Optical orbital angular momentum identification method and system based on binary neural network model

Abstract

The invention provides an optical orbital angular momentum identification method and system based on a binary neural network model, and is used for solving the technical problems that an OAM identification method based on a spatial light modulator deep learning method is high in calculation complexity, large in memory occupation and difficult to meet the application requirements of light weight and real-time performance of space laser communication. According to the optical orbital angular momentum identification method based on the binary neural network model provided by the invention, the identification of different topological charges OAM is realized by constructing the binary neural network model, and compared with a traditional OAM identification method based on a spatial light modulator deep learning method, the method provided by the invention adopts a binary weight and a binary activation function, so that the identification accuracy is improved. And the parameter storage requirement is obviously reduced. Meanwhile, according to the binary neural network model, PopCount processing is adopted to replace convolution operation, the calculation amount is reduced, the real-time performance is enhanced, and the requirement of application scenes such as inter-satellite communication for rapid recognition can be met.

Description

Technical Field

[0001] The present invention belongs to the technical field of space laser communication, and in particular relates to an optical orbital angular momentum identification method and system based on a binary neural network model. Background Art

[0002] With the deepening of space exploration activities and the growing demand for global information transmission, space laser communication technology has attracted widespread attention due to its advantages such as high speed, large capacity, and low latency. As an important physical quantity of light waves, optical orbital angular momentum (OAM) has unique application value in the field of optical communications. By utilizing the diversity of OAM, multiplexing can be achieved, thereby greatly improving the capacity and efficiency of optical communications. Accurate identification of OAM is one of the key technologies to realize its application. Therefore, it is of great significance to study efficient new OAM identification methods.

[0003] Existing OAM identification methods mainly include interference method, coordinate transformation method and deep learning method based on spatial light modulator. Interference method uses the number of spiral fringes or staggered fringes generated by the interference of OAM light beam with plane wave or spherical wave to realize OAM identification; coordinate transformation method is a method based on mathematical transformation, which realizes OAM identification by mapping the complex amplitude field of the light beam to a transformation space; deep learning method based on spatial light modulator uses deep neural network to extract and classify features of OAM images, so as to realize OAM identification. Deep learning method based on spatial light modulator has the advantages of wide recognition range and high recognition accuracy, but its computational complexity is high and memory usage is large, which makes it difficult to meet the application requirements of lightweight and real-time space laser communication. Summary of the invention

[0004] The purpose of the present invention is to solve the technical problems that the OAM identification method based on the deep learning method of the spatial light modulator has high computational complexity, large memory usage, and is difficult to meet the application requirements of lightweight and real-time space laser communication, and to provide an optical orbital angular momentum identification method and system based on a binary neural network model.

[0005] In order to achieve the above purpose, the technical solution provided by the present invention is as follows:

[0006] An optical orbital angular momentum identification method based on a binary neural network model is characterized in that it comprises the following steps:

[0007] S1, build a binary neural network model;

[0008] The binary neural network model includes an input layer, at least one processing block and an output layer connected in sequence according to input and output; the input layer is used to receive the interference image data of OAM and perform binarization processing on the interference image data of OAM; the processing block includes a binary convolution layer, a normalization layer, and a binary activation layer arranged in sequence; the binary convolution layer is used to perform convolution operation on the interference image data after binarization processing through XOR and PopCount processing to extract feature information; the normalization layer is used to perform normalization processing on the data output by the binary convolution layer; the binary activation layer is used to perform nonlinear mapping on the data output by the normalization layer; the output layer is used to output the final recognition result of OAM;

[0009] S2, collecting interference image data of different OAMs and preprocessing them;

[0010] S3, establishing a training data set and a test data set, and inputting the pre-processed OAM interference image data into the training data set and the test data set respectively in proportion;

[0011] S4, training of binary neural network model;

[0012] The training parameters of the binary neural network model are set, and the OAM interference image data in the training data set is input into the binary neural network model to train the binary neural network model, so as to iteratively update the weight of the binary neural network model and obtain a trained binary neural network model;

[0013] S5, OAM identification;

[0014] The OAM interference image data in the test data set is input into the binary neural network model trained in step S4 for classification and recognition to obtain an OAM recognition result.

[0015] Further, in step S2, the OAM interference image data is obtained by:

[0016] The signal light and the local oscillator light are respectively subjected to wavefront phase modulation to obtain modulated signal light and local oscillator light, the modulated signal light and the local oscillator light are combined, and the interference image data of OAM is obtained by detection.

[0017] Further, in step S4, the OAM interference image data in the training data set is input into the binary neural network model to train the binary neural network model specifically as follows:

[0018] After rotating, stretching and twisting the training data set, the OAM interference image data in the training data set are input into the binary neural network model in random order, and the binary neural network model is trained using the back propagation algorithm and the Adam adaptive learning rate optimization method.

[0019] Furthermore, in step S1, the binary neural network model adopts a binarization function as an activation function; and the number of the processing blocks is four.

[0020] Furthermore, in step S3, the pre-processed OAM interference image data is input into the training data set and the test data set respectively in a ratio of 7:3.

[0021] Furthermore, step S2 specifically includes: collecting interference image data of different OAMs, and performing filtering and downsampling processing on them in sequence.

[0022] In addition, the present invention also provides an optical orbital angular momentum identification system based on a binary neural network model, which is used to implement the above-mentioned optical orbital angular momentum identification method based on a binary neural network model, comprising a spatial light modulation unit, a spatial light interference unit and an OAM identification unit; the spatial light modulation unit is used to perform wavefront phase modulation on the generated signal light and emit it; the spatial light interference unit comprises a receiving optical antenna, a local oscillator laser, a second collimator, a second polarizer, a spectroscope and a CCD detection module; the receiving optical antenna is used to receive and focus the signal light emitted by the spatial light modulation unit; the second collimator and the second polarizer are sequentially arranged on the laser emission optical path of the local oscillator laser, and are used to sequentially collimate and filter the polarization state of the emitted local oscillator light;

[0023] The special feature is that the spatial light interference unit also includes a second control module and a second wavefront modulation module; the second control module is used to generate a wavefront phase distribution to be modulated on the local oscillator light; the output end of the second control module is connected to the input end of the second wavefront modulation module, and the second wavefront modulation module is located on the optical path of the local oscillator light after being filtered by the polarization state of the second polarizer, and is used to load the generated wavefront phase distribution to be modulated on the local oscillator light onto the second wavefront modulation module to achieve wavefront phase modulation of the local oscillator light; the spectroscope is simultaneously located on the optical path of the local oscillator light modulated by the second wavefront modulation module and the signal light transmitted by the receiving optical antenna, and is used to combine the modulated local oscillator light and the received signal light; the CCD detection module (16) is located on the output optical path of the spectroscope, and is used to detect the OAM interference image of the signal light and the local oscillator light;

[0024] The OAM recognition unit includes an image acquisition module, an image preprocessing module and a classification recognition module;

[0025] The input end of the image acquisition module is connected to the output end of the CCD detection module, and is used to collect OAM interference image data; the input end of the image preprocessing module is connected to the output end of the image acquisition module, and is used to preprocess the collected OAM interference image data; the input end of the classification and recognition module is connected to the output end of the image preprocessing module, and a binary neural network model is arranged inside the classification and recognition module, which is used to extract the characteristic information of the OAM interference image data through the binary neural network model and integrate and classify it to obtain the OAM recognition result;

[0026] The binary neural network model includes an input layer, at least one processing block and an output layer which are connected in sequence according to input and output; the input layer is used to receive the interference image data of OAM and perform binarization processing on the interference image data of OAM; the processing block includes a binary convolution layer, a normalization layer and a binary activation layer which are arranged in sequence; the binary convolution layer is used to perform convolution operation on the interference image data after binarization processing through XOR and PopCount processing to extract feature information; the normalization layer is used to perform normalization processing on the data output by the binary convolution layer; the binary activation layer is used to perform nonlinear mapping on the data output by the normalization layer; and the output layer is used to output the final recognition result of OAM.

[0027] Further, the spatial light modulation unit includes a signal laser, an optical power amplifier, a first collimator, a first polarizer, a first control module, a first wavefront modulation module, a reflector, and a transmitting optical antenna;

[0028] The signal laser is used to generate signal light based on continuous laser; the optical power amplifier, the first collimator and the first polarizer are sequentially arranged on the output optical path of the signal light, and are used to perform power amplification, collimation and polarization state filtering on the signal light in sequence; the first control module is used to generate a wavefront phase distribution to be modulated on the signal light; the output end of the first control module is connected to the input end of the first wavefront modulation module, and the first wavefront modulation module is located on the optical path of the signal light after polarization state filtering by the first polarizer, and is used to load the generated wavefront phase distribution to be modulated on the signal light onto the first wavefront modulation module to realize wavefront phase modulation of the signal light; the reflector is located on the optical path of the signal light after modulated by the first wavefront modulation module, and is used to change the direction of the light beam; the transmitting optical antenna is located on the reflecting optical path of the reflector, and is used to perform beam processing on the modulated signal light and then emit it;

[0029] The first wavefront modulation module is a reflective spatial light modulator (SLM) or a transmissive liquid crystal phased array;

[0030] The second wavefront modulation module is a reflective spatial light modulator (SLM) or a transmissive liquid crystal phased array.

[0031] Furthermore, the image preprocessing module includes a filtering module and a downsampling module;

[0032] The input end of the filtering module is connected to the output end of the image acquisition module, and is used to filter the input OAM interference image data;

[0033] The input end of the downsampling module is connected to the output end of the filtering module, and is used to extract and interpolate the filtered OAM interference image data;

[0034] The input end of the classification recognition module is connected to the output end of the downsampling module.

[0035] Furthermore, the filtering module uses a Gaussian filter, a mean filter, or a median filter to implement filtering; the downsampling module uses a nearest neighbor interpolation method or a bilinear interpolation method to implement interpolation processing of OAM interference image data;

[0036] The binary neural network model adopts a binarization function as an activation function; the number of the processing blocks is four.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] 1. The present invention provides an optical orbital angular momentum identification method based on a binary neural network model. By constructing a binary neural network model, the identification of OAMs with different topological charges is realized. Compared with the traditional OAM identification method based on the deep learning method of the spatial light modulator, the present invention adopts binary weights and binary activation functions, and the parameter storage requirements are significantly reduced. It is suitable for resource-constrained space laser communication systems. At the same time, the binary neural network model adopts PopCount processing instead of convolution operation, which reduces the amount of calculation and enhances real-time performance, and can meet the needs of rapid identification in application scenarios such as intersatellite communication.

[0039] 2. The present invention provides an optical orbital angular momentum recognition method based on a binary neural network model, which first performs a binarization process on the interference image data of OAM input into the binary neural network model, and has a high recognition accuracy and good robustness for the noise of the interference image data of OAM.

[0040] 3. The optical orbital angular momentum identification system based on the binary neural network model provided by the present invention realizes wavefront phase modulation of the local oscillator light through the second wavefront modulation module and the second control mode, thereby realizing the removal or reduction of the OAM topological charge of the local oscillator light. Only a smaller topological charge needs to be identified to obtain the topological charge result of the signal light received by the receiving end, which reduces the complexity of identifying large OAM topological charge numbers and has better flexibility. At the same time, a binary neural network model is set in the OAM identification unit, which greatly reduces the amount of calculation and enhances real-time performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a structural schematic diagram of an embodiment of an optical orbital angular momentum identification system based on a binary neural network model of the present invention;

[0042] Figure 2 is a schematic diagram of the structure of an OAM identification unit in an embodiment of the present invention;

[0043] Figure 3 3 are interference images of the signal light and the local oscillator light that has not undergone wavefront phase modulation when the OAM topological charge of the signal light in the embodiment of the present invention is 1, 2, and 3, respectively, wherein (a) is the interference image of the signal light and the local oscillator light that has not undergone wavefront phase modulation when the OAM topological charge is 1, (b) is the interference image of the signal light and the local oscillator light that has not undergone wavefront phase modulation when the OAM topological charge is 2, and (c) is the interference image of the signal light and the local oscillator light that has not undergone wavefront phase modulation when the OAM topological charge is 3;

[0044] Figure 4 1 is an intensity distribution diagram and a phase distribution diagram of the signal light after wavefront phase modulation in an embodiment of the present invention, wherein (a) is an intensity distribution diagram of the signal light after wavefront phase modulation, and (b) is a phase distribution diagram of the signal light after wavefront phase modulation;

[0045] Figure 5 : are intensity distribution diagrams and phase distribution diagrams of the local oscillation light after wavefront phase modulation in an embodiment of the present invention, wherein (a) is the intensity distribution diagram of the local oscillation light after wavefront phase modulation, and (b) is the phase distribution diagram of the local oscillation light after wavefront phase modulation;

[0046] Figure 6 : are interference images and phase distribution diagrams of the signal light after wavefront phase modulation and the local oscillator light after wavefront phase modulation in an embodiment of the present invention, wherein (a) is the interference image of the signal light after wavefront phase modulation and the local oscillator light after wavefront phase modulation, and (b) is the phase distribution diagram of the interference image;

[0047] Figure 7 The figure is a schematic diagram of the classification result of the confusion matrix obtained by performing a classification experiment on the OAM interference image data of the test data set using the binary neural network model of an embodiment of the present invention.

[0048] The specific reference numerals are as follows:

[0049] 1-Signal laser; 2-Optical power amplifier; 3-First collimator; 4-First polarizer; 5-First control module; 6-First wavefront modulation module; 7-Reflector; 8-Transmitting optical antenna; 9-Receiving optical antenna; 10-Local oscillator laser; 11-Second collimator; 12-Second polarizer; 13-Second control module; 14-Second wavefront modulation module; 15-Beam splitter; 16-CCD detection module. DETAILED DESCRIPTION

[0050] In order to make the advantages and features of the present invention more clear, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0051] like Figure 1 As shown, an optical orbital angular momentum identification system based on a binary neural network model includes a spatial light modulation unit, a spatial light interference unit and an OAM identification unit. The spatial light modulation unit is used as a transmitting end, and includes a signal laser 1, an optical power amplifier 2, a first collimator 3, a first polarizer 4, a first control module 5, a first wavefront modulation module 6, a reflector 7 and a transmitting optical antenna 8. The spatial light interference unit is used as a receiving end, and includes a receiving optical antenna 9, a local oscillator laser 10, a second collimator 11, a second polarizer 12, a second control module 13, a second wavefront modulation module 14, a spectroscope 15 and a CCD detection module 16.

[0052] The signal laser 1 is used to generate a stable continuous laser and use the generated continuous laser as the signal light. The optical power amplifier 2, the first collimator 3, and the first polarizer 4 are sequentially arranged on the output optical path of the signal light. Among them, the optical power amplifier 2 is used to amplify the power of the signal light to meet the spatial transmission requirements of a certain distance; the first collimator 3 is used to collimate the power-amplified signal light and radiate the collimated signal light from the optical fiber into the space; the first polarizer 4 is used to filter the polarization state of the collimated signal light.

[0053] The first control module 5 is used to generate a wavefront phase distribution to be modulated onto the signal light, the output end of the first control module 5 is connected to the input end of the first wavefront modulation module 6, and the first wavefront modulation module 6 is located on the optical path of the signal light after the polarization state is filtered by the first polarizer 4, and is used to load the generated wavefront phase distribution to be modulated onto the signal light onto the first wavefront modulation module 6 to achieve wavefront phase modulation of the signal light. The reflector 7 is located on the optical path of the signal light modulated by the first wavefront modulation module 6, and is used to change the direction of the light beam. The transmitting optical antenna 8 is located on the reflecting optical path of the reflector, and is used to transmit the modulated signal light after beam processing.

[0054] The receiving optical antenna 9 receives and focuses the signal light emitted by the transmitting optical antenna 8 and transmitted through the transmission channel. The local oscillator laser 10 is used to generate local oscillator light, and the second collimator 11 and the second polarizer 12 are sequentially arranged on the output optical path of the local oscillator light. Among them, the second collimator 11 is used to collimate the local oscillator light and couple it from the optical fiber into the space; the second polarizer 12 is used to filter the polarization state of the collimated local oscillator light.

[0055] The second control module 13 is arranged in the computer, and is used to generate the wavefront phase distribution to be modulated on the local oscillator light. The output end of the second control module 13 is connected to the input end of the second wavefront modulation module 14, and the second wavefront modulation module 14 is located on the optical path of the local oscillator light after the polarization state is filtered by the second polarizer 12, and is used to load the generated wavefront phase distribution to be modulated on the local oscillator light onto the second wavefront modulation module 14, so as to realize the wavefront phase modulation of the local oscillator light. The spectroscope 15 is located on the optical path of the local oscillator light modulated by the second wavefront modulation module 14 and the signal light emitted by the receiving optical antenna 9, and is used to combine the modulated local oscillator light and the received signal light. At this time, the signal light and the local oscillator light will produce coaxial interference. The CCD detection module 16 is located on the output optical path of the spectroscope 15, and is used to detect the interference image of the signal light and the local oscillator light.

[0056] The first wavefront modulation module 6 and the second wavefront modulation module 14 can be, but not limited to, reflective spatial light modulators (SLMs), transmissive liquid crystal phased arrays, etc. In this embodiment, the first wavefront modulation module 6 and the second wavefront modulation module 14 are both reflective spatial light modulators, denoted as SLM1 and SLM2, respectively.

[0057] The OAM identification unit is set in the computer, which includes an image acquisition module, an image preprocessing module and a classification and identification module. The input end of the image acquisition module is connected to the output end of the CCD detection module 16, which is used to collect OAM interference image data. The image preprocessing module includes a filtering module and a downsampling module; the input end of the filtering module is connected to the output end of the image acquisition module, which is used to filter the input OAM interference image data, and specifically, a Gaussian filter, a mean filter or a median filter can be used to implement filtering; the input end of the downsampling module is connected to the output end of the filtering module, which is used to extract and interpolate the filtered OAM interference image data, and specifically, the nearest neighbor interpolation method or the bilinear interpolation method can be used to implement the interpolation processing of the OAM interference image data. The input end of the classification and identification module is connected to the output end of the image preprocessing module, and a binary neural network model is set inside it, which is used to extract the feature information of the OAM interference image data through the binary neural network model and integrate and classify it to obtain the OAM identification result. In this embodiment, the second control module 13 and the OAM identification unit are operated using the same computer.

[0058] like Figure 2 As shown, the binary neural network model includes an input layer, four processing blocks and an output layer connected in sequence according to input and output. The input layer is used to receive the interference image data of OAM, and perform binary processing on the interference image data of OAM, converting it into image data with only black and white grayscales for subsequent processing. The present invention can use a global threshold method, a local threshold method or an adaptive threshold method to achieve binary processing. The processing block includes a binary convolution layer, a normalization layer and a binary activation layer arranged in sequence. Among them, the binary convolution layer is used to perform an XOR operation on the interference image data after the binary processing, and count the number of 1s in the XOR result through PopCount processing, thereby realizing the function of convolution operation and extracting feature information from the interference image data of OAM; the normalization layer is used to perform normalization processing on the data output by the binary convolution layer; the binary activation layer is used to perform nonlinear mapping on the data output by the normalization layer. The output layer is used to output the final recognition result of OAM.

[0059] The signal light generated in this embodiment is a vortex plane wave, and the local oscillation light is a spherical wave. When the vortex plane wave of the signal light and the spherical wave of the local oscillation light perform coaxial interference, assuming that the OAM topological charge of the signal light is l1, and the OAM topological charge of the local oscillation light is l2, the topological charge of the OAM interference image generated after the interference of the signal light and the local oscillation light is l3. Due to the spiral phase structure of the vortex light beam, the number of spiral interference fringes in the interference image is l3, and the orbital angular momentum mode can be effectively identified by the number and rotation direction of the spiral fringes in the interference image.

[0060] Specifically, the electric field E of the signal light vortex plane wave in this embodiment is sig for:

[0061]

[0062] The electric field E of the LO light spherical wave in this embodiment is loc for:

[0063]

[0064] Among them, A sig represents the amplitude of the signal light vortex plane wave, A loc represents the amplitude of the local oscillator light spherical wave, k is the wave number, ω0 is the beam waist radius, r 2 =(x-x0) 2 +(y-y0) 2 , x0 and y0 are the horizontal and vertical coordinates of the center of the local oscillator light spherical wave, respectively, and z is the axial coordinate during coaxial interference. i is an imaginary unit, is the phase of the signal light.

[0065] When the wavefront phase modulation is not performed on the LO light, the OAM topological charge number l2 of the LO light is 0, and the electric field of the LO light degenerates to:

[0066]

[0067] The light intensity distribution I of the interference image on the plane z = 0 is:

[0068]

[0069] It can be seen from the above formula that the interference fringes in the interference image on the plane z≠0 are spiral-shaped. The number of interference fringes l3 in the figure is equal to the OAM topological charge l1 of the signal light minus the OAM topological charge l2 of the local oscillator light, that is, l3=l1-l2. Among them, since l2=0 in this embodiment, the rotation direction of the interference fringes is related to the positive and negative OAM topological charge l1 of the signal light vortex plane wave. When l1 is negative, the interference fringes rotate clockwise; when l1 is positive, the interference fringes rotate counterclockwise.

[0070] like Figure 3 As shown, the interference images of the signal light and the local oscillator light are respectively 1, 2, and 3 when the OAM topological charge number l1 of the signal light is 1, 2, and 3, wherein (a) is the interference image of the signal light and the local oscillator light that has not undergone wavefront phase modulation when the OAM topological charge number of the signal light is 1, (b) is the interference image of the signal light and the local oscillator light that has not undergone wavefront phase modulation when the OAM topological charge number is 2, and (c) is the interference image of the signal light and the local oscillator light that has not undergone wavefront phase modulation when the OAM topological charge number is 3. In this embodiment, the wavelength λ of the signal light and the local oscillator light are both set to 1550nm, the horizontal and vertical coordinates of the spatial range of the observed interference image are set to [-100λ, +100λ], and the phase center distance between the received signal light and SLM2 is 500λ.

[0071] Figure 4 are the intensity distribution diagram and phase distribution diagram of the signal light after wavefront phase modulation, Figure 5 are the intensity distribution diagram and phase distribution diagram of the local oscillator light after wavefront phase modulation, Figure 6 Figure 2 is the interference image and phase distribution diagram of the signal light after wavefront phase modulation and the local oscillator light after wavefront phase modulation. It can be seen that the OAM topological charge of the signal light after wavefront phase modulation is l1=3, and the OAM topological charge of the local oscillator light after wavefront phase modulation is l2=1. The number of interference fringes on the interference image is l3=l1-l2=3-1=2, that is, the third-order OAM topological charge is reduced to the second-order OAM topological charge, which effectively reduces the complexity of identifying large OAM topological charges.

[0072] Based on the above optical orbital angular momentum identification system based on the binary neural network model, the present invention provides an optical orbital angular momentum identification method based on the binary neural network model, which specifically includes the following steps:

[0073] S1, build a binary neural network model.

[0074] S1.1, build the above-mentioned binary neural network model, wherein the number of processing blocks can be first set according to experience and then adjusted according to the output result of the binary neural network model.

[0075] S1.2, determine the number of layers, dimensions, convolution kernel size and other hyperparameters of each layer of the binary neural network model: In this embodiment, the number of processing blocks is four, the binary convolution layer uses a 9x9 convolution kernel for convolution operation, and uses a normalization layer to normalize the data output by the binary convolution layer. In this embodiment, the fully connected layer is implemented in the form of a convolution layer in the fourth processing block.

[0076] S1.3, select activation function: In this embodiment, a binarization function is used as the activation function to complete the nonlinear mapping, that is, f(x) = sign(x); where x is the input value, f(x) is the output value of the activation function, and the binarization function limits the output value to between 0 and 1.

[0077] The number of processing blocks and the size of the convolution kernel in the binary neural network structure of the present invention can be determined as needed. The embodiments of the present invention can also use different training data sets, such as OAM with different topological charges, OAM interference image data under different noise conditions, etc.

[0078] S2, collects interference image data of different OAMs and preprocesses them.

[0079] The interference image data of different OAMs are collected by the CCD detection module 16, and the collected OAM interference image data are filtered and downsampled by the image preprocessing module to improve the robustness and generalization ability of the algorithm, specifically including:

[0080] a) Filtering: Filter the input OAM interference image data to remove noise interference in the image. The filter may be, but is not limited to, a Gaussian filter, a mean filter, or a median filter.

[0081] b) Downsampling: The filtered OAM interference image data is subjected to decimation and interpolation processing to reduce the amount of image data and improve computational efficiency. The interpolation methods include but are not limited to the nearest neighbor interpolation method and the bilinear interpolation method.

[0082] S3, establishing a training data set and a test data set, inputting the interference image data preprocessed in step S2 into the training data set and the test data set respectively in proportion, and marking the OAM of the interference image data in the training data set respectively.

[0083] In this embodiment, the signal light generates a vortex plane wave, and the local oscillation light generates a spherical wave. The interference image data of the vortex plane wave with an orbital angular momentum topological charge of 1 to 8 and the spherical wave are used as the acquisition data of the sample data set. The orbital angular momentum topological charge corresponding to the light intensity distribution icon is used as the label, and the data in the sample data set is divided into a training data set and a test data set in a ratio of 7:3.

[0084] S4, training of binary neural network model;

[0085] S4.1, set training parameters: set training optimization strategy, initial learning rate, number of iterations and other hyperparameters.

[0086] The training optimization strategy adopts the Adam adaptive learning rate optimization method. All learnable parameters in the binary neural network model can be updated by setting the training optimization strategy. The initial learning rate is set to 0.01, which is the step size for adjusting parameters of the binary neural network model during training; the number of iterations is set to 100. In addition, a learning rate scheduler is set. When the loss of the binary neural network model training set does not improve within two cycles, the learning rate scheduler will reduce the learning rate so that the new learning rate is half of the current learning rate. The above parameter settings are conducive to overfitting of the binary neural network model during training and accelerate model convergence.

[0087] S4.2, model training.

[0088] After rotating, stretching and twisting the training data set, the interference image data of OAM in the training data set are input into the binary neural network model in random order. The binary neural network model is trained using the back propagation algorithm and the Adam adaptive learning rate optimization method to iteratively update the weights of the binary neural network model and obtain a trained binary neural network model.

[0089] S5, OAM identification;

[0090] The OAM interference image data in the test data set is input into the trained binary neural network model, and the OAM interference image data in the test data set is classified and identified to obtain an OAM identification result, that is, an estimated value of the topological charge number l3 of the OAM interference image in the test data set is obtained. In this embodiment, the value of l3 is between 1 and 8.

[0091] Figure 7The figure is a schematic diagram of the classification result of the confusion matrix obtained by using the binary neural network model of the embodiment of the present invention to classify the OAM interference image of the test data set. The experimental results show that the optical orbital angular momentum identification method and system based on the binary neural network model of the present invention can effectively identify OAM with different topological charges and has a high recognition accuracy.

[0092] At the same time, the present invention does not require complex calculations in convolution and nonlinear processing, has high computational efficiency, and the weights and activation functions of the binary neural network model are both binarized, so its memory occupancy is significantly reduced compared to traditional deep learning methods.

[0093] The above description is only used to illustrate the technical solution of the present invention rather than to limit it. For ordinary professional and technical personnel in the field, the specific technical solution recorded in the above embodiment can be modified, or some of the technical features therein can be replaced by equivalents, and these modifications or replacements do not make the essence of the corresponding technical solution deviate from the scope of the technical solution protected by the present invention.

Claims

1. An optical orbital angular momentum identification method based on a binary neural network model, characterized in that: The following steps are involved: S1, build a binary neural network model; The binary neural network model includes an input layer, at least one processing block and an output layer connected in sequence according to input and output; the input layer is used to receive the interference image data of OAM and perform binary processing on the interference image data of OAM; the processing block includes a binary convolution layer, a normalization layer, and a binary activation layer arranged in sequence; the binary convolution layer is used to perform convolution operation on the interference image data after the binary processing through XOR and PopCount processing to extract feature information; The normalization layer is used to normalize the data output by the binary convolution layer; the binary activation layer is used to perform nonlinear mapping on the data output by the normalization layer; The output layer is used to output the final recognition result of OAM; S2, collecting interference image data of different OAMs and preprocessing them; S3, establishing a training data set and a test data set, and inputting the pre-processed OAM interference image data into the training data set and the test data set respectively in proportion; S4, training of binary neural network model; The training parameters of the binary neural network model are set, and the OAM interference image data in the training data set is input into the binary neural network model to train the binary neural network model, so as to iteratively update the weight of the binary neural network model and obtain a trained binary neural network model; S5, OAM identification; The OAM interference image data in the test data set is input into the binary neural network model trained in step S4 for classification and recognition to obtain an OAM recognition result.

2. The optical orbital angular momentum identification method based on a binary neural network model according to claim 1, characterized in that: In step S2, the OAM interference image data is obtained by: The signal light and the local oscillator light are respectively subjected to wavefront phase modulation to obtain modulated signal light and local oscillator light, the modulated signal light and the local oscillator light are combined, and the interference image data of OAM is obtained by detection.

3. The optical orbital angular momentum identification method based on a binary neural network model according to claim 1 or 2, characterized in that: In step S4, the OAM interference image data in the training data set is input into the binary neural network model to train the binary neural network model specifically as follows: After rotating, stretching and twisting the training data set, the OAM interference image data in the training data set are input into the binary neural network model in random order, and the binary neural network model is trained using the back propagation algorithm and the Adam adaptive learning rate optimization method.

4. The optical orbital angular momentum identification method based on a binary neural network model according to claim 3 is characterized in that: In step S1, the binary neural network model uses a binarization function as an activation function; The number of the processing blocks is four.

5. The optical orbital angular momentum identification method based on a binary neural network model according to claim 4 is characterized in that: In step S3, the pre-processed OAM interference image data is input into the training data set and the test data set in a ratio of 7:

3.

6. The optical orbital angular momentum identification method based on a binary neural network model according to claim 5, characterized in that: Step S2 is specifically as follows: The interference image data of different OAMs are collected and filtered and downsampled in turn.

7. An optical orbital angular momentum identification system based on a binary neural network model, used to implement the optical orbital angular momentum identification method based on a binary neural network model as described in any one of claims 1 to 6, comprising a spatial light modulation unit, a spatial light interference unit and an OAM identification unit; the spatial light modulation unit is used to perform wavefront phase modulation on the generated signal light and emit it; the spatial light interference unit comprises a receiving optical antenna (9), a local oscillator laser (10), a second collimator (11), a second polarizer (12), a spectroscope (15) and a CCD detection module (16); the receiving optical antenna (9) is used to receive and focus the signal light emitted by the spatial light modulation unit; the second collimator (11) and the second polarizer (12) are sequentially arranged on the laser emission optical path of the local oscillator laser (10) to sequentially collimate and filter the polarization state of the emitted local oscillator light; Features: The spatial light interference unit also includes a second control module (13) and a second wavefront modulation module (14); The second control module (13) is used to generate a wavefront phase distribution to be modulated onto the local oscillator light; the output end of the second control module (13) is connected to the input end of the second wavefront modulation module (14), and the second wavefront modulation module (14) is located on the optical path of the local oscillator light after the polarization state is filtered by the second polarizer (12), and is used to load the generated wavefront phase distribution to be modulated onto the local oscillator light onto the second wavefront modulation module (14), so as to realize wavefront phase modulation of the local oscillator light; the beam splitter (15) is simultaneously located on the optical path of the local oscillator light modulated by the second wavefront modulation module (14) and the signal light transmitted by the receiving optical antenna (9), and is used to combine the modulated local oscillator light and the received signal light; the CCD detection module (16) is located on the output optical path of the beam splitter (15), and is used to detect the OAM interference image of the signal light and the local oscillator light; The OAM recognition unit includes an image acquisition module, an image preprocessing module and a classification recognition module; The input end of the image acquisition module is connected to the output end of the CCD detection module (16) and is used to acquire OAM interference image data; the input end of the image preprocessing module is connected to the output end of the image acquisition module and is used to preprocess the acquired OAM interference image data; the input end of the classification and recognition module is connected to the output end of the image preprocessing module and is provided with a binary neural network model inside thereof, which is used to extract characteristic information of the OAM interference image data through the binary neural network model and integrate and classify it to obtain an OAM recognition result; The binary neural network model includes an input layer, at least one processing block and an output layer connected in sequence according to input and output; the input layer is used to receive the interference image data of OAM and perform binary processing on the interference image data of OAM; the processing block includes a binary convolution layer, a normalization layer, and a binary activation layer arranged in sequence; the binary convolution layer is used to perform convolution operation on the interference image data after the binary processing through XOR and PopCount processing to extract feature information; The normalization layer is used to normalize the data output by the binary convolution layer; the binary activation layer is used to perform nonlinear mapping on the data output by the normalization layer; The output layer is used to output the final recognition result of OAM.

8. The optical orbital angular momentum identification system based on the binary neural network model according to claim 7, characterized in that: The spatial light modulation unit comprises a signal laser (1), an optical power amplifier (2), a first collimator (3), a first polarizer (4), a first control module (5), a first wavefront modulation module (6), a reflector (7) and a transmitting optical antenna (8); The signal laser (1) is used to generate signal light based on continuous laser; the optical power amplifier (2), the first collimator (3) and the first polarizer (4) are sequentially arranged on the output optical path of the signal light, and are used to perform power amplification, collimation and polarization state filtering on the signal light in sequence; the first control module (5) is used to generate a wavefront phase distribution to be modulated on the signal light; the output end of the first control module (5) is connected to the input end of the first wavefront modulation module (6), and the first wavefront modulation module (6) is located on the optical path of the signal light after polarization state filtering by the first polarizer (4), and is used to load the generated wavefront phase distribution to be modulated on the signal light onto the first wavefront modulation module (6), so as to realize wavefront phase modulation of the signal light; the reflector (7) is located on the optical path of the signal light after modulation by the first wavefront modulation module (6), and is used to change the direction of the light beam; the transmitting optical antenna (8) is located on the reflected optical path of the reflector (7), and is used to perform beam processing on the modulated signal light and then transmit it; The first wavefront modulation module (6) is a reflective spatial light modulator or a transmissive liquid crystal phased array; The second wavefront modulation module (14) is a reflective spatial light modulator or a transmissive liquid crystal phased array.

9. The optical orbital angular momentum identification system based on the binary neural network model according to claim 8, characterized in that: The image preprocessing module includes a filtering module and a downsampling module; The input end of the filtering module is connected to the output end of the image acquisition module, and is used to filter the input OAM interference image data; The input end of the downsampling module is connected to the output end of the filtering module, and is used to extract and interpolate the filtered OAM interference image data; The input end of the classification recognition module is connected to the output end of the downsampling module.

10. The optical orbital angular momentum identification system based on the binary neural network model according to claim 9, characterized in that: The filtering module uses a Gaussian filter, a mean filter or a median filter to achieve filtering; The downsampling module uses the nearest neighbor interpolation method or the bilinear interpolation method to implement the interpolation processing of the OAM interference image data; The binary neural network model uses a binarization function as an activation function; The number of the processing blocks is four.

Images