Weighing method and system based on fiber grating sensor

Abstract

The invention provides a weighing method and system based on fiber grating sensors, the method and system are applied to a weighing device, the weighing device comprises a bearing platform and a plurality of fiber grating sensors arranged at intervals along the length direction of the bearing platform, and when a measured vehicle runs over the bearing platform, a computing device receives spectral data transmitted by the fiber grating sensors and calculates the spectral data according to the spectral data. Analyzing and calculating the spectral data to obtain a central wavelength drift characteristic and a spectral shape characteristic; the central wavelength drift feature and the spectral shape feature are input into the neural network model, accurate weight data of the detected vehicle can be directly obtained, in the process, a compensation device used for supplementing the influence of the change of the environment temperature and the vehicle speed on the system detection precision does not need to be additionally added, the complexity of the system is simplified, and the detection accuracy of the vehicle is improved. And the vehicle weight dynamic detection precision is effectively improved.

Description

Technical Field

[0001] The invention relates to the technical field of vehicle weighing, and in particular to a weighing method and a system thereof based on a fiber grating sensor. Background Art

[0002] The use of fiber grating sensors to weigh vehicles can achieve vehicle weighing without stopping the vehicle. Additional sensors need to be set up in the relevant weighing system to obtain the ambient temperature and vehicle speed to correct the weighing results. The weighing system is complex and the weighing accuracy is not high. Summary of the invention

[0003] The invention provides a weighing method and a weighing system based on a fiber grating sensor, which are used to solve the technical problem of how to simplify the weighing system and improve the weighing accuracy.

[0004] In order to solve the above technical problems, the following technical solutions are proposed:

[0005] A weighing method based on a fiber Bragg grating sensor, the weighing method is applied to a weighing device, the weighing device comprises a load-bearing platform and a plurality of fiber Bragg grating sensors arranged at intervals along the length direction of the load-bearing platform, the weighing method comprises:

[0006] Acquiring spectral data of the fiber grating sensor;

[0007] Calculate the center wavelength drift characteristics and the spectrum shape characteristics according to the spectrum data;

[0008] The center wavelength drift feature and the spectrum shape feature are combined to obtain the weight data of the tested vehicle.

[0009] In some embodiments, before calculating the central wavelength drift characteristics and the spectral shape characteristics of the fiber Bragg grating sensor according to the spectral data, the method further includes: performing noise reduction processing on the spectral data through a noise reduction layer.

[0010] In some embodiments, the denoising layer includes a convolutional layer, a batch normalization layer, a ReLU activation function, a plurality of residual shrinkage modules, a batch normalization layer, a ReLU function layer, a global average pooling layer and a fully connected layer arranged in sequence.

[0011] In some implementations, acquiring the spectral data of the fiber Bragg grating sensor includes:

[0012] When the vehicle under test passes through the supporting platform, the spectrum data of the fiber grating sensor is collected multiple times to obtain the spectrum data corresponding to each fiber grating sensor at different time moments.

[0013] In some embodiments, calculating the central wavelength drift characteristics of the fiber Bragg grating sensor according to the spectral data includes:

[0014] According to the spectral data corresponding to each of the fiber grating sensors at different time moments, the sub-center wavelength drift characteristics are calculated respectively;

[0015] All the sub-center wavelength drift characteristics are passed through a fully connected layer to obtain the center wavelength drift characteristic.

[0016] In some embodiments, the spectral data corresponding to each of the fiber grating sensors at different time moments are calculated through a central wavelength drift feature extraction layer to obtain a sub-central wavelength drift feature;

[0017] The central wavelength drift feature extraction layer includes a convolution layer, a global pooling layer, a convolution layer, a global pooling layer, a Flatten layer, a ReLU function layer and a fully connected layer which are arranged in sequence.

[0018] In some embodiments, calculating the spectral shape feature according to the spectral data comprises:

[0019] Screening the spectral data in a time sequence to obtain spectral data sets corresponding to the fiber Bragg grating sensor at different times;

[0020] Inputting the spectral data sets at different times into the spectral shape feature extraction layer to obtain the position features corresponding to the different times;

[0021] Inputting the position features at different times into a spectral shape feature extraction layer to obtain spectral shape features;

[0022] Among them, the spectral shape feature extraction layer includes a forward LSTM structure layer, a reverse LSTM structure layer, a Concat layer, a forward LSTM structure layer, a reverse LSTM structure layer, a Concat layer and a fully connected layer which are arranged in sequence.

[0023] In some embodiments, the weight data is obtained by fusing the central wavelength drift feature and the spectrum shape feature through a weighing model;

[0024] Wherein, the weighing model training method comprises:

[0025] Acquiring a spectral data set, the spectral data set comprising spectral data and actual weight data, actual vehicle speed, and actual ambient temperature corresponding to the spectral data;

[0026] Inputting the spectral data into a neural network model so that the neural network model outputs a predicted weight, a predicted vehicle speed, and a predicted ambient temperature;

[0027] Calculating a loss value based on the predicted weight, predicted vehicle speed, predicted ambient temperature, and corresponding actual weight data, actual vehicle speed, and actual ambient temperature;

[0028] Adjusting the weight and threshold of each node in the neural network model according to the loss value until the loss value reaches a minimum;

[0029] Different spectral data are input into the neural network model for iterative training until the predicted weight, predicted vehicle speed, predicted ambient temperature output by the neural network model for multiple times in succession, and the loss values ​​calculated from the corresponding actual weight data, actual vehicle speed and actual ambient temperature are all less than the preset loss value threshold.

[0030] In some embodiments, the spectral data set further includes vehicle types, the spectral data is screened according to the vehicle types, and the weighing models of different vehicle types are generated;

[0031] When acquiring the spectrum data of the fiber grating sensor, it is also necessary to identify the vehicle type of the vehicle being tested, and acquire a weighing model matching the vehicle type according to the vehicle type.

[0032] The embodiment of the present invention further provides a weighing system based on a fiber grating sensor, comprising:

[0033] A weighing device, the weighing device comprising a carrying platform and a plurality of fiber grating sensors spaced apart along the length direction of the carrying platform;

[0034] An acquisition model, wherein the acquisition module is used for the spectral data of the fiber Bragg grating sensor;

[0035] A processing module is used to process the spectral data acquired by the acquisition model according to any of the above-mentioned weighing methods based on fiber grating sensors to obtain weight data of the vehicle under test.

[0036] Beneficial effects: The present application provides a weighing method and system based on a fiber grating sensor, which is applied to a weighing device, wherein the weighing device includes a load-bearing platform and a plurality of fiber grating sensors arranged at intervals along the length direction of the load-bearing platform. When the vehicle under test passes over the load-bearing platform, the fiber grating sensor senses the pressure change of the load-bearing platform and converts it into a measurable optical signal, and uses a photoelectric conversion device to convert the optical signal into spectral data. After obtaining the spectral data, the computing device obtains the center wavelength drift characteristics and the spectral shape characteristics by analyzing and calculating the spectral data; the center wavelength drift characteristics and the spectral shape characteristics are input into a neural network model, and the weight data of the vehicle under test can be directly predicted by the neural network model. In this embodiment, the weight data of the vehicle under test can be directly obtained by analyzing and calculating the spectral data obtained by the fiber grating sensor, without the need to add an additional compensation device for supplementing the changes in ambient temperature and vehicle speed that will affect the detection accuracy of the system, and the accurate weight data of the vehicle under test can be directly obtained, which not only simplifies the complexity of the system, but also effectively improves the accuracy of dynamic detection of vehicle weight. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 A structural schematic diagram of a weighing device provided by an embodiment of the present invention;

[0038] Figure 2 A flowchart of a weighing method based on a fiber grating sensor provided in an embodiment of the present invention;

[0039] Figure 3 The overall block diagram of each module for inputting spectral data provided by the embodiment of the present invention;

[0040] Figure 4 A network structure framework diagram of a feature fusion module provided in an embodiment of the present invention;

[0041] Figure 5 A network structure framework diagram of a noise processing layer provided by an embodiment of the present invention;

[0042] Figure 6 A network structure framework diagram of a residual shrinkage module provided in an embodiment of the present invention;

[0043] Figure 7 A network structure framework diagram of a central wavelength drift feature extraction layer provided in an embodiment of the present invention;

[0044] Figure 8 A network structure framework diagram of a spectral shape feature extraction layer provided in an embodiment of the present invention;

[0045] Fig. 9 A schematic diagram of spectral data acquired by a fiber grating sensor provided in an embodiment of the present invention.

[0046] Description of Reference Numerals

[0047] 10. Weighing device; 100. Carrying platform; 200. Fiber Bragg grating sensor. DETAILED DESCRIPTION

[0048] The following embodiments of the technical solution of the present application will be described in detail in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present application, and are therefore only used as examples, and cannot be used to limit the scope of protection of the present application.

[0049] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" in the specification and claims of this application and the above-mentioned figure descriptions and any variations thereof are intended to cover non-exclusive inclusions.

[0050] In the description of the embodiments of the present application, the technical terms "first", "second", etc. are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "multiple" is more than two, unless otherwise clearly and specifically defined.

[0051] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0052] In the description of the embodiments of the present application, the term "and / or" is only a description of the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0053] In the description of the embodiments of the present application, the term "multiple" refers to more than two (including two). Similarly, "multiple groups" refers to more than two groups (including two groups), and "multiple pieces" refers to more than two pieces (including two pieces).

[0054] In the description of the embodiments of the present application, the technical terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc. indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, which are only for the convenience of describing the embodiments of the present application and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the embodiments of the present application.

[0055] In the description of the embodiments of the present application, unless otherwise clearly specified and limited, technical terms such as "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, and it can be the internal connection of two elements or the interaction relationship between two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the embodiments of the present application can be understood according to the specific circumstances.

[0056] In some embodiments, Figure 1 As shown, the weighing device includes: a load-bearing platform and a fiber grating sensor. A plurality of fiber grating sensors are arranged on the load-bearing platform at intervals along the length direction of the load-bearing platform. When the vehicle passes over the load-bearing platform, different parts of the load-bearing platform produce elastic deformation in sequence. The deformation amount of the corresponding position of the load-bearing platform can be obtained through the change of the spectral signal obtained by the fiber grating sensors arranged at different positions of the load-bearing platform, so that the weight of the vehicle can be obtained through the change of the deformation amount of each part of the load-bearing platform. Among them, the fiber grating sensor can be arranged on the top surface of the load-bearing platform, and can also be buried inside the load-bearing platform.

[0057] The following is an explanation of the steps of calculating the weight data of the vehicle to be tested using the spectrum data obtained by the fiber grating sensor. Figure 2 As shown, Figure 2 A flowchart of a first weighing method based on a fiber grating sensor provided in an embodiment of the present invention, the weighing method comprises the following steps:

[0058] S1, obtaining spectral data of the fiber Bragg grating sensor;

[0059] S2. Calculate the center wavelength drift characteristics and spectrum shape characteristics based on the spectrum data.

[0060] S3. The center wavelength drift characteristics and the spectrum shape characteristics are integrated to obtain the weight data of the measured vehicle.

[0061] For details, please refer to Figure 2-Figure 4Fiber Bragg grating sensor is a sensor that uses fiber Bragg grating as a sensitive element. It can sense the change of external physical quantity and convert it into a measurable light signal change. Therefore, when the vehicle under test passes over the load platform, the fiber Bragg grating sensor senses the pressure change of the load platform and converts it into a measurable light signal. The light signal is converted into spectral data by using a photoelectric conversion device. After obtaining the spectral data, the computing device obtains the center wavelength drift characteristics and the spectral shape characteristics by analyzing and calculating the spectral data; the center wavelength drift characteristics and the spectral shape characteristics are input into the neural network model, and the weight data of the vehicle under test can be directly predicted by the neural network model. In this embodiment, the weight data of the vehicle under test can be directly obtained by analyzing and calculating the spectral data obtained by the fiber Bragg grating sensor. There is no need to add an additional compensation device to supplement the changes in ambient temperature and vehicle speed that will affect the detection accuracy of the system. The accurate weight data of the vehicle under test can be directly obtained, which not only simplifies the complexity of the system, but also effectively improves the accuracy of dynamic detection of vehicle weight.

[0062] For further information, please refer to Figure 5 In an embodiment of the present application, before the above step S2, it also includes: performing noise reduction processing on the spectral data through a noise reduction layer.

[0063] In the present embodiment, after the computing device acquires the spectral data through the fiber grating sensor, it is necessary to perform noise reduction processing on the spectral data after the computing device acquires the spectral data through the fiber grating sensor, considering that the spectral data may be affected by environmental noise, system noise and other noises during the process of acquiring the spectral data through the fiber grating sensor, so as to increase the signal-to-noise ratio and reliability of the spectral signal, thereby improving the accuracy of the subsequent data analysis structure.

[0064] For further information, please refer to Figure 5 In an embodiment of the present application, the denoising layer includes a convolutional layer, a batch normalization layer, a ReLU activation function, a plurality of residual shrinkage modules, a batch normalization layer, a ReLU function layer, a global average pooling layer and a fully connected layer which are arranged in sequence.

[0065] Specifically, in this embodiment, the spectral data acquired by the fiber Bragg grating sensors at different times and positions need to be subjected to noise reduction processing separately. Therefore, in this embodiment, the spectral data R acquired by the j-th fiber Bragg grating sensor at the i-th time is ti-FBGj For example, the spectral data R ti-FBGj After being input into the denoising layer, the spectral data will pass through the Conv convolution layer, the batch normalization BN layer, the ReLU activation function, and then through multiple residual shrinkage modules, and then through the batch normalization BN layer, the ReLU activation function, the global average pooling GAP, and finally through the fully connected layer FC output to obtain the spectral data R' after noiseti-FBGj .

[0066] In this embodiment, the residual shrinkage modules all adopt Figure 6 The network structure shown in Figure 1. First, the data is processed by 1×1Conv convolution layer, batch normalization BN layer, ReLU activation function, and then by 1×1Conv convolution layer, batch normalization layer BN, and activation function ReLU. Then, the eigenvalue is obtained by global average pooling layer GAP. At the same time, the features output by the global average pooling layer GAP are passed through the fully connected layer FC, the batch normalization layer BN, the activation function ReLU, the fully connected layer FC, the activation function ReLU, and the feature weight α ti-FBGj . The feature value output after global average pooling of the GAP layer With feature weight α ti-FBGj Multiply them together to get the weight threshold, which is used to perform soft thresholding of spectral data.

[0067] It should be noted that, in this embodiment, the weight threshold value can be adjusted dynamically according to the spectral data.

[0068] For further information, please refer to Fig. 9 In the embodiment of the present application, during the process of the tested vehicle passing through the load platform, the spectral data of the fiber grating sensor is collected multiple times to obtain the spectral data corresponding to each fiber grating sensor at different time moments, and then the spectral data of the fiber grating sensor at different positions at different times are analyzed, and then the corresponding sub-features are obtained, and then these sub-features are analyzed and calculated again, and finally the central wavelength drift feature and the spectral shape feature can be obtained. In this implementation, the analysis structure of the spectral data of the fiber grating sensor at different positions at different times is fused together through the data fusion method to form a calculation result under macroscopic conditions, and then the weight data of the tested vehicle can be calculated. In this case, not only the complexity of the system is simplified, but also the accuracy of dynamic detection of vehicle weight is effectively improved.

[0069] For further information, please refer to Figure 7 In the embodiment of the present application, the central wavelength drift characteristics of the fiber Bragg grating sensor calculated based on the spectral data include:

[0070] The sub-center wavelength drift characteristics are calculated respectively according to the spectral data corresponding to each fiber grating sensor at different time moments;

[0071] All sub-center wavelength drift features are passed through the fully connected layer to obtain the center wavelength drift feature.

[0072] Specifically, since the measured vehicle passes through the load platform, the use of data at different times and different positions can avoid the influence of a single data deviation. Therefore, in this embodiment, the spectral data obtained by the fiber grating sensors at different positions at different times are calculated separately. After obtaining the spectral data of all fiber grating sensors to calculate the sub-center wavelength drift characteristics, all sub-center wavelength drift characteristics are passed through the fully connected layer to obtain the center wavelength drift characteristics. The above data processing method takes into account the full process and all-round data, and can analyze the spectral data more comprehensively, which is conducive to improving the measurement accuracy of the vehicle weight. In this embodiment, the strategy of processing the spectral data of all fiber grating sensors separately and then fusing them through the fully connected layer can obtain more comprehensive center wavelength drift characteristics, effectively improving the generalization ability and accuracy of the model.

[0073] For further information, please refer to Figure 8 In the embodiment of the present application, the spectral data corresponding to each fiber grating sensor at different time moments are calculated through the central wavelength drift feature extraction layer to obtain the sub-central wavelength drift feature;

[0074] In this embodiment, the central wavelength drift feature extraction layer adopts a 1D-CNN network structure, including a convolution layer, a global pooling layer, a convolution layer, a global pooling layer, a Flatten layer, a ReLU function layer and a fully connected layer arranged in sequence.

[0075] Specifically, the spectral data R obtained by the j-th fiber Bragg grating sensor at the i-th time ti-FBGj For example, the spectral data R ti-FBGj After being input into the center wavelength drift feature extraction layer, the Conv convolution layer and the global pooling layer in the center wavelength drift feature extraction layer alternately perform spectral data R ti-FBGj After data processing, Flatten is used to expand the data and ReLU function is used to activate it. Finally, the fully connected layer FC is used to extract the sub-center wavelength drift feature A of the j-th FBG at the i-th moment. ti-FBGj .

[0076] After completing the analysis and calculation of the spectral data acquired by one group of the fiber Bragg grating sensors, repeat the above method to obtain the sub-center wavelength drift characteristics A of all fiber Bragg grating sensors at all times. t1-FBGj , A t2-FBGj , ... A tM-FBGj .

[0077] Finally, the sub-center wavelength drift characteristics of all fiber grating sensors at all times mentioned above are input into the fully connected layer to obtain the center wavelength drift characteristics.

[0078] Furthermore, in an embodiment of the present application, in the process of calculating the spectral shape characteristics based on the spectral data, the spectral data are screened in chronological order to obtain spectral data sets corresponding to the fiber grating sensor at different times; the spectral data sets at different times are input into the spectral shape feature extraction layer to obtain the position characteristics corresponding to the different times; and then the position characteristics at different times are input into the spectral shape feature extraction layer to obtain the spectral shape characteristics.

[0079] In this embodiment, the corresponding time features and space features are extracted from the spectral data at the time level and the space level respectively, and the spectral shape features with time features and position features are obtained and input into the neural network model, so as to analyze the spatiotemporal features of the spectral data and then calculate the dynamic weighing data of the tested vehicle. In this process, the dynamic influence of the vehicle speed on the spectral data is taken into account. Therefore, in this embodiment, the weighing result corrected based on the vehicle speed can be directly obtained without the need for an additional vehicle speed sensor, thereby further simplifying the weighing system.

[0080] In this embodiment, the spectral shape feature extraction layer includes a forward LSTM structure layer, a reverse LSTM structure layer, a Concat layer, a forward LSTM structure layer, a reverse LSTM structure layer, a Concat layer and a fully connected layer which are arranged in sequence.

[0081] Specifically, the spectral data are screened in time order to obtain the spectral data sets corresponding to the fiber Bragg grating sensor at different times. Taking any time ti as an example, the spectral data set corresponding to the fiber Bragg grating sensor at time ti is {R ti-FBG1 , R ti-FBG2 ,……,R ti-FBGn}, the above spectral data set {R ti-FBG1 , R ti-FBG2 ,……,R ti-FBGn} They are sequentially passed through the forward LSTM structure and the reverse LSTM structure, and then the outputs of the two LSTM structures are spliced, and then the output results are passed through the forward LSTM structure and the reverse LSTM structure again, and the outputs of the two LSTM structures are spliced ​​twice to obtain the output results. Finally, the output results of all spectral data in the spectral dataset are passed through the fully connected layer FC to output the position feature R at time i ti .

[0082] Repeat the above steps to obtain the position features R corresponding to different times t1 , R t2 , R t3 , …, R ti .

[0083] The position features R corresponding to all the above moments t1 , R t2, R t3 , …, R ti They pass through the forward LSTM structure and the reverse LSTM structure respectively, and then the outputs of the two LSTM structures are spliced. Then the output results pass through the forward LSTM structure and the reverse LSTM structure again, and the outputs of the two LSTM structures are spliced ​​again to obtain the output results. Finally, the output results of all spectral data in the spectral dataset are output through the fully connected layer FC to obtain the spectral shape characteristics.

[0084] In some embodiments, the center wavelength drift feature and the spectrum shape feature are fused through a weighing model to obtain weight data;

[0085] Among them, the weighing model training method includes:

[0086] Acquire a spectral data set, the spectral data set including spectral data and actual weight data, actual vehicle speed, and actual ambient temperature corresponding to the spectral data;

[0087] Inputting the spectral data into the neural network model so that the neural network model outputs predicted weight, predicted vehicle speed and predicted ambient temperature;

[0088] Calculate the loss value based on the predicted weight, the predicted vehicle speed, the predicted ambient temperature, and the corresponding actual weight data, the actual vehicle speed, and the actual ambient temperature;

[0089] According to the loss value, the weight and threshold of each node in the neural network model are adjusted until the loss value reaches the minimum;

[0090] Different spectral data are input into the neural network model for iterative training until the predicted weight, predicted vehicle speed, predicted ambient temperature output by the neural network model for multiple times in succession, and the loss values ​​calculated from the corresponding actual weight data, actual vehicle speed and actual ambient temperature are all less than the preset loss value threshold.

[0091] In the above weighing model training method, the loss value is calculated by the loss function Cost, and the loss function Cost is as follows:

[0092]

[0093] Among them, W i and W i′ are the actual and predicted weight of the vehicle, V i and V i′ are the actual and predicted speeds of the vehicle, T i and T i′ are the actual and predicted ambient temperatures respectively. a, b, c are weighting coefficients, which are determined by experience according to the actual situation.

[0094] Taking into account that different types of vehicles have different body lengths and the time it takes to pass through the weighing platform is also different, in order to improve the accuracy and efficiency of the model, it is necessary to train weighing models for different types of vehicles separately. In the training process of the weighing model, the spectral data set also includes the vehicle type. The spectral data set is screened according to the vehicle type to obtain the spectral data sets corresponding to different vehicle types, and the weighing models of different vehicle types are generated based on the above spectral data sets.

[0095] In actual use, while obtaining the spectral data of the fiber grating sensor, it is also necessary to identify the vehicle type of the vehicle being tested, and obtain a weighing model that matches the vehicle type based on the vehicle type, and then calculate the weight data of the vehicle being tested through the weighing model that matches the vehicle type.

[0096] The embodiment of the present invention further provides a weighing system based on a fiber grating sensor, comprising:

[0097] A weighing device, comprising a load-bearing platform and a plurality of fiber grating sensors spaced apart along the length direction of the load-bearing platform;

[0098] Acquisition model, acquisition module for spectral data of fiber Bragg grating sensor;

[0099] The processing module is used to process the spectral data acquired by the acquisition model according to any of the above-mentioned weighing methods based on fiber grating sensors to obtain weight data of the vehicle under test.

[0100] It should be noted that the present application is not limited to the above-mentioned embodiments. The above-mentioned embodiments are only examples, and the embodiments having the same structure as the technical idea and exerting the same effect within the scope of the technical solution of the present application are all included in the technical scope of the present application. In addition, without departing from the scope of the main purpose of the present application, various modifications that can be thought of by those skilled in the art to the embodiments and other methods of combining some of the constituent elements in the embodiments are also included in the scope of the present application.

Claims

1. A weighing method based on fiber grating sensor, characterized in that: The weighing method is applied to a weighing device, the weighing device comprises a carrying platform and a plurality of fiber grating sensors arranged at intervals along the length direction of the carrying platform, and the weighing method comprises: Acquiring spectral data of the fiber grating sensor; Calculate the center wavelength drift characteristics and the spectrum shape characteristics according to the spectrum data; The center wavelength drift feature and the spectrum shape feature are combined to obtain the weight data of the tested vehicle.

2. The weighing method based on fiber grating sensor according to claim 1 is characterized in that: Before calculating the central wavelength drift characteristics and the spectrum shape characteristics of the fiber Bragg grating sensor according to the spectrum data, the method further includes performing noise reduction processing on the spectrum data through a noise reduction layer.

3. The weighing method based on fiber grating sensor according to claim 2 is characterized in that: The denoising layer includes a convolutional layer, a batch normalization layer, a ReLU activation function, a plurality of residual shrinkage modules, a batch normalization layer, a ReLU function layer, a global average pooling layer and a fully connected layer which are arranged in sequence.

4. The weighing method based on fiber grating sensor according to claim 1, characterized in that: The step of acquiring the spectral data of the fiber grating sensor comprises: When the vehicle under test passes through the supporting platform, the spectrum data of the fiber grating sensor is collected multiple times to obtain the spectrum data corresponding to each fiber grating sensor at different time moments.

5. The weighing method based on fiber grating sensor according to claim 4 is characterized in that: The calculating of the center wavelength drift characteristics of the fiber Bragg grating sensor according to the spectral data comprises: According to the spectral data corresponding to each of the fiber grating sensors at different time moments, the sub-center wavelength drift characteristics are calculated respectively; All the sub-center wavelength drift characteristics are passed through a fully connected layer to obtain the center wavelength drift characteristic.

6. The weighing method based on fiber grating sensor according to claim 5, characterized in that: The spectrum data corresponding to each fiber grating sensor at different time moments is calculated through a center wavelength drift feature extraction layer to obtain a sub-center wavelength drift feature; The central wavelength drift feature extraction layer includes a convolution layer, a global pooling layer, a convolution layer, a global pooling layer, a Flatten layer, a ReLU function layer and a fully connected layer which are arranged in sequence.

7. The weighing method based on fiber grating sensor according to claim 4, characterized in that: The spectrum shape feature calculated according to the spectrum data comprises: Screening the spectral data in a time sequence to obtain spectral data sets corresponding to the fiber Bragg grating sensor at different times; Inputting the spectral data sets at different times into the spectral shape feature extraction layer to obtain the position features corresponding to the different times; Inputting the position features at different times into a spectral shape feature extraction layer to obtain spectral shape features; Among them, the spectral shape feature extraction layer includes a forward LSTM structure layer, a reverse LSTM structure layer, a Concat layer, a forward LSTM structure layer, a reverse LSTM structure layer, a Concat layer and a fully connected layer which are arranged in sequence.

8. The weighing method based on fiber grating sensor according to claim 1, characterized in that: The weight data is obtained by fusing the central wavelength drift feature and the spectrum shape feature through a weighing model; Wherein, the weighing model training method comprises: Acquiring a spectral data set, the spectral data set comprising spectral data and actual weight data, actual vehicle speed, and actual ambient temperature corresponding to the spectral data; Inputting the spectral data into a neural network model so that the neural network model outputs a predicted weight, a predicted vehicle speed, and a predicted ambient temperature; Calculating a loss value based on the predicted weight, predicted vehicle speed, predicted ambient temperature, and corresponding actual weight data, actual vehicle speed, and actual ambient temperature; Adjusting the weight and threshold of each node in the neural network model according to the loss value until the loss value reaches a minimum; Different spectral data are input into the neural network model for iterative training until the predicted weight, predicted vehicle speed, predicted ambient temperature output by the neural network model for multiple times in succession, and the loss values ​​calculated from the corresponding actual weight data, actual vehicle speed and actual ambient temperature are all less than the preset loss value threshold.

9. The weighing method based on fiber grating sensor according to claim 8, characterized in that: The spectral data set also includes vehicle types, the spectral data is screened according to the vehicle types, and the weighing models of different vehicle types are generated; When acquiring the spectrum data of the fiber grating sensor, it is also necessary to identify the vehicle type of the vehicle being tested, and acquire a weighing model matching the vehicle type according to the vehicle type.

10. A weighing system based on fiber grating sensor, characterized in that: include: A weighing device, the weighing device comprising a carrying platform and a plurality of fiber grating sensors spaced apart along the length direction of the carrying platform; An acquisition model, wherein the acquisition module is used for the spectral data of the fiber Bragg grating sensor; A processing module is used to process the spectral data acquired by the acquisition model according to the weighing method based on the fiber grating sensor according to any one of claims 1 to 9 to obtain weight data of the vehicle under test.

Images