A light wave measurement device and method based on multimode optical waveguide

Abstract

The present invention discloses an optical wave measurement device and method based on a multimode optical waveguide, the device comprising: an optical access port; a temperature modulation module, used for performing temperature modulation on the multimode optical waveguide; a multimode optical waveguide, wherein the first end of the multimode optical waveguide is coupled with the optical access port, receives the optical wave to be measured and forms a speckle at the second end of the multimode optical waveguide; a one-to-many optical waveguide structure, used for coupling the speckle emitted by the multimode optical waveguide into a multi-channel dielectric optical waveguide; the multi-channel dielectric optical waveguide is used for coupling the speckle into a photodetector array; and the photodetector array is used for analyzing the speckle to determine at least one attribute of the optical wave to be measured and / or determine the change of at least one attribute of the optical wave to be measured.

Description

Technical Field

[0001] The present invention relates to the field of optical fiber sensing technology, and in particular to an optical wave measuring device and method based on a multimode optical waveguide. Background Art

[0002] In recent years, the discovery and application of laser speckle effect has injected new research ideas into the field of fiber optic sensing. Laser speckle effect is a very useful physical effect, which is widely used in various fields such as non-destructive detection, speckle imaging, and spectral measurement. Using the speckle effect of coherent light, various physical quantities superimposed on the coherent light can be measured.

[0003] At present, in the field of fiber optic sensing, the light used is mostly near-infrared light. Currently, the only optoelectronic material with the best detection efficiency in the near-infrared band is InGaAs, but InGaAs area array or linear array CCDs are extremely expensive, ranging from tens of thousands to hundreds of thousands, which makes the price of near-infrared fiber optic sensor devices using speckle high. Using a photodetector array as a speckle detector will compress the original rich information of the speckle into the same number of features of the photodetector array used, resulting in a large loss of information.

[0004] In order to solve the above technical problems, the present application provides an optical wave measurement device and method based on a multimode optical waveguide. Summary of the invention

[0005] The embodiments of the present application provide a light wave measurement device and method based on a multimode optical waveguide, which solves the technical problem in the prior art that a lot of information is lost when using a photodetector array for speckle detection, and achieves the technical effect of reducing information loss and obtaining rich data information when using a radio and television detector to detect speckle.

[0006] The embodiment of the present application provides a light wave measurement device based on a multimode optical waveguide, the device comprising:

[0007] an optical access port including a single-mode optical waveguide;

[0008] A temperature modulation module, used for performing temperature modulation on a multimode optical waveguide;

[0009] A multimode optical waveguide, wherein a first end of the multimode optical waveguide is coupled with a single-mode optical waveguide of the optical access port, receives the optical wave to be measured and forms a speckle at a second end of the multimode optical waveguide;

[0010] A one-to-many optical waveguide structure is used to couple the speckle emitted by the multimode optical waveguide into a multi-channel dielectric optical waveguide;

[0011] a multi-channel dielectric optical waveguide for coupling the speckle into a photodetector array;

[0012] A photodetector array is used to analyze the speckle to determine at least one property of the light wave to be measured and / or to determine a change in at least one property of the light wave to be measured.

[0013] Furthermore, the temperature modulation module includes:

[0014] A heating unit, used for heating the multimode optical waveguide;

[0015] The temperature detection unit is used to detect the temperature change of the multimode optical waveguide in real time.

[0016] Furthermore, the device also includes a data processing module.

[0017] The photoelectric detector analyzes speckles formed by the same light wave to be measured at different temperatures, and the data processing module determines at least one attribute of the light wave to be measured and / or determines a change in at least one attribute of the light wave to be measured based on the speckles formed at each temperature.

[0018] Furthermore, the data processing module includes a deep learning network model.

[0019] The photodetector array analyzes the speckles at M different temperatures to obtain a one-dimensional array with a length of N corresponding to each temperature;

[0020] The data processing module normalizes the one-dimensional array with a length of N, converts M normalized one-dimensional arrays with a length of N into two-dimensional arrays of M*N, and converts the two-dimensional array into a grayscale image;

[0021] The data processing module inputs the grayscale image into a deep learning network model to determine at least one attribute of the light wave to be measured and / or determine a change in at least one attribute of the light wave to be measured.

[0022] Furthermore, the data processing module is also used to train the deep learning network model.

[0023] The multimode optical waveguide receives Q optical waves to be measured with different central wavelengths, and sends P pulses to the multimode optical waveguide at each central wavelength;

[0024] Temperature modulation is performed on the multimode optical waveguide, a photodetector array analyzes the speckles at M different temperatures, and the data processing module receives speckle information with a data volume of P*Q*M*N;

[0025] The data processing module performs normalization processing on a one-dimensional array of length N detected by the photoelectric detector array each time;

[0026] For each pulse of the light wave to be measured of each wavelength, the data processing module converts M normalized one-dimensional arrays of length N into M*N two-dimensional arrays, and converts the two-dimensional arrays into grayscale images to obtain P*Q grayscale images;

[0027] The data processing module uses P*Q grayscale images as training sets, selects at least part of them as training data sets, and the rest as verification training sets, inputs the training sets into the deep learning network model, and trains the deep learning network model.

[0028] Further, the one-to-many optical waveguide structure includes a first tapered waveguide and a plurality of second tapered waveguides, and the number of the second tapered waveguides corresponds to the number of photodetectors in the photodetector array;

[0029] The first end face of the first tapered waveguide is coupled to the second end of the multimode optical waveguide, the second end face of the first tapered waveguide is coupled to the first end face of any second tapered waveguide, the second end face of any second tapered waveguide is coupled to the first end of a dielectric optical waveguide, the first end face area of ​​the first tapered waveguide is smaller than the second end face area of ​​the first tapered waveguide, and the first end face area of ​​the second tapered waveguide is larger than the second end face area of ​​the second tapered waveguide;

[0030] Wherein, when the areas of the first end surfaces of the second tapered waveguides are substantially equal, the second tapered waveguides are unevenly distributed on the second end surface of the first tapered waveguide;

[0031] Alternatively, the plurality of second tapered waveguides included in the one-to-many optical waveguide structure include at least two types of second tapered waveguides, and each type of second tapered waveguide has a different first end surface area.

[0032] Furthermore, the attributes include one or more of the following:

[0033] The wavelength of the light wave to be measured, the polarization of the light and the spatial mode.

[0034] The present application also provides a method for measuring optical waves based on a multimode optical waveguide, the method comprising the following steps:

[0035] performing temperature modulation on the multimode optical waveguide;

[0036] The light wave to be measured forms speckles through the multimode optical waveguide, and the speckles at M different temperatures are analyzed by a photodetector array to obtain a one-dimensional array with a length of N corresponding to each temperature;

[0037] Normalizing the one-dimensional array with a length of N;

[0038] M normalized one-dimensional arrays of length N are converted into M*N two-dimensional arrays, and the two-dimensional arrays are converted into grayscale images, and the grayscale images are input into a deep learning network model to determine at least one attribute of the light wave to be measured and / or determine the change of at least one attribute of the light wave to be measured.

[0039] Furthermore, the method also includes the following training steps of the deep learning network model:

[0040] Inputting Q optical waves to be measured with different central wavelengths, and sending P pulses to the multimode optical waveguide at each central wavelength;

[0041] Temperature modulation is performed on the multimode optical waveguide, and a photodetector array analyzes the speckles at M different temperatures to obtain speckle information with a data volume of P*Q*M*N;

[0042] Normalizing the one-dimensional array with a length of N detected by the photodetector array each time;

[0043] For each pulse of the light wave to be measured of each wavelength, M normalized one-dimensional arrays of length N are converted into M*N two-dimensional arrays, and the two-dimensional arrays are converted into grayscale images to obtain P*Q grayscale images;

[0044] Take P*Q grayscale images as training sets, select at least part of them as training data sets, and the rest as verification training sets, input the training sets into the deep learning network model, and train the deep learning network model.

[0045] Furthermore, the deep learning network model includes:

[0046] A convolutional layer, used to extract features from the grayscale image and output a feature vector;

[0047] A pooling layer, used for performing a pooling operation on the feature vector output by the convolutional layer to reduce the dimension of the feature vector;

[0048] A fully connected layer, connected to the pooling layer, for integrating the features output by the pooling layer;

[0049] The output of the fully connected layer is classified using a softmax function to determine at least one attribute of the light wave to be measured and / or determine a change in at least one attribute of the light wave to be measured.

[0050] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0051] 1. Since a photoelectric detector array is used for speckle detection, the technical problem of expensive photoelectric materials in the prior art is effectively solved, thereby reducing the expensive cost of photoelectric materials.

[0052] 2. Since a heater is used to heat the multimode optical waveguide and to record the information of the speckle at different temperatures in real time, the technical problem of losing a lot of two-dimensional image information when using a photoelectric detection array as a speckle detector is effectively solved. The change of the speckle in the temperature dimension is used to greatly increase the acquired data information.

[0053] 3. Since semiconductor silicon photonic chip technology is used to integrate multimode optical waveguides, temperature modulation modules and photodetector arrays on the optical chip, the speckle is directly coupled to the photosensitive surface of the photodetector array on the chip, avoiding spatial coupling. At the same time, the multimode optical waveguide is used as a scattering medium, which increases the volume and stability of the scattering medium. The integration of temperature modulation realizes multiple changes of the speckle at the end of the multimode optical waveguide with temperature modulation under the same incident light, enriching the amount of speckle information obtained by the photodetector array and improving the resolution and accuracy of demodulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 A schematic diagram of the structure of a light wave measurement device based on a multimode optical waveguide provided in an embodiment of the present application;

[0055] Figure 2 This is a schematic diagram of the structure of the temperature detection module in the embodiment of the present application;

[0056] Figure 3 This is a schematic diagram of the structure of the internal optical path of the optical chip in the embodiment of the present application;

[0057] Figure 4 A schematic diagram of the hierarchical structure inside the optical chip in the embodiment of the present application;

[0058] Figure 5 This is a schematic diagram of a one-to-many optical waveguide structure in an embodiment of the present application;

[0059] Figure 6 A schematic diagram of a heating pulse signal applied to a heater in an optical chip in an embodiment of the present application;

[0060] Figure 7 This is a schematic diagram of a temperature change curve collected by a temperature detector in an optical chip when a heater obtains a heating pulse in an embodiment of the present application;

[0061] Figure 8 This is a schematic diagram of an M×N two-dimensional array collected by N photodetectors at M different temperatures in an embodiment of the present application;

[0062] Fig. 9A flow chart of a light wave measurement method based on a multimode optical waveguide provided in an embodiment of the present application. DETAILED DESCRIPTION

[0063] In order to make those skilled in the art better understand the scheme of the present invention, the technical scheme in the specific implementation mode of the present application will be clearly and completely described below in conjunction with the drawings in the implementation mode of the present application. It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other in the absence of conflict. The present application will be described in detail below in conjunction with the embodiments.

[0064] like Figure 1 As shown, an embodiment of the present application provides an optical wave measurement device 100 based on a multimode optical waveguide, comprising: an optical access port 11, a multimode optical waveguide 12, a temperature modulation module 13, a one-to-many optical waveguide structure 14, a multi-channel dielectric optical waveguide 15, and a photodetector array 16.

[0065] The optical access port 11 is where the light wave to be measured enters. The optical access port may include an optical fiber 111 portion extending outside the device and a single-mode optical waveguide 112 portion inside the device.

[0066] The multimode optical waveguide 12 has a first end coupled with the single-mode optical waveguide 112 of the optical access port 11 , receives the optical wave to be measured and forms a speckle at the second end of the multimode optical waveguide 12 .

[0067] The temperature modulation module 13 is used to perform temperature modulation on the multimode optical waveguide 12 .

[0068] The one-to-many optical waveguide structure 14 is used to couple the speckle emitted by the multimode optical waveguide 12 into the multi-path dielectric optical waveguide 15 .

[0069] The multi-channel dielectric optical waveguide 15 is used to couple the speckle into the photodetector array 16.

[0070] The photodetector array 16 is used to analyze the speckle to determine at least one property of the light wave to be measured and / or to determine the change of at least one property of the light wave to be measured.

[0071] As an optional implementation method, the light to be measured enters the multimode optical waveguide 12 and the temperature modulation module 13 through the optical access port 11, and forms an interference speckle at the end of the multimode optical waveguide 12, and the temperature modulation module 13 can perform temperature modulation on the multimode optical waveguide 12 to change the output speckle.

[0072] The one-to-many waveguide structure 14 couples the light intensity at different positions of the speckle pattern into the multi-path dielectric optical waveguide 15, and the light is directly coupled into the photosensitive surface of the photodetector at the end of the dielectric optical waveguide; finally, the signal of the photodetector array 16 is collected by the signal collection system, and the information of the light to be measured is demodulated after processing.

[0073] The device provided by the present application integrates a single-mode optical waveguide 112 and a multi-mode optical waveguide 12 on an optical chip, and introduces the optical wave to be measured into the device through the single-mode optical waveguide 112, thereby avoiding the influence of external vibration on the speckle distribution. Among them, the single-mode optical waveguide 112 can use a single-mode optical fiber with an FC / APC interface, and the end of the single-mode optical fiber and the multi-mode optical waveguide 12 use end-face coupling to couple light into the multi-mode optical waveguide 12.

[0074] As an optional implementation, in the multi-channel dielectric optical waveguide 15, any channel of the dielectric optical waveguide can be single-mode or multi-mode, which is not limited here.

[0075] like Figure 2 As shown, as an optional implementation, the temperature modulation module 13 includes: a heating unit 131, a temperature detection unit 132.

[0076] The heating unit 131 is used to heat the multimode optical waveguide 12 .

[0077] The temperature detection unit 132 is used to detect the temperature change of the multi-mode optical waveguide 12 in real time.

[0078] Combination Figure 1 and Figure 2 When the light to be measured enters the multimode optical waveguide 12 from the optical access port 11, the temperature modulation module 13 performs temperature modulation on the multimode optical waveguide 12, that is, the heating unit 131 heats the multimode optical waveguide 12 to change the temperature of the multimode optical waveguide 12, thereby changing the speckle information output by the multimode optical waveguide 12. The temperature detection unit 132 can detect the temperature of the multimode optical waveguide 12 in real time, detect the real-time temperature of the multimode optical waveguide 12 after being heated by the heating unit 131, and analyze the speckle information obtained at the temperature through the photodetector array 16.

[0079] As an optional implementation method, in an embodiment of the present application, the optical access port 11, the multi-mode optical waveguide 12, the temperature modulation module 13, the one-to-many waveguide structure 14, the multi-channel dielectric optical waveguide 15 and the photodetector array 16 can all be integrated on an optical chip using semiconductor silicon photonic chip technology.

[0080] As an optional implementation, the optical chip may be an optical waveguide chip using a PLC process. Figure 3As shown, the light to be measured enters the multimode optical waveguide 12 and the temperature modulation module 13 through the access port 11, the temperature modulation module 13 is not shown in the figure, and forms a speckle at the end of the multimode optical waveguide 12, and the temperature modulation module can perform temperature modulation on the multimode optical waveguide, so that the output speckle changes; the one-to-many waveguide structure 14 couples the light intensity at different positions of the speckle pattern into the multi-path dielectric optical waveguide 15, and the light is directly coupled into the photosensitive surface of the photodetector at the end of the dielectric optical waveguide; finally, the signal of the photodetector array 16 is collected by the signal collection system, and the information of the light to be measured is demodulated after processing.

[0081] The multimode optical waveguide 12 may adopt a double helix structure but is not limited to it, so as to increase the length of the waveguide in a small area, and the core layer size of the waveguide must meet the standard of multimode transmission. Under the same conditions, the larger the core layer size, the more modes it can support, and the number of scattering patterns at the end of the multimode optical waveguide 12 will also increase, thereby improving the resolution of the device.

[0082] The temperature modulation module 13 can perform temperature modulation on the multimode optical waveguide 12, change the temperature of the multimode optical waveguide 12, so that the output speckle information changes with the change of the temperature of the multimode optical waveguide 12, thereby effectively increasing the amount of data obtained. Analyzing the speckle information obtained under different temperature conditions can effectively improve the accuracy of the analysis.

[0083] As an optional implementation, the heating unit 131 can heat the multimode optical waveguide 12 to a certain temperature, record the output speckle information, and then stop heating the multimode optical waveguide 12. Instead, the temperature of the multimode optical waveguide 12 naturally decreases or the temperature decrease rate of the multimode optical waveguide 12 is actively intervened, and the output speckle information when the temperature drops to different temperatures is recorded in real time, so that rich speckle information data can be recorded.

[0084] As an optional implementation, when the temperature of the multimode optical waveguide 12 is modulated, the multimode optical waveguide 12 may be heated by the heating unit 131, so that the temperature of the multimode optical waveguide 12 is increased multiple times, and the speckle information outputted during the temperature increase of the multimode optical waveguide 12 is recorded. When recording the outputted speckle information, the outputted speckle information may be recorded once at a same interval, or the outputted speckle information may be recorded in real time.

[0085] As an optional implementation, the temperature detection unit 132 may be a thermal sensor detector. Figure 4As shown, it shows a schematic diagram of the hierarchical structure of the optical chip. The optical chip includes a temperature detection layer 401, a waveguide layer 402 and a substrate 403. Among them, the waveguide layer 402 is located above the substrate 403, and the temperature detection layer 401 is located above the waveguide layer 402. The multimode optical waveguide 12 is in the waveguide layer 402, and the thermistor detector is in the temperature detection layer 401, which is closely connected to the waveguide layer 402, so as to quickly sense the temperature change of the multimode optical waveguide 12. The heating unit 131 is placed in the waveguide layer 402 and is closely connected to the multimode optical waveguide 12, so that the heat emitted by the heating unit 131 can be quickly conducted to the multimode optical waveguide 12.

[0086] As an optional implementation, the first end of the one-to-many optical waveguide structure 14 is coupled with the second end of the multimode optical waveguide 12 to receive the speckle, and the distribution of the speckle is amplified at the second end of the one-to-many optical waveguide structure 14, the first end of each dielectric optical waveguide in the multi-channel dielectric optical waveguide 15 is connected to the second end of the one-to-many optical waveguide structure 14, and the second end of each dielectric optical waveguide is respectively connected to the photodetector in the light spot detector array 16;

[0087] As an optional implementation, Figure 5 As shown, the one-to-many optical waveguide structure 14 may include an approximately tapered waveguide. The area of ​​the first end face thereof is smaller than the area of ​​the second end face, so that the speckle distribution can be amplified. The multi-channel dielectric optical waveguide 15 is connected to the second end of the one-to-many optical waveguide structure 14. For the sake of convenience, the present application refers to the approximately tapered waveguide for amplifying the speckle distribution as the first tapered waveguide 141.

[0088] like Figure 5 As shown, as an optional implementation, the one-to-many optical waveguide structure also includes a plurality of smaller approximately conical waveguides, and each of the smaller approximately conical waveguides connects a dielectric optical waveguide 152 to the second end face of the first conical waveguide. For the sake of convenience, the application refers to the smaller approximately conical waveguide as the second conical waveguide.

[0089] For each second tapered waveguide 142 , the first end face of the second tapered waveguide 142 is connected to the second end face of the first tapered waveguide 141 , the second end face of the second tapered waveguide 142 is connected to a dielectric optical waveguide 151 , and the first end face area of ​​the second tapered waveguide 142 is larger than the second end face area.

[0090] As an optional implementation, the number of the second tapered waveguides 142 corresponds to the number of photodetectors in the photodetector array.

[0091] As an optional implementation, the plurality of second tapered waveguides 142 included in the one-to-many optical waveguide structure 14 include at least two types of second tapered waveguides 142 , and each type of second tapered waveguide 142 has a different first end surface area.

[0092] For each second tapered waveguide 142 having a different first end surface area, the arrangement position of the second tapered waveguide 142 on the second end surface of the first tapered waveguide 141 can be set according to actual needs.

[0093] As an optional implementation, the area of ​​the first end face of the second tapered waveguide 142 located at the center of the first tapered waveguide 141 may be smaller than the area of ​​the first end face of the second tapered waveguide 142 located at the edge of the first tapered waveguide 141. That is, the thicker second tapered waveguide 142 is connected to the position where the second end face of the first tapered waveguide 141 approaches the edge, and the thinner second tapered waveguide 142 is connected to the position where the second end face of the first tapered waveguide 141 approaches the center. Thus, on the second end face of the one-to-many optical waveguide structure 14, the connection area between the one-to-many optical waveguide structure 14 and the multi-channel dielectric optical waveguide 15 increases along the center of the one-to-many optical waveguide structure 14 to the edge direction. Thus, the multi-channel dielectric optical waveguide 15 can receive more speckle information located at the edge of the first tapered waveguide 141.

[0094] As another optional implementation, the second tapered waveguide 142 with a larger first end face area may be arranged at a position close to the center of the second end face of the first tapered waveguide 141, and the second tapered waveguide 142 with a smaller first end face area may be arranged at a position close to the edge of the second end face of the first tapered waveguide 141, so as to better receive the speckle information located at the center of the first tapered waveguide 141.

[0095] In another optional implementation, the areas of the first end surfaces of the second tapered waveguides 142 are substantially equal. In this embodiment, the second tapered waveguides 142 are unevenly distributed on the second end surface of the first tapered waveguide 141 .

[0096] For example, on the second end face of the first tapered waveguide 141, the distribution density of the second tapered waveguide 142 close to the edge of the second end face is greater than the distribution density of the second tapered waveguide 142 close to the center of the second end face. That is, when the one-to-many optical waveguide structure 14 is connected to the multi-channel dielectric optical waveguide 15, a smaller number of dielectric optical waveguides 151 are connected close to the center of the second end face, and a larger number of dielectric optical waveguides 151 are connected close to the edge of the second end face. This design can also enable the multi-channel dielectric optical waveguide 15 to receive more speckle information located at the edge of the first tapered waveguide 141.

[0097] Alternatively, on the second end face of the first tapered waveguide 141, the distribution density of the second tapered waveguide 142 close to the edge of the second end face is less than the distribution density of the second tapered waveguide 142 close to the center of the second end face. This design enables the multi-channel dielectric optical waveguide 15 to receive more speckle information located at the center of the first tapered waveguide 141.

[0098] In the embodiment of the present application, the second tapered waveguide 142 may also be randomly distributed on the second end surface of the first tapered waveguide 141. How the second tapered waveguide 142 is arranged on the first end surface of the first tapered waveguide 141 needs to be set according to actual needs.

[0099] For example, in some cases, the speckle resolution is determined by the difference formula It can be seen that when the waveguide length L is fixed, its resolution is determined by the high-order mode propagation constant β m The higher order mode is determined by the propagation constant β1, and the higher order mode tends to exist at the edge of the waveguide. Therefore, the speckle resolution at the edge of the multimode optical waveguide 12 is better, and more speckle information at the edge of the multimode optical waveguide 12 is more conducive to demodulation. Therefore, in this case, the second tapered waveguide 142 is arranged for the purpose of receiving as much speckle information as possible at the edge of the first tapered waveguide 141 (for example, the second tapered waveguide 142 is more densely distributed at the edge of the first tapered waveguide 141, or the first end surface area of ​​the second tapered waveguide 142 located at the edge is larger than the first end surface area of ​​the second tapered waveguide 142 located at the center).

[0100] As an optional implementation, the number of photodetectors in the multi-channel dielectric optical waveguide 15 and the photodetector array 16 should be consistent to ensure that the light coupled into the multi-channel dielectric optical waveguide 15 by the multi-mode optical waveguide 12 through the one-to-many optical waveguide structure 14 can be finally coupled to the photosensitive surface of the photodetector. The number of photodetectors can be designed according to actual needs, and can be greater than or equal to 2, preferably 16.

[0101] Different from the conventional method of using an InGaAs area array or linear array CCD as a speckle collection device, in the embodiment of the present application, a photodetector array 16 is used as a speckle collection device, which can effectively reduce the cost.

[0102] In addition, the present application uses the photodetector array 16 as a speckle collection device, which can improve the collection speed of the speckle information by the collection device. At the same time, in the embodiment of the present application, the temperature modulation module 13 is used to actively perform temperature modulation on the multimode optical waveguide 12, so that the photodetector array 16 can obtain speckle information under different temperature conditions, which greatly enriches the amount of speckle information obtained by the photodetector array 16, thereby solving the problem of losing a lot of information due to the limitation of the number of photodetectors in the photodetector array 16 when compressing the original speckle-rich information on the photodetector array 16 used.

[0103] As an optional implementation, the device 100 also includes a data processing module.

[0104] The photoelectric detector analyzes the speckles formed by the same light wave to be measured at different temperatures, and the data processing module determines at least one attribute of the light wave to be measured and / or determines the change of at least one attribute of the light wave to be measured according to the speckles formed at each temperature. The heater heats the multimode optical waveguide, thereby changing the temperature of the multimode optical waveguide, and can obtain the speckle information formed by the same light wave to be measured on the multimode optical waveguide at different temperatures. The photoelectric detector can obtain at least one attribute of the light wave to be measured and / or determine the change of at least one attribute of the light wave to be measured according to the speckle information at different temperatures.

[0105] As an optional implementation, the properties of the light wave include one or more of the following:

[0106] The wavelength of the light wave to be measured, the polarization of the light, and the spatial mode. The grayscale image is input into the deep learning network and after training, the wavelength of the light wave and / or the wavelength change can be obtained. According to the detected wavelength of the light wave to be measured and / or the wavelength change, the polarization and spatial mode of the light wave to be measured and / or the change of polarization and spatial mode can be obtained.

[0107] As an optional implementation, the data processing module includes a deep learning network model. The data processing module first trains the deep learning network model, and uses the trained deep learning network model to parse the speckle information to determine at least one attribute of the light wave to be measured and / or determine a change in at least one attribute of the light wave to be measured.

[0108] As an optional implementation method, the data processing module trains the deep learning network model including the following process:

[0109] The multimode optical waveguide receives Q optical waves to be measured with different central wavelengths, and sends P pulses to the multimode optical waveguide at each central wavelength.

[0110] The multimode optical waveguide is temperature modulated, the photodetector array analyzes the speckles at M different temperatures, and the data processing module receives the speckle information with a data volume of P*Q*M*N.

[0111] The data processing module normalizes the one-dimensional array with a length of N detected by the photoelectric detector array each time.

[0112] For each pulse of the light wave to be measured of each wavelength, the data processing module converts M normalized one-dimensional arrays of length N into M*N two-dimensional arrays, and converts the two-dimensional arrays into grayscale images to obtain P*Q grayscale images.

[0113] Specifically, Figure 6 As shown, it shows a schematic diagram of the heating pulse signal applied to the heater in the optical chip provided by the present application. First, the FBG (Fiber Bragg Grating) is controlled to reflect and output Q different central wavelengths of light, namely λ1, λ2, ...λ Q , input light into an optical wave measuring device based on the multimode optical waveguide 12, and input an identical pulse to the heating unit 131 in the device at each wavelength to perform temperature modulation on the multimode optical waveguide 12, so that the temperature of the multimode optical waveguide changes rapidly.

[0114] like Figure 7 As shown, it shows a schematic diagram of the temperature change curve collected by the temperature detector in the optical chip when the heater obtains a heating pulse in the embodiment of the present application. The temperature detection unit 132 detects the temperature of the multimode optical waveguide in real time. The collection system is when the temperature of the multimode optical waveguide is T1, T2, ...T M When collecting data from the photodetector array 16, the number of photodetectors is N, then this pulse will collect M×N data, and P pulses are sent to the multimode optical waveguide 12 at each wavelength. When collecting training data, the amount of data is P×Q×M×N.

[0115] like Figure 8 As shown, the collected training data is spliced ​​and expanded. The data collected by the photodetector at one time is one-dimensional data with a length of N. The light intensity collected by each photodetector is compared with the total light intensity of all detectors to eliminate the influence of the change of the measured light intensity. When the multimode optical waveguide 12 is temperature modulated at the same central wavelength and data is collected at different temperatures, M groups of one-dimensional data with a length of N are obtained, which becomes a two-dimensional array of M×N.

[0116] Since the temperature modulation speed of the multimode optical waveguide 12 is very fast and the photodetector array 16 is quickly converted, the time to obtain a two-dimensional array will be very short after inputting a central wavelength and then performing a temperature modulation. The data of the M×N two-dimensional array is mapped to 0-255, and the two-dimensional array can be regarded as a grayscale image. Then, the image size is changed by the nearest neighbor interpolation and bilinear interpolation method, so that its pixels become 224*224. The amount of data collected is P×Q×M×N, so P×Q grayscale images with a pixel size of 224*224 are obtained.

[0117] The data processing module uses P*Q grayscale images as a training set, selects at least part of them as a training data set, and the rest as a validation training set, inputs the training set into a deep learning network model, and trains the deep learning network model. Exemplarily, 0.8×P×Q groups of images and their corresponding central wavelength target sequences are randomly selected from P×Q grayscale images as a training data set, where 0.8×P×Q images are used as input data, and the central wavelength corresponding to each image is used as target data, and the remaining 0.2×P×Q images and their corresponding central wavelengths are used as a validation data set.

[0118] When training the deep learning network model, 0.8×P×Q groups of grayscale images are used as training sets, and the model is iteratively trained. After each iteration, the verification set is input into the deep learning network model for verification. When the accuracy rate reaches more than 98%, the training can be stopped; after the training is completed, the trained deep learning network model parameters are saved.

[0119] As an optional implementation method, the data processing module uses the trained deep learning network model to analyze the speckle information, including the following process:

[0120] The photodetector array 16 analyzes the speckles at M different temperatures to obtain a one-dimensional array with a length of N corresponding to each temperature.

[0121] The data processing module normalizes the one-dimensional array with a length of N, converts M normalized one-dimensional arrays with a length of N into M*N two-dimensional arrays, and converts the two-dimensional array into a grayscale image.

[0122] The data processing module inputs the grayscale image into the deep learning network model to determine at least one attribute of the light wave to be measured and / or determine a change in at least one attribute of the light wave to be measured.

[0123] like Fig. 9 As shown, the embodiment of the present application also provides a light wave measurement method based on a multimode optical waveguide, the method comprising the following steps:

[0124] S1. Temperature modulation of the multimode optical waveguide.

[0125] S2. The light wave to be measured forms speckles through a multimode optical waveguide, and the speckles at M different temperatures are analyzed by a photodetector array to obtain a one-dimensional array of length N corresponding to each temperature.

[0126] S3. Normalize the one-dimensional array with a length of N.

[0127] S4. Convert M normalized one-dimensional arrays of length N into M*N two-dimensional arrays, convert the two-dimensional array into a grayscale image, and input the grayscale image into a deep learning network model to determine at least one attribute of the light wave to be measured and / or determine a change in at least one attribute of the light wave to be measured.

[0128] As an optional implementation, the light to be measured enters the multimode optical waveguide 12 and the temperature modulation module 13 through the optical access port 11, and forms speckles on the second end face of the multimode optical waveguide 12. The heating unit 131 heats the multimode optical waveguide 12 to change the temperature of the multimode optical waveguide 12, thereby changing the speckle information formed by the light to be measured on the multimode optical waveguide 12. The temperature detection unit 132 detects the temperature of the multimode optical waveguide 12 heated by the heating unit 131 in real time, and can record and obtain the speckle information at M different temperatures. The photodetector array 16 analyzes the speckles at M different temperatures, and obtains a stack of arrays corresponding to a wavelength of N at each temperature. After normalizing the stack of M arrays with a length of N, the arrays are converted into a two-dimensional array of M*N, and the two-dimensional array is converted into a grayscale image. Finally, the grayscale image is input into the deep learning network model, and at least one attribute and / or a change in at least one attribute of the light to be measured can be obtained.

[0129] As an optional implementation, the method further includes the following deep learning network model training steps:

[0130] Q optical waves to be measured with different central wavelengths are input, and P pulses are sent to the multimode optical waveguide at each central wavelength.

[0131] The temperature of the multimode optical waveguide is modulated, and the photodetector array analyzes the speckles at M different temperatures to obtain speckle information with a data volume of P*Q*M*N.

[0132] A one-dimensional array with a length of N obtained by each detection of the photodetector array is normalized.

[0133] For each pulse of the light wave to be measured of each wavelength, M normalized one-dimensional arrays of length N are converted into M*N two-dimensional arrays, and the two-dimensional arrays are converted into grayscale images to obtain P*Q grayscale images.

[0134] Take P*Q grayscale images as training sets, select at least part of them as training data sets, and the rest as verification training sets, input the training sets into the deep learning network model, and train the deep learning network model.

[0135] As an optional implementation method, the deep learning network model includes: convolution layer, pooling layer, and pooling layer.

[0136] The convolution layer is used to extract features from the grayscale image and output a feature vector; the pooling layer is used to perform a pooling operation on the feature vector output by the convolution layer to reduce the dimension of the feature vector; the fully connected layer is connected to the pooling layer to integrate the features output by the pooling layer, and finally the output of the fully connected layer is classified using the softmax function to determine at least one attribute of the light wave to be measured and / or determine the change of at least one attribute of the light wave to be measured. Among them, the deep learning network also includes: a preprocessing layer, a normalization layer, and an activation layer. Among them, the activation layer enables the network to solve nonlinear problems and uses the ReLU activation function; the pooling layer can remove redundant information, reduce the amount of calculation, and use maximum pooling downsampling.

[0137] In summary, the embodiments of the present application provide a light wave measurement device and method based on a multimode optical waveguide, which can effectively reduce the cost of the device and improve the collection efficiency by collecting speckle information through a photodetector array. In addition, by actively heating the multimode optical waveguide, the photodetector array collects the speckle information formed after the same light wave passes through the multimode optical waveguide at different temperatures, thereby solving the problem of losing a lot of information due to the limited number of photodetectors in the photodetector array when compressing the original speckle-rich information on the photodetector array used.

[0138] The above disclosure is only the preferred embodiment of the present application, but it is not intended to limit the scope of rights of the present application. A person of ordinary skill in the art can understand that without departing from the spirit and scope of the present application and the appended claims, changes, modifications, substitutions, combinations, and simplifications should all be equivalent replacement methods and still fall within the scope covered by the application.

Claims

1. A light wave measurement device based on a multimode optical waveguide, characterized in that: The device comprises: an optical access port including a single-mode optical waveguide; A temperature modulation module, used for performing temperature modulation on a multimode optical waveguide; A multimode optical waveguide, wherein a first end of the multimode optical waveguide is coupled with a single-mode optical waveguide of the optical access port, receives the optical wave to be measured and forms a speckle at a second end of the multimode optical waveguide; A one-to-many optical waveguide structure is used to couple the speckle emitted by the multimode optical waveguide into a multi-channel dielectric optical waveguide; a multi-channel dielectric optical waveguide for coupling the speckle into a photodetector array; a photodetector array for analyzing the speckle to determine at least one property of the light wave to be measured and / or to determine a change in at least one property of the light wave to be measured; The temperature modulation module comprises: A heating unit, used for heating the multimode optical waveguide; A temperature detection unit is used to detect the temperature change of the multimode optical waveguide in real time. The device further includes a data processing module, wherein the photodetector analyzes speckles formed by the same light wave to be measured at different temperatures, and the data processing module determines at least one attribute of the light wave to be measured and / or determines a change in at least one attribute of the light wave to be measured according to the speckles formed at each temperature; The photodetector array analyzes the speckles at M different temperatures to obtain a one-dimensional array with a length of N corresponding to each temperature; The data processing module normalizes the one-dimensional array with a length of N, converts M normalized one-dimensional arrays with a length of N into two-dimensional arrays of M*N, and converts the two-dimensional array into a grayscale image; The data processing module inputs the grayscale image into a deep learning network model to determine at least one attribute of the light wave to be measured and / or determine a change in at least one attribute of the light wave to be measured.

2. The device according to claim 1, characterized in that The data processing module is also used to train the deep learning network model. The multimode optical waveguide receives Q optical waves to be measured with different central wavelengths, and sends P pulses to the multimode optical waveguide at each central wavelength; Temperature modulation is performed on the multimode optical waveguide, a photodetector array analyzes the speckles at M different temperatures, and the data processing module receives speckle information with a data volume of P*Q*M*N; The data processing module performs normalization processing on a one-dimensional array of length N detected by the photoelectric detector array each time; For each pulse of the light wave to be measured of each wavelength, the data processing module converts M normalized one-dimensional arrays of length N into M*N two-dimensional arrays, and converts the two-dimensional arrays into grayscale images to obtain P*Q grayscale images; The data processing module uses P*Q grayscale images as training sets, selects at least part of them as training data sets, and the rest as verification training sets, inputs the training sets into the deep learning network model, and trains the deep learning network model.

3. The device according to claim 1, characterized in that The one-to-many optical waveguide structure includes a first tapered waveguide and a plurality of second tapered waveguides, wherein the number of the second tapered waveguides corresponds to the number of photodetectors in the photodetector array; The first end face of the first tapered waveguide is coupled to the second end of the multimode optical waveguide, the second end face of the first tapered waveguide is coupled to the first end face of any second tapered waveguide, the second end face of any second tapered waveguide is coupled to the first end of a dielectric optical waveguide, the first end face area of ​​the first tapered waveguide is smaller than the second end face area of ​​the first tapered waveguide, and the first end face area of ​​the second tapered waveguide is larger than the second end face area of ​​the second tapered waveguide; Wherein, when the areas of the first end surfaces of the plurality of second tapered waveguides are substantially equal, the second tapered waveguides are unevenly distributed on the second end surface of the first tapered waveguide; Alternatively, the plurality of second tapered waveguides included in the one-to-many optical waveguide structure include at least two types of second tapered waveguides, and each type of second tapered waveguide has a different first end surface area.

4. The device according to claim 1, characterized in that The attributes include one or more of the following: The wavelength of the light wave to be measured, the polarization of the light and the spatial mode.

5. A light wave measurement method based on a multimode optical waveguide, the light wave measurement method being applied to the light wave measurement device based on a multimode optical waveguide as claimed in any one of claims 1 to 4, characterized in that: The method comprises the following steps: performing temperature modulation on the multimode optical waveguide; The light wave to be measured forms speckles through the multimode optical waveguide, and the speckles at M different temperatures are analyzed by a photodetector array to obtain a one-dimensional array with a length of N corresponding to each temperature; Normalizing the one-dimensional array with a length of N; M normalized one-dimensional arrays of length N are converted into M*N two-dimensional arrays, and the two-dimensional arrays are converted into grayscale images, and the grayscale images are input into a deep learning network model to determine at least one attribute of the light wave to be measured and / or determine the change of at least one attribute of the light wave to be measured.

6. The method according to claim 5, characterized in that The method also includes the following training steps of the deep learning network model: Inputting Q optical waves to be measured with different central wavelengths, and sending P pulses to the multimode optical waveguide at each central wavelength; Temperature modulation is performed on the multimode optical waveguide, and a photodetector array analyzes the speckles at M different temperatures to obtain speckle information with a data volume of P*Q*M*N; Normalizing the one-dimensional array with a length of N detected by the photodetector array each time; For each pulse of the light wave to be measured of each wavelength, M normalized one-dimensional arrays of length N are converted into M*N two-dimensional arrays, and the two-dimensional arrays are converted into grayscale images to obtain P*Q grayscale images; Take P*Q grayscale images as training sets, select at least part of them as training data sets, and the rest as verification training sets, input the training sets into the deep learning network model, and train the deep learning network model.

7. The method according to claim 5 or 6, characterized in that: The deep learning network model includes: A convolutional layer, used to extract features from the grayscale image and output a feature vector; A pooling layer, used for performing a pooling operation on the feature vector output by the convolutional layer to reduce the dimension of the feature vector; A fully connected layer, connected to the pooling layer, for integrating the features output by the pooling layer; The output of the fully connected layer is classified using a softmax function to determine at least one attribute of the light wave to be measured and / or determine a change in at least one attribute of the light wave to be measured.

Images