Multi-dimensional optical information detection method and system

Abstract

The invention relates to the technical field of optical detection, in particular to a multi-dimensional optical information detection method and system. Comprising the following steps: acquiring reference light information of a plurality of target light signals as a mark set, taking a light current response matrix as a signal set, establishing a multilayer perceptron model to output multi-dimensional light information, and randomly sampling the data set into a training set, a test set and a verification set according to the same random seed function; training the multi-layer perceptron model, and outputting the multi-layer perceptron model as a model file; and verifying the trained model file by adopting a verification set, if the obtained accuracy is greater than a first threshold range, determining that the model file is qualified, otherwise, repeating the steps by adopting a new random seed function to carry out iterative model training and verification until the accuracy is greater than the first threshold range. According to the invention, high-precision detection of multi-dimensional optical information such as polarization and wavelength is realized, the measurement precision and the data processing efficiency are remarkably improved, and the limitation that a traditional detection system can only realize two-dimensional parameter detection is effectively solved.

Description

Technical Field

[0001] The present invention relates to the field of light detection technology, and in particular to a multi-dimensional light information detection method and system. Background Art

[0002] Existing optical detection systems can usually only detect wavelength parameters, and it is difficult to obtain polarization information of light at the same time. Existing multi-dimensional optical detection systems usually require complex optical components and precise calibration processes, resulting in complex system structures and high costs. Summary of the invention

[0003] The present invention aims to at least improve one of the technical problems existing in the prior art. To this end, the present invention proposes a multi-dimensional optical information detection method and system.

[0004] The technical solution of the present invention is as follows: A multi-dimensional optical information detection method, comprising: S1, obtaining reference light information and photocurrent response matrix of multiple target light signals; S2, using the reference light information as a label set and the photocurrent response matrix as a signal set, establishing a multilayer perceptron model to output multidimensional light information; S3, according to the same random seed function, randomly sampling the corresponding data set consisting of the label set and the signal set in a first range ratio and dividing them into a training set, a test set and a validation set; S4, train the multi-layer perceptron model, and the loss function is as follows: , Where S is the reference polarization information, λ is the reference wavelength information, is the polarization information predicted by the model, is the wavelength information predicted by the model, is the L2 regularization equation, and β is the regularization factor; its function is to enhance network optimization and prevent overfitting of the network; When the minimum value of the loss function obtained after training the training set is less than the set threshold, the output is used as the model file; S5, use the verification set to verify the trained model file, if the accuracy is greater than the first threshold range, it is qualified, otherwise use a new random seed function to repeat S3 to S4 for iterative model training and verification until the accuracy is greater than the first threshold range.

[0005] In a possible technical solution, further, in S1, each element in the photocurrent response matrix is ​​determined by the polarization information S and wavelength information λ of the reference light information, and a photocurrent response matrix representing multi-dimensional light information is formed by applying different bias voltages and collecting current signals of each electrode of the multi-electrode pyroelectric thin film detector. The reference light information of the i-th incident light is ; Photocurrent response matrix It is expressed as follows: , Where S is the polarization information of the reference light information, λ is the wavelength information of the reference light information; U1, U2, ..., U M are different bias voltage values; I1, I2, I3, ..., I N is the current signal collected by the detector, and M and N are the electrode numbers.

[0006] In a possible technical solution, further, in S2, in the multilayer perceptron model, the input layer receives the vectorized data of the photocurrent response matrix; each hidden layer contains a connection matrix, an activation function and a bias term, wherein the neural activity of the Lth layer is obtained by the output of the previous layer through the weighting of the connection matrix, superimposing the bias and then passing through the activation function; the output layer reads the matrix and maps the activation function, and finally outputs multidimensional light information. The multilayer perceptron model includes an input layer, an output layer and multiple hidden layers. Each hidden layer includes a fully connected layer and an optimization model layer, wherein the input layer for: , The neural activity in the Lth layer in the hidden layer is expressed as: , in is the connection matrix from layer L-1 to layer L, is the activation function of the Lth layer, is the bias of the adjacent hidden layer, is the neural activity in layer L-1; The output layer is: in To read the matrix, is the activation function of the output layer, is the bias of the output unit, which is used to correct the bias of the output layer.

[0007] In a possible technical solution, further, each hidden layer includes a fully connected layer and a batch normalization layer, and a ReLU activation function is used in the hidden layer.

[0008] In a possible technical solution, further, in S1, the reference light information includes polarization S and wavelength λ of the light.

[0009] According to the multi-dimensional optical information detection method of the embodiment of the present invention, the polarization light sensitivity and wavelength sensitivity generated by the lattice symmetry breaking of the thermoelectric film and the substrate interface are combined with a deep learning neural network to achieve high-precision detection of multi-dimensional optical information such as polarization and wavelength. The system does not require complex optical components and spectrometer structures, and is suitable for multi-dimensional optical information detection from visible light to mid-infrared bands. It has the characteristics of simple structure, low cost and high measurement accuracy. At the same time, the use of deep learning algorithms for optical information decoding significantly improves the measurement accuracy and data processing efficiency, and effectively solves the limitation that traditional detection systems can only achieve two-dimensional parameter detection.

[0010] A multi-dimensional optical information detection system, used to implement the multi-dimensional optical information detection method as described above, comprising: a light source for emitting a light beam; An optical component for modulating the polarization state of the light beam; A multi-electrode pyroelectric thin film detector receives light beams of different polarization states to obtain reference light information, and generates current signals of different electrodes according to the reference light information; A signal acquisition and control unit, used for bias modulating the multi-electrode pyroelectric thin film detector and collecting current signals to form a photocurrent response matrix; A building module, used to use the reference light information as a label set and the photocurrent response matrix as a signal set to build a multilayer perceptron model; A sorting module is used to randomly sample the corresponding data set according to the same random seed function and divide it into a training set, a test set and a validation set in a first range ratio; The training module is used to train the multi-layer perceptron model. When the minimum value of the loss function obtained through training of the training set is less than the set threshold, the model file is output; A verification module is used to verify the trained model file. If the accuracy rate is greater than the first threshold range, the model is qualified. Otherwise, a new random seed function is used to repeat iterative model training and verification until the accuracy rate is greater than the first threshold range. An output unit outputs multi-dimensional light information.

[0011] In a possible technical solution, further, the multi-electrode thermoelectric thin film detector is a single single crystal thin film detector with thermoelectric effect and lattice symmetry breaking at the interface with the substrate material, having N signal output electrodes (N>3), a bias electrode and a common ground electrode, and is modulated by M biases of a certain step length.

[0012] In a possible technical solution, further, the signal acquisition and control unit includes: A bias modulation circuit is used for bias modulation of a multi-electrode pyroelectric thin film detector; The current signal acquisition module is used to acquire current signals.

[0013] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the above-mentioned multi-dimensional optical information detection method when executing the computer program.

[0014] A computer storage medium, wherein instructions are stored in the computer storage medium, and when the instructions are executed on a computer, the computer executes the multi-dimensional optical information detection method as described above.

[0015] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 is a flow chart of a multi-dimensional optical information detection method according to an embodiment of the present invention; Figure 2 is a schematic diagram of a multi-dimensional optical information detection system according to an embodiment of the present invention; Figure 3 is a graph showing polarization response characteristics of a pyroelectric thin film detector in a multi-dimensional optical information detection method according to an embodiment of the present invention; Figure 4 is a schematic diagram of a multi-layer perceptron in a multi-dimensional optical information detection method according to an embodiment of the present invention; Figure 5 It is a schematic diagram of a deep learning framework training loss curve in a multi-dimensional optical information detection method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The embodiments of the present invention are described in detail below. The embodiments described with reference to the accompanying drawings are exemplary. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0019] It should be noted that when an element is referred to as being "fixed to" another element, it may be directly on the other element or there may be an intermediate element. When an element is considered to be "connected to" another element, it may be directly connected to the other element or there may be an intermediate element at the same time.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention belongs. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0021] The terms "first", "second", "third", etc. in the specification and claims of the present application and the drawings are used to distinguish different objects rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a series of steps or units are included, or optionally, steps or units not listed are included, or optionally, other steps or units inherent to these processes, methods, products or devices are included.

[0022] Only the part relevant to the present application is shown in the accompanying drawings, but not all of the content. Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processing or methods depicted as flow charts. Although the flow chart describes each operation (or step) as a sequential process, many of the operations therein can be implemented in parallel, concurrently or simultaneously. In addition, the order of each operation can be rearranged. When its operation is completed, the process can be terminated, but it can also have additional steps not included in the accompanying drawings. The process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0023] The terms "component", "module", "system", "unit", etc. used in this specification are used to represent computer-related entities, hardware, firmware, a combination of hardware and software, software, or software in execution. For example, a unit can be, but is not limited to, a process running on a processor, a processor, an object, an executable file, an execution thread, a program, and / or distributed between two or more computers. In addition, these units can be executed from various computer-readable media having various data structures stored thereon. Units can communicate through local and / or remote processes, for example, based on signals having one or more data packets (e.g., data from a second unit interacting with another unit in a local system, a distributed system, and / or a network. For example, the Internet interacts with other systems via signals).

[0024] Example 1 As attached Figure 1 As shown, this embodiment provides a multi-dimensional optical information detection method, which includes: S1, obtain the reference light information and photocurrent response matrix of multiple target light signals, where the reference light information of the i-th incident light is ; Photocurrent response matrix It is expressed as follows: , Where S is the polarization information of the reference light information, λ is the wavelength information of the reference light information; U1, U2, ..., U M are different bias voltage values; I1, I2, I3, ..., I N is the current signal collected by the detector, and M and N are the electrode numbers.

[0025] S2, using the reference light information as a label set and the photocurrent response matrix as a signal set, a multilayer perceptron model is established to output multidimensional light information, wherein the multidimensional light information of the i-th incident light beam is decoded as: , in is a multi-layer perceptron, is the photocurrent response matrix of the i-th incident light; S3, according to the same random seed function, randomly sampling the corresponding data set consisting of the label set and the signal set in a first range ratio and dividing them into a training set, a test set and a validation set; S4, train the multi-layer perceptron model, and the loss function is as follows: , Where S is the reference polarization information, λ is the reference wavelength information, is the polarization information predicted by the model, is the wavelength information predicted by the model, is the L2 regularization equation, and β is the regularization factor; its function is to enhance network optimization and prevent overfitting of the network; When the minimum value of the loss function obtained after training the training set is less than the set threshold, the output is used as the model file; S5, use the verification set to verify the trained model file, if the accuracy is greater than the first threshold range, it is qualified, otherwise use a new random seed function to repeat S3 to S4 for iterative model training and verification until the accuracy is greater than the first threshold range.

[0026] It should be noted that in S2, each element in the photocurrent response matrix is ​​determined by the polarization information S and wavelength information λ of the reference light information. By applying different bias voltages and collecting the current signals of each electrode of the multi-electrode thermoelectric thin film detector, a photocurrent response matrix representing multi-dimensional light information is formed.

[0027] It should be noted that in S2, in the multilayer perceptron model, the input layer receives the vectorized data of the photocurrent response matrix; each hidden layer contains a connection matrix, an activation function and a bias term, wherein the neural activity of the Lth layer is obtained by the output of the previous layer through the weighting of the connection matrix, superimposing the bias and then passing through the activation function; the output layer reads the matrix and maps the activation function, and finally outputs multidimensional light information. In this embodiment, the multilayer perceptron model includes an input layer, an output layer and multiple hidden layers. The input layer for: , The neural activity in the Lth layer in the hidden layer is expressed as: , in is the connection matrix from layer L-1 to layer L, is the activation function of the Lth layer, is the bias of the adjacent hidden layer, is the neural activity in layer L-1; The output layer is: in To read the matrix, is the activation function of the output layer, is the bias of the output unit, which is used to correct the bias of the output layer.

[0028] It should be noted that, in S2 of this embodiment, each hidden layer includes a fully connected layer and a batch normalization layer, and a ReLU activation function is used in the hidden layer.

[0029] It should be noted that, in S1 of this embodiment, the reference light information includes the polarization S and wavelength λ of the light.

[0030] In this embodiment, a group of specific implementation cases are provided, in which a multi-dimensional optical information detection system is used to detect multi-dimensional light. The multi-electrode thermoelectric thin film detector in the detection system is an 8-electrode system, and the bias voltage is modulated in steps of 0.02V. The modulation range is from -0.1V to 0.1V, then M is 11 and N is 8.

[0031] Step 1: Use a multi-dimensional optical detection system with optical hot spot effect to collect multiple single-wavelength polarization modulated optical signals to obtain reference light information, and obtain the photocurrent response matrix of the electrode pyroelectric thin film detector ; like Figure 3 As shown, it is a polarization response characteristic curve of the pyroelectric thin film detector; it can be seen from the figure that the normalized response characteristics of the 8-electrode pyroelectric thin film detector to polarized light modulated at different angles with a central wavelength of 1550nm through a λ / 4 polarizer, indicating that the 8-electrode pyroelectric thin film detector has the ability to detect the polarization state of light; The reference light information of the i-th incident light is: ; Photocurrent response matrix It is expressed as follows: , The reference light information is the wavelength λ and polarization S of the single-wavelength polarization modulated light signal.

[0032] Step 2: Using the reference light information as a label set and the photocurrent response matrix as a signal set, a multi-layer perceptron model (MLP) is established to output multi-dimensional light information, wherein the multi-dimensional light information of the i-th incident light beam is decoded as: , in is a multi-layer perceptron, is the photocurrent response matrix; Then the photocurrent response matrix Input Figure 4 In the multilayer perceptron shown, there are only 3 hidden layers in the multilayer perceptron, and each hidden layer has 256 neurons. The multilayer perceptron model consists of an input layer, an output layer, and a hidden layer. Each hidden layer includes a fully connected layer and an optimized model layer. The ReLU activation function is used in the hidden layer, and the Sigmoid and Softmax functions are used as activation functions in the output layer. The input layer is: , Neural activity in layer L: , in is the connection matrix from layer L-1 to layer L, is the activation function of the Lth layer, is the bias of the hidden layer; The output layer is: , in is the read matrix, is the activation function of the output layer, is the bias of the output unit, which is used to correct the bias of the output layer.

[0033] Step 3: According to the same random seed function, the corresponding data set consisting of the label set and the signal set is randomly sampled into a training set, a test set, and a validation set in a ratio of 4:3:3; Step 4: Train the multi-layer perceptron (MLP) model. The loss function is as follows: , Where S is the reference polarization information, λ is the reference wavelength information, is the polarization information predicted by the model, is the wavelength information predicted by the model, is the L2 regularization equation, β is the regularization factor. In this embodiment, the multi-layer perceptron (MLP) model is used for training to enhance network optimization and prevent overfitting of the network. When the minimum value of the loss function obtained after training the training set is less than the set threshold, the output is used as the model file. In this embodiment, Figure 5 , which is a schematic diagram of the deep learning framework training loss curve in the multi-dimensional optical information detection method of the present invention. This example is trained through 5000 iterations, and the threshold value is 0.0001, and finally a model file is obtained.

[0034] Step 5: Use the validation set to verify the trained model file. If the accuracy is greater than 98.6%, it is qualified. Otherwise, use a new random seed function to repeat steps 3 to 4 for iterative model training and verification until the accuracy is greater than 98.6%.

[0035] According to the multi-dimensional optical information detection method of the embodiment of the present invention, the polarization light sensitivity and wavelength sensitivity generated by the lattice symmetry breaking of the thermoelectric film and the substrate interface are combined with a deep learning neural network to achieve high-precision detection of multi-dimensional optical information such as polarization and wavelength. The system does not require complex optical components and spectrometer structures, and is suitable for multi-dimensional optical information detection from visible light to mid-infrared bands. It has the characteristics of simple structure, low cost and high measurement accuracy. At the same time, the use of deep learning algorithms for optical information decoding significantly improves the measurement accuracy and data processing efficiency, and effectively solves the limitation that traditional detection systems can only achieve two-dimensional parameter detection.

[0036] Example 2 like Figure 2 As shown, this embodiment provides a multi-dimensional optical information detection system, which is used to implement the multi-dimensional optical information detection method as described above, including: The light source is used to emit a light beam. In the present embodiment, the light source is a broad-spectrum light source, which covers the wavelengths from visible light to mid-infrared light. It may be a thermal radiation light source or a laser, and its polarization state may be modulated by a polarizer.

[0037] An optical component for modulating the polarization state of the light beam; A multi-electrode pyroelectric thin film detector receives light beams of different polarization states to obtain reference light information, and generates current signals of different electrodes according to the reference light information; A signal acquisition and control unit, used for bias modulating the multi-electrode pyroelectric thin film detector and collecting current signals to form a photocurrent response matrix; A module is established, which is used to use the reference light information as a label set and the photocurrent response matrix as a signal set to establish a multi-layer perceptron model, wherein the multi-dimensional light information of the i-th incident light beam is decoded as: , in is a multi-layer perceptron, is the photocurrent response matrix; A sorting module is used to randomly sample the corresponding data set according to the same random seed function and divide it into a training set, a test set and a validation set in a first range ratio; The training module is used to train the multi-layer perceptron model. The loss function is as follows: , Where S is the reference polarization information, λ is the reference wavelength information, is the polarization information predicted by the model, is the wavelength information predicted by the model, is the L2 regularization equation, β is the regularization factor; the purpose of using the multi-layer perceptron (MLP) model for training is to enhance network optimization and prevent overfitting of the network.

[0038] When the minimum value of the loss function obtained after training the training set is less than the set threshold, the output is used as the model file; A verification module is used to verify the trained model file. If the accuracy rate is greater than the first threshold range, the model is qualified. Otherwise, a new random seed function is used to repeat iterative model training and verification until the accuracy rate is greater than the first threshold range. An output unit outputs multi-dimensional light information.

[0039] It should be noted that, in this embodiment, the optical component includes: Polarizer, used to select light of a specific polarization direction, allowing light in a certain direction to pass through and blocking light in other directions; Wave plates, which change the polarization state of light through the birefringence effect, such as converting linearly polarized light into circularly polarized light, or adjusting the polarization direction, preferably λ / 4 wave plates or λ / 2 wave plates; Filters are used to select light of a specific wavelength, allowing a certain wavelength band to pass through and blocking other wavelength bands. The polarization state and wavelength of the incident light are modulated by combining the above optical components.

[0040] It should be noted that, in this embodiment, the multi-electrode pyroelectric thin film detector adopts anisotropic pyroelectric materials and has the characteristic of being sensitive to light of different polarizations and wavelengths.

[0041] As a further preference of this embodiment, the material of the multi-electrode thermoelectric thin film detector is lead selenide (PbSe), and the substrate material is strontium titanate (SrTiO3). The multi-electrode thermoelectric thin film detector is a single single crystal thin film detector with thermoelectric effect and lattice symmetry breaking at the interface with the substrate material, and has N signal output electrodes (N>3), a bias electrode and a common ground electrode, and is modulated by M biases of a certain step length.

[0042] It should be noted that the training process of deep learning neural network includes: Dataset classification: The labeled set signal set is shuffled through the same random seed function, and the training set, test set and validation set are sampled; Deep learning training: Training is performed based on a deep learning framework that combines data-driven and physical models to obtain a training model, which is then tested on a test set. Verification test: The trained model is verified using the verification set. If the accuracy is greater than 99%, it passes. Otherwise, unqualified data in the verification set is extracted into the training set for iterative verification.

[0043] It should be noted that, in this embodiment, the signal acquisition and control unit includes: A bias modulation circuit is used for bias modulation of a multi-electrode pyroelectric thin film detector; The current signal acquisition module is used to acquire current signals.

[0044] A multidimensional optical information detection system in an embodiment of the present application is suitable for multidimensional optical information detection from visible light to mid-infrared bands, and has the advantages of simple structure, low cost and high measurement accuracy.

[0045] A multi-dimensional optical information detection system in the embodiment of the present application may be a device, or a component, integrated circuit, or chip in a terminal. The device may be a mobile electronic device or a non-mobile electronic device. Exemplarily, the mobile electronic device may be a mobile phone, a tablet computer, a laptop computer, a PDA, an ultra-mobile personal computer (UMPC), a netbook, or a personal digital assistant (PDA), etc., and the non-mobile electronic device may be a server, a network attached storage (NAS), a personal computer (PC), etc., which is not specifically limited in the embodiment of the present application.

[0046] A multi-dimensional optical information detection system in the embodiment of the present application may be a device having an operating system. The operating system may be an Android operating system, an iOS operating system, or other possible operating systems, which are not specifically limited in the embodiment of the present application.

[0047] The multi-dimensional optical information detection system provided in the embodiment of the present application can realize Figure 1 The various processes of the method embodiment of a multi-dimensional optical information detection method are implemented, and to avoid repetition, they are not repeated here.

[0048] The multidimensional optical information detection system according to the embodiment of the present invention realizes high-precision detection of multidimensional optical information such as polarization and wavelength through the polarization light sensitivity and wavelength sensitivity generated by the lattice symmetry breaking of the thermoelectric film and the substrate interface, combined with a deep learning neural network. The system does not require complex optical components and spectrometer structures, and is suitable for multidimensional optical information detection from visible light to mid-infrared bands. It has the characteristics of simple structure, low cost and high measurement accuracy. At the same time, the use of deep learning algorithms for optical information decoding significantly improves the measurement accuracy and data processing efficiency, and effectively solves the limitation that traditional detection systems can only realize two-dimensional parameter detection.

[0049] Optionally, an embodiment of the present application also provides an electronic device, including a processor, a memory, and a program or instruction stored in the memory and executable on the processor. When the program or instruction is executed by the processor, each process of the above-mentioned multi-dimensional optical information detection method embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0050] An embodiment of the present application also provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, each process of the above-mentioned multi-dimensional optical information detection method embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0051] The processor is the processor in the electronic device in the above embodiment. The readable storage medium includes a computer readable storage medium, such as a computer read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0052] Example 3 This embodiment provides a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and the processor implements the multi-dimensional optical information detection method as described above when executing the computer program.

[0053] Example 4 This embodiment provides a computer storage medium, wherein the computer storage medium stores instructions, and when the instructions are executed on a computer, the computer executes the multi-dimensional optical information detection method as described above.

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

[0055] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example.

[0056] Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Mentioning "embodiment" in this article means that the specific features, structures or characteristics described in conjunction with the embodiment may be included in at least one embodiment of the present embodiment application. The appearance of this phrase in various positions in the specification is not necessarily the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It can be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0057] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A multi-dimensional optical information detection method, characterized in that: include: S1, obtaining reference light information and photocurrent response matrix of multiple target light signals; S2, using the reference light information as a label set and the photocurrent response matrix as a signal set, and establishing a multilayer perceptron model to output multidimensional light information; S3, according to the same random seed function, randomly sampling the corresponding data set consisting of the label set and the signal set in a first range ratio and dividing them into a training set, a test set and a validation set; S4, training the multilayer perceptron model, when the minimum value of the loss function obtained after training the training set is less than the set threshold, output it as a model file; S5, use the verification set to verify the trained model file, if the accuracy is greater than the first threshold range, it is qualified, otherwise use a new random seed function to repeat S3 to S4 for iterative model training and verification until the accuracy is greater than the first threshold range.

2. The multi-dimensional optical information detection method according to claim 1, characterized in that: In S2, each element in the photocurrent response matrix is ​​determined by the polarization information S and wavelength information λ of the reference light information. By applying different bias voltages and collecting the current signals of each electrode of the multi-electrode thermoelectric thin film detector, a photocurrent response matrix representing multi-dimensional light information is formed.

3. The multi-dimensional optical information detection method according to claim 1, characterized in that: In S2, in the multi-layer perceptron model, the input layer receives the vectorized data of the photocurrent response matrix; each hidden layer contains a connection matrix, an activation function, and a bias term, where the neural activity of the Lth layer is obtained by the output of the previous layer through the weighting of the connection matrix, superimposing the bias, and then passing through the activation function; The output layer reads the matrix and the activation function mapping, and finally outputs multi-dimensional light information.

4. The multi-dimensional optical information detection method according to claim 2, characterized in that: In S1, the reference light information of the i-th incident light is , the photocurrent response matrix It is expressed as follows: ， Where S is the polarization information of the reference light information, λ is the wavelength information of the reference light information; U1, U2, ..., U M are different bias voltage values; I1, I2, I3, ..., I N is the current signal collected by the detector, and M and N are the electrode numbers.

5. The multi-dimensional optical information detection method according to claim 3, characterized in that: In S2, the multilayer perceptron model includes an input layer, an output layer, and multiple hidden layers. for: ， The neural activity in the Lth layer in the hidden layer is expressed as: ， in is the connection matrix from layer L-1 to layer L, is the activation function of the Lth layer, is the bias of the adjacent hidden layer, is the neural activity in layer L-1; The output layer is: in To read the matrix, is the activation function of the output layer, is the bias of the output unit, which is used to correct the bias of the output layer.

6. A multi-dimensional optical information detection system, used to implement the multi-dimensional optical information detection method according to any one of claims 1 to 5, characterized in that: include: a light source for emitting a light beam; An optical component for modulating the polarization state of the light beam; A multi-electrode pyroelectric thin film detector receives light beams of different polarization states to obtain reference light information, and generates current signals of different electrodes according to the reference light information; A signal acquisition and control unit, used for bias modulating the multi-electrode pyroelectric thin film detector and collecting current signals to form a photocurrent response matrix; A building module, used to use the reference light information as a label set and the photocurrent response matrix as a signal set to build a multilayer perceptron model; A sorting module is used to randomly sample the corresponding data set according to the same random seed function and divide it into a training set, a test set and a validation set in a first range ratio; The training module is used to train the multi-layer perceptron model. When the minimum value of the loss function obtained through training of the training set is less than the set threshold, the model file is output; A verification module is used to verify the trained model file. If the accuracy rate is greater than the first threshold range, the model is qualified. Otherwise, a new random seed function is used to repeat iterative model training and verification until the accuracy rate is greater than the first threshold range. An output unit outputs multi-dimensional light information.

7. The multi-dimensional optical information detection system according to claim 6, characterized in that: The multi-electrode thermoelectric thin film detector is a single crystal thin film detector with thermoelectric effect and lattice symmetry breaking at the interface with the substrate material. It has N signal output electrodes: N>3, a bias electrode and a common ground electrode, and is modulated by M biases of a certain step length.

8. The multi-dimensional optical information detection system according to claim 6, characterized in that: The signal acquisition and control unit comprises: A bias modulation circuit is used for bias modulation of a multi-electrode pyroelectric thin film detector; The current signal acquisition module is used to acquire current signals.

9. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the multi-dimensional optical information detection method as claimed in any one of claims 1 to 5 when executing the computer program.

10. A computer storage medium, characterized in that: The computer storage medium stores instructions, and when the instructions are executed on a computer, the computer executes the multi-dimensional optical information detection method according to any one of claims 1 to 5.

Images