral feature map; perform global mean pooling processing on each feature matrix along the channel dimension in the upsampled local spectral feature map to obtain an upsampled spectral semantic channel feature vector; perform point convolution encoding on the upsampled spectral semantic channel feature vector to obtain the spectral global semantic feature vector.
[0078] Among them, the spectral local semantic feature extraction unit 832 is further used to: upsample the local spectral feature map to obtain an upsampled local spectral feature map; perform point convolution encoding on the upsampled local spectral feature map to obtain a channel modulation upsampled local spectral feature map; perform two-dimensional convolution encoding on the channel modulation upsampled local spectral feature map to obtain the spectral local semantic feature vector.
[0079] It should be understood that since each local spectral feature focused representation vector in the sequence of the local spectral feature focused representation vector contains the semantic features of the local spectral region after multi-scale feature aggregation, the semantic features of these local spectral regions have a correlation relationship and associated semantic information based on the overall spectrum graph. At the same time, the semantic features of these local spectral regions have different importance and relevance when performing the subsequent pressure analysis task. Based on this, in the technical solution of the present application, the sequence of the local spectral feature focused representation vector is further passed through a node message propagation fusion network that fuses the node importance to obtain a global spectral feature transfer aggregation representation vector. Through the processing of the node message propagation fusion network that fuses the node importance, each local feature focused representation vector can be integrated into a global feature representation, so that the overall change trend and pattern of the spectrum graph can be captured, rather than being limited to the local area. In addition, the node message propagation fusion network can also learn and identify which nodes or features are more important in feature representation and subsequent pressure analysis tasks, which is crucial for extracting the information most relevant to pressure changes. By applying a node message propagation fusion network, the influence of noise and irrelevant features can be reduced at the global level, thereby enhancing the robustness of features in order to better describe the characteristics of the overall spectral graph.
[0080] Accordingly, if Figure 7 As shown, the global spectral feature information transfer aggregation representation module 840 includes: a neighbor node semantic fusion unit 841, which is used to calculate the positional sum of the first (s-1) local spectral feature focus representation vectors in the sequence of the local spectral feature focus representation vector to obtain a neighbor node semantic fusion feature vector; a weighted optimization unit 842, which is used to weight the feature values of each position in the s-th local spectral feature focus representation vector using the sum of a trainable preset hyperparameter and one as a weighting coefficient to obtain a weighted optimized local spectral feature focus representation vector; a multi-layer perception unit 843, which is used to process the neighbor node semantic fusion feature vector and the weighted optimized local spectral feature focus representation vector by a multi-layer perceptron after positionally summing the feature vector to obtain the spectral feature information transfer aggregation feature vector corresponding to the s-th local spectral feature focus representation vector; an aggregation unit 844, which is used to calculate the positional sum between the spectral feature information transfer aggregation feature vectors corresponding to each local spectral feature focus representation vector in the sequence of the local spectral feature focus representation vector to obtain the global spectral feature transfer aggregation representation vector.
[0081] The global spectral feature information transmission aggregation representation module 840 is used to: process the sequence of the local spectral feature focus representation vectors through the node message propagation fusion network of the fusion node importance using the following node message propagation fusion formula to obtain the global spectral feature transmission aggregation representation vector; wherein the node message propagation fusion formula is:
[0082]
[0083] Among them, V s is the sth local spectral feature focused representation vector in the sequence of local spectral feature focused representation vectors, V k is the kth local spectral feature focused representation vector in the sequence of local spectral feature focused representation vectors, ∈ is a trainable preset hyperparameter, MLP(·) represents a multi-layer perceptron, n is the number of feature vectors in the sequence of local spectral feature focused representation vectors, and V is the global spectral feature transfer aggregation representation vector.
[0084] Then, the global spectral feature transfer aggregation representation vector is passed through the decoder-based pressure analyzer to obtain the pressure measurement value. In other words, the global feature representation formed by the transfer and aggregation of the local semantic feature information of the spectrum is used to perform decoding regression to measure the pressure value. In this way, the recognition and analysis capabilities of the sensor output signal can be further improved, thereby improving the performance of the optical fiber pressure sensor and achieving high-precision measurement of pressure changes.
[0085] Correspondingly, the air pressure measurement module 850 is used to: pass the global spectral feature aggregation representation vector through a decoder-based air pressure parser to obtain the air pressure measurement value.
[0086] In the technical solution of the present application, each local spectral feature focused representation vector in the sequence of local spectral feature focused representation vectors respectively represents the spectral image semantic features of the local area determined by image segmentation of the spectral graph. When the sequence of local spectral feature focused representation vectors is passed through a node message propagation fusion network of fused node importance, its essence is to use the node message propagation fusion network of fused node importance to perform distributed fusion of the image semantics of the local image domain nodes under the full spectrum image domain. However, considering that each local spectral feature focused representation vector in the sequence of local spectral feature focused representation vectors has a more significant inconsistency in feature distribution, this will cause the overall global spectral feature transfer aggregation representation vector obtained to cause local overflow features due to the difference in image semantic propagation of local image domain nodes under the full spectrum image domain, thereby affecting the accuracy of the pressure measurement value obtained by the global spectral feature transfer aggregation representation vector through the decoder-based pressure parser.
[0087] Preferably, the air pressure measurement module 850 is further used to: calculate the self-mean representation matrix and the self-variance representation matrix of the global spectral feature transfer aggregation representation vector, wherein the value of each position of the self-mean representation matrix is the mean of a pair of eigenvalues of the two positions of the global spectral feature transfer aggregation representation vector corresponding to the position coordinates, and the value of each position of the self-variance representation matrix is the variance of a pair of eigenvalues of the two positions of the global spectral feature transfer aggregation representation vector corresponding to the position coordinates; matrix multiply the transposed vector of the global spectral feature transfer aggregation representation vector with the self-mean representation matrix to obtain a first global spectral feature transfer aggregation intermediate vector and matrix multiply the self-variance representation matrix with the global spectral feature transfer aggregation representation vector to obtain a second global spectral feature transfer aggregation intermediate vector, wherein The global spectral feature transfer aggregate representation vector is a column vector; the point sum of the first global spectral feature transfer aggregate intermediate vector and the transposed vector of the second global spectral feature transfer aggregate intermediate vector is calculated to obtain a third global spectral feature transfer aggregate intermediate vector; the matrix product of the self-mean representation matrix and the self-variance representation matrix is calculated, and the transposed vector of the global spectral feature transfer aggregate representation vector and the matrix product are matrix multiplied to obtain a fourth global spectral feature transfer aggregate intermediate vector; the point sum of the third global spectral feature transfer aggregate intermediate vector and the transposed vector of the fourth global spectral feature transfer aggregate intermediate vector is calculated to obtain a corrected global spectral feature transfer aggregate representation vector; and the corrected global spectral feature transfer aggregate representation vector is passed through a decoder-based pressure parser to obtain an air pressure measurement value.
[0088] Based on this, the self-mean representation matrix and the self-variance representation matrix of the group aggregation statistical evaluation of the local distribution of the feature value granularity of the global spectral feature transfer aggregation representation vector are used as the retrieval and response distribution enhancement of the global spectral feature transfer aggregation representation vector, and a reference-free distribution retrieval response framework based on the group aggregation feature distribution open domain of the global spectral feature transfer aggregation representation vector is constructed to avoid the distribution response redundancy caused by the local overflow characteristics of the global spectral feature transfer aggregation representation vector through response superposition, so as to realize the faithfulness constraint of the self-aggregation statistics correlation of the response of the global spectral feature transfer aggregation representation vector to the target decoding regression domain, and improve the accuracy of the air pressure measurement value obtained by the decoder-based air pressure parser of the corrected global spectral feature transfer aggregation representation vector. In this way, the intelligent air pressure recognition and analysis of the optical fiber air pressure sensor can be realized, thereby improving the performance of the optical fiber air pressure sensor and achieving high-precision measurement of air pressure changes.
[0089] Based on the above embodiments, see Figure 8FIG. 1 is a flow chart of a method for manufacturing an optical fiber pressure sensor according to an embodiment of the present disclosure. Figure 8 As shown, the manufacturing method of the optical fiber pressure sensor according to the embodiment of the present disclosure includes the following steps: S510, selecting a first capillary tube and a first single-mode optical fiber of predetermined length and flatness; S520, removing the end coating layer of the first single-mode optical fiber, and cutting the end face of the first single-mode optical fiber flat to obtain the first single-mode optical fiber after the end face is cut; S530, filling ultraviolet glue at one end of the first capillary tube until the colloid completely fills the first capillary tube to obtain the first capillary tube filled with ultraviolet glue; S540, inserting the first single-mode optical fiber after the end face is cut into the first capillary tube filled with ultraviolet glue, and inserting the other end of the first capillary tube filled with ultraviolet glue into the first single-mode optical fiber after the end face is cut, so that the space between the two end faces of the first single-mode optical fiber after the end face is filled with ultraviolet glue. Outer glue; S550, use an ultraviolet curing lamp to irradiate the ultraviolet glue-filled first capillary inserted with the first single-mode optical fiber after end face cutting, so that the ultraviolet glue is cured under the influence of ultraviolet rays to obtain a reference cavity; S560, select a second capillary and a second single-mode optical fiber of predetermined length and flatness; S570, remove the end coating layer of the second single-mode optical fiber, and cut the end face of the second single-mode optical fiber flat to obtain the second single-mode optical fiber after end face cutting; S580, insert the second single-mode optical fiber after end face cutting into one end of the second capillary, drip in ultraviolet glue and cure to obtain the cured second capillary; S590, insert the second single-mode optical fiber after end face cutting into the other end of the cured second capillary to form an air cavity, and drip in ultraviolet glue and cure to obtain a sensing cavity.
[0090] It is worth mentioning that the following points should be paid special attention to during the preparation of the optical fiber pressure sensor: (1) The end face of the optical fiber inserted into the capillary tube serves as a reflective surface, and the flatness of the end face of the optical fiber needs to be ensured. Therefore, when the single-mode optical fiber contacts the capillary tube, damage to the end face of the optical fiber should be avoided; (2) When making the glue cavity, the UV glue filled is in a non-viscous liquid state to avoid the UV glue not being able to be completely filled in the capillary tube. At the same time, the experimental environment should be carried out in a dark place as much as possible to avoid the phenomenon of colloid solidification during the preparation process; (3) When making the reference cavity and the sensor cavity, it is necessary to ensure that the lengths of the two cavities are similar but not equal (in the present invention, the difference in the lengths of the two cavities is in the range of 0.02635-0.02636um). The length of the cavity is adjusted by observing the free spectral range formed by the two (the cavity length is related to the free spectral range formed) until the free spectral range formed by the two is similar (in the present invention, the difference in the number of peaks of the two spectra in the spectral band with an interval of 20nm is less than one), so as to avoid the spectrum corresponding to the cursor effect not appearing after parallel connection.
[0091] Here, those skilled in the art will appreciate that the specific operations of each step in the above-mentioned method for manufacturing the optical fiber air pressure sensor have been described in detail above. Figures 1 to 7 The optical fiber air pressure sensor has been described in detail, and therefore, its repeated description will be omitted.
[0092] Furthermore, a method for testing an optical fiber pressure sensor is provided, which specifically comprises the following steps: connecting the optical signal input end of the reference cavity 100 to the first port of the optical signal output end of the spectrum analyzer 800, connecting the Ethernet port output end of the spectrum analyzer 800 to the Ethernet port input end of the computer 600, placing the reference cavity 100 in a pressure gauge 500, and displaying the interference fringe pattern generated by the reference cavity 100 acting alone under normal atmospheric pressure or the spectrum pattern of the wavelength changing with the atmospheric pressure when the reference cavity 100 acts alone through the computer 600; connecting the sensing cavity 2 The optical signal input end of the reference cavity 100 is connected to the second port of the optical signal output end of the spectrum analyzer 800, the Ethernet port output end of the spectrum analyzer 800 is connected to the Ethernet port input end of the computer 600, the sensor cavity 200 is placed in the pressure gauge 500, and the interference fringe pattern generated by the sensor cavity 200 acting alone under normal atmospheric pressure or the spectrum diagram of the wavelength changing with the atmospheric pressure when the sensor cavity 200 acts alone is displayed by the computer 600; the optical signal input ends of the reference cavity 100 and the sensor cavity 200 are connected to the optical signal of the spectrum analyzer 800 through the coupler 300 respectively. The first port of the output end is connected, the Ethernet port output end of the spectrum analyzer 800 is connected to the Ethernet port input end of the computer 600, and finally the reference cavity 100 and the sensor cavity 200 are placed in the pressure gauge 500, and the computer 600 records and displays the interference fringe pattern generated after the cursor effect under normal atmospheric pressure or the spectrum diagram of the wavelength changing with the air pressure after the cursor effect; by observing the spectrum diagram of the wavelength changing with the air pressure when the reference cavity 100 acts alone, it can be known that the structure is insensitive to the change of pressure, so it has a certain stability, which meets the premise that the reference cavity is insensitive to the measurement environment required by the vernier effect; observe the spectrum diagram of the wavelength changing with the air pressure when the sensor cavity 200 acts alone, record the position of the peak or trough in the spectrum under different air pressure conditions, and obtain the sensitivity and linearity of the structure through data analysis; observe the spectrum diagram of the wavelength changing with the air pressure after the vernier effect, record the position of the peak or trough in the spectrum under different air pressure conditions, and obtain the sensitivity and linearity of the structure through data analysis; compare the sensitivity of the sensor cavity 200 when it acts alone and when the vernier effect is generated, and obtain the multiple of the sensor sensitivity amplification.
[0093] Based on the above embodiments, another exemplary embodiment of an electronic device is also provided in the embodiments of the present disclosure. In some possible implementations, the electronic device in the embodiments of the present disclosure may include a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the steps of the method for manufacturing the optical fiber air pressure sensor in the above embodiments can be implemented when the processor executes the program.
[0094] The embodiment of the present disclosure also provides a computer-readable storage medium, on which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, the manufacturing method of the optical fiber pressure sensor according to the embodiment of the present disclosure described with reference to the above figures can be executed. The computer-readable storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. The volatile memory may, for example, include a random access memory (RAM) and / or a cache memory (cache), etc. The non-volatile memory may, for example, include a read-only memory (ROM), a hard disk, a flash memory, etc.
[0095] The embodiments of the present disclosure also provide a computer program product or a computer program, which includes computer executable instructions, and the computer executable instructions are stored in a computer readable storage medium. The processor of the computer device reads the computer executable instructions from the computer readable storage medium, and the processor executes the computer executable instructions, so that the computer device executes the manufacturing method of the optical fiber air pressure sensor according to the embodiment of the present disclosure.
[0096] Those skilled in the art will appreciate that the contents disclosed in this disclosure may be subject to various modifications and improvements. For example, the various devices or components described above may be implemented by hardware, or by software, firmware, or a combination of some or all of the three.
[0097] In addition, although the present disclosure makes various references to certain units in the system according to embodiments of the present disclosure, any number of different units may be used and run on the client and / or server. The units are illustrative only, and different aspects of the systems and methods may use different units.
[0098] Those skilled in the art will appreciate that all or part of the steps in the above method can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk. Optionally, all or part of the steps in the above embodiment can also be implemented using one or more integrated circuits. Accordingly, each module / unit in the above embodiment can be implemented in the form of hardware or in the form of a software functional module. The present disclosure is not limited to any particular form of combination of hardware and software.
[0099] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present disclosure belongs. It should also be understood that terms such as those defined in common dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology and should not be interpreted in an idealized or highly formal sense, unless explicitly defined as such herein.
[0100] The above is an explanation of the present disclosure and should not be considered as a limitation thereof. Although several exemplary embodiments of the present disclosure have been described, it will be readily understood by those skilled in the art that many modifications may be made to the exemplary embodiments without departing from the novel teachings and advantages of the present disclosure. Therefore, all such modifications are intended to be included within the scope of the present disclosure as defined in the claims. It should be understood that the above is an explanation of the present disclosure and should not be considered as being limited to the specific embodiments disclosed, and modifications to the disclosed embodiments and other embodiments are intended to be included within the scope of the appended claims. The present disclosure is defined by the claims and their equivalents.
Claims
1. An optical fiber air pressure sensor, characterized in that: include: A reference cavity and a sensing cavity connected in parallel, wherein the reference cavity and the sensing cavity each have an optical signal input end; A spectrum analyzer, wherein the optical signal output end of the spectrum analyzer is connected to the optical signal input ends of the reference cavity and the sensor cavity respectively; The spectrum analyzer is used to analyze the spectrum graphs generated by the reference cavity and the sensing cavity to obtain the air pressure measurement value.
2. The optical fiber air pressure sensor according to claim 1, characterized in that: The spectrum analyzer comprises: A spectral feature enhancement module, used for performing contrast normalization on the spectrum to obtain a detail-enhanced spectrum; A spectral local region feature extraction module, used for performing local spectral region feature extraction on the detail enhanced spectral image to obtain a sequence of local spectral feature images; A spectral local feature multi-scale aggregation module, used for respectively passing each local spectral feature graph in the sequence of local spectral feature graphs through the feature multi-scale aggregation module to obtain a sequence of local spectral feature focus representation vectors; A global spectral feature information transfer aggregation representation module is used to obtain a global spectral feature transfer aggregation representation vector by transmitting the sequence of the local spectral feature focus representation vectors through a node message propagation fusion network of fusion node importance; The air pressure measurement module is used to determine the air pressure measurement value based on the global spectral feature transfer aggregation representation vector.
3. The optical fiber air pressure sensor according to claim 2, characterized in that: The spectral local area feature extraction module comprises: A spectral image segmentation unit, used for performing image segmentation on the detail enhanced spectral image to obtain a sequence of local spectral images; The local spectral feature extraction unit is used to pass each local spectral graph in the sequence of local spectral graphs through a spectral feature extractor based on a hole convolutional neural network model to obtain the sequence of local spectral feature graphs.
4. The optical fiber air pressure sensor according to claim 3, characterized in that: The spectral local feature multi-scale aggregation module includes: A spectral global semantic feature extraction unit, used for extracting a spectral global semantic feature vector from the local spectral feature map; A spectral local semantic feature extraction unit, used for extracting a spectral local semantic feature vector from the local spectral feature map; The spectral multi-scale semantic focus representation unit is used to fuse the spectral global semantic feature vector and the spectral local semantic feature vector to obtain the local spectral feature focus representation vector.
5. The optical fiber air pressure sensor according to claim 4, characterized in that: The spectral global semantic feature extraction unit is further used to: Upsampling the local spectral feature map to obtain an upsampled local spectral feature map; Performing global mean pooling processing on each feature matrix along the channel dimension in the upsampled local spectral feature map to obtain an upsampled spectral semantic channel feature vector; Point convolution encoding is performed on the upsampled spectral semantic channel feature vector to obtain the spectral global semantic feature vector.
6. The optical fiber air pressure sensor according to claim 5, characterized in that: The spectral local semantic feature extraction unit is further used to: Upsampling the local spectral feature map to obtain an upsampled local spectral feature map; Performing point convolution encoding on the upsampled local spectral feature map to obtain a channel-modulated upsampled local spectral feature map; Two-dimensional convolution encoding is performed on the channel-modulated upsampled local spectral feature map to obtain the spectral local semantic feature vector.
7. The optical fiber air pressure sensor according to claim 6, characterized in that: The global spectral feature information transmission aggregation representation module includes: A neighbor node semantic fusion unit, used for calculating the positional sum of the first (s-1) local spectral feature focused representation vectors in the sequence of the local spectral feature focused representation vectors to obtain a neighbor node semantic fusion feature vector; A weighted optimization unit, configured to perform weighted processing on each position feature value in the s-th local spectral feature focus representation vector by using the sum of a trainable preset hyperparameter and one as a weighting coefficient to obtain a weighted optimized local spectral feature focus representation vector; A multi-layer perception unit, used for adding the neighbor node semantic fusion feature vector and the weighted optimized local spectral feature focus representation vector by position and then processing them through a multi-layer perceptron to obtain a spectral feature information transfer aggregation feature vector corresponding to the sth local spectral feature focus representation vector; The aggregation unit is used to calculate the positional sum between the spectral feature information transfer aggregation feature vectors corresponding to each local spectral feature focus representation vector in the sequence of the local spectral feature focus representation vector to obtain the global spectral feature transfer aggregation representation vector.
8. The optical fiber air pressure sensor according to claim 7, characterized in that: The air pressure measurement module is used to: pass the global spectral feature aggregation representation vector through a decoder-based air pressure parser to obtain the air pressure measurement value.
9. A method for manufacturing an optical fiber air pressure sensor, characterized in that: include: Selecting a first capillary tube and a first single-mode optical fiber of predetermined length and flatness; Removing the coating layer at the end of the first single-mode optical fiber, and cutting and flattening the end face of the first single-mode optical fiber to obtain a first single-mode optical fiber after end face cutting; Filling one end of the first capillary with ultraviolet glue until the first capillary is completely filled with the colloid to obtain the first capillary filled with ultraviolet glue; Inserting the first single-mode optical fiber after end-face cutting into the first capillary filled with UV glue, and inserting the other end of the first capillary filled with UV glue into the first single-mode optical fiber after end-face cutting, so that the space between the two end faces of the first single-mode optical fiber after end-face cutting is filled with UV glue; Using an ultraviolet curing lamp to irradiate the ultraviolet glue-filled first capillary tube into which the first single-mode optical fiber after end-face cutting is inserted, so that the ultraviolet glue is cured under the influence of ultraviolet rays to obtain a reference cavity; Selecting a second capillary tube and a second single-mode optical fiber of predetermined length and flatness; Removing the coating layer of the end of the second single-mode optical fiber, and cutting and flattening the end face of the second single-mode optical fiber to obtain a second single-mode optical fiber after end face cutting; Inserting the second single-mode optical fiber after end-face cutting into one end of the second capillary, dripping ultraviolet glue into it and curing it to obtain a cured second capillary; The other end of the cured second capillary is inserted into the second single-mode optical fiber after the end face is cut to form an air cavity, and ultraviolet glue is dripped into it for curing to obtain a sensing cavity.
