A method of spectral detection

Abstract

The application relates to the technical field of optical equipment, and provides a spectrum detection method, which comprises an artificial synapse component and a signal processing component.The artificial synapse component comprises a plurality of artificial synapse devices, which are used for responding to different wavelengths of incident light and outputting corresponding induced currents; the signal processing component is connected with the plurality of artificial synapse devices, is used for receiving the induced currents and performing amplification, analog-digital conversion and analysis, so as to reconstruct spectrum information of the incident light, the artificial synapse component is directly irradiated by to-be-detected light, the corresponding induced currents are outputted by the artificial synapse component, the induced currents are transmitted to the signal processing component, and the spectrum information of the to-be-detected light can be outputted after processing, so that the detection architecture is simplified, the power consumption is reduced, and integrated application is facilitated without relying on complex scanning components.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of optical equipment, in particular to a spectrum detection method. BACKGROUND

[0002] Traditional spectrum detectors usually use prism, grating and other light splitting elements to decompose composite light into monochromatic light, and the core is based on light splitting detection principle, and light signals are received by CCD, CMOS and other array detectors. Due to the difference in wavelengths of different wavebands, different angles of refraction or diffraction will occur when passing through the prism or grating, and the sensing components capture these angle differences to realize the identification of the corresponding waveband. However, in order to realize the above-mentioned accurate detection process, such spectrum detectors are often complex in structure, and rely on precise mechanical scanning components, resulting in high system power consumption, and it is difficult to realize miniaturization and integration application. SUMMARY

[0003] The present application aims to improve at least one technical problem in the background art.

[0004] The present application provides a spectrum detection device, comprising:

[0005] An artificial synapse component comprising a plurality of artificial synapse devices for responding to different wavelengths of incident light and outputting corresponding induced currents;

[0006] A signal processing component connected to the plurality of artificial synapse devices for receiving the induced currents and amplifying, analog-to-digital converting and analyzing to reconstruct the spectral information of the incident light.

[0007] According to some technical solutions of the present application, the artificial synapse device comprises a first electrode, a second electrode and a functional layer, the functional layer is arranged between the first electrode and the second electrode, and the first electrode and / or the second electrode is connected to the signal processing component for leading the induced current to the signal processing component.

[0008] According to some technical solutions of the present application, the artificial synapse device comprises a first electrode, a second electrode and a functional layer, the first electrode and the second electrode are respectively arranged at both ends of the functional layer, and the first electrode and / or the second electrode is connected to the signal processing component for leading the induced current to the signal processing component.

[0009] According to some technical solutions of the present application, the material of the functional layer is selected from at least one of metal oxide, two-dimensional material, perovskite material or heterojunction thereof.

[0010] According to some technical solutions of the present application, the signal processing component comprises:

[0011] a current amplification module connected with the plurality of artificial synapse devices, configured to amplify the induced current to an analog current signal in a processable range;

[0012] an analog-digital conversion module connected with the current amplification module, configured to convert the amplified analog current signal into a digital signal;

[0013] a data processing module connected with the analog-digital conversion module, configured to analyze the wavelength distribution and intensity information of the incident light according to the digital signal;

[0014] a spectrum output module connected with the data processing module, configured to output the obtained wavelength distribution and intensity information.

[0015] The application further provides a spectrum detection method, comprising the spectrum detection device as described above, and the spectrum detection method comprises:

[0016] irradiating the artificial synapse assembly with the light to be detected;

[0017] adjusting the original response weight of the artificial synapse device according to the irradiation intensity of the light to be detected after irradiation;

[0018] outputting the corresponding induced current signal based on the adjusted original response weight;

[0019] processing the induced current signal to obtain a digital current signal;

[0020] inputting the digital current signal into a preset spectrum reconstruction model for analysis to obtain the spectrum information of the light to be detected.

[0021] According to some technical solutions of the application, the adjusting the original response weight of the artificial synapse device according to the irradiation intensity of the light to be detected after irradiation specifically comprises:

[0022] comparing the light intensity of the light to be detected with a preset threshold;

[0023] if the light intensity of the light to be detected is equal to the preset threshold, maintaining the original response weight;

[0024] if the light intensity of the light to be detected is greater than the preset threshold, increasing the original response weight;

[0025] if the light intensity of the light to be detected is less than the preset threshold, decreasing the original response weight.

[0026] According to some technical solutions of the application, the outputting the corresponding induced current signal based on the adjusted original response weight specifically comprises:

[0027] comparing the adjusted original response weight with a preset response weight;

[0028] If the adjusted original response weight is greater than the preset response weight, the induced current signal is enhanced;

[0029] If the adjusted original response weight is less than the preset response weight, the induced current signal is suppressed.

[0030] According to some technical solutions of the present application, the processing of the induced current signal to obtain a digital current signal specifically includes:

[0031] The induced current signal is amplified to obtain an analog current signal;

[0032] The analog current signal is converted to a digital current signal.

[0033] According to some technical solutions of the present application, before the digital current signal is input into the preset spectral reconstruction model for analysis to obtain the spectral information of the light to be detected, it further includes:

[0034] A plurality of known wavelength and light intensity optical signals are sequentially irradiated on the artificial synapse component;

[0035] The test current generated by the artificial synapse component on the known wavelength and light intensity optical signal is collected and recorded;

[0036] Based on the recorded test current and the corresponding standard spectral data, a spectral training data set is constructed;

[0037] The spectral reconstruction model is trained by a preset algorithm using the spectral training data set.

[0038] The spectral detection device provided by the present application has at least the following beneficial effects: the light to be detected directly irradiates the artificial synapse component, and the corresponding induced current is output by the artificial synapse component; the induced current is transmitted to the signal processing component, and after processing, the spectral information of the light to be detected can be output, so that the spectral detection device does not need to rely on a light splitting element and complex scanning components, simplifies the architecture of the detection device, reduces power consumption, and is convenient for integrated application. BRIEF DESCRIPTION OF DRAWINGS

[0039] Figure 1 A structural diagram of the spectral detection device provided by the present application is provided;

[0040] Figure 2 A structural diagram of the artificial synapse device provided by the present application is provided;

[0041] Figure 3 Another structural diagram of the artificial synapse device provided by the present application is provided;

[0042] Figure 4A response curve of an artificial synapse device provided by an embodiment of the present application to incident light of different wavelengths;

[0043] Figure 5 A flowchart of a spectrum detection method provided by an embodiment of the present application. DETAILED DESCRIPTION

[0044] Embodiments of the present application are described in detail below with reference to the accompanying drawings, in which the same or similar notations used throughout the drawings denote the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application and cannot be understood as a limitation of the present application.

[0045] In the description of the present application, it should be understood that the orientation description, such as the orientation or position relationship indicated by up, down, front, back, left, right, etc., is based on the orientation or position relationship shown in the drawings, and is only for the convenience of describing the present application and simplifying the description, and does not indicate or imply that the device or element referred to must have a particular orientation and be constructed, operated in a particular orientation, and therefore cannot be understood as a limitation of the present application.

[0046] In the description of the present application, unless otherwise explicitly limited, the words such as arrangement, installation, connection, etc. should be broadly understood, and those skilled in the art can reasonably determine the specific meaning of the above words in the present application in combination with the specific content of the technical solution.

[0047] The following will be described in combination with Figures 1 to 5 The embodiments of the present application are described.

[0048] The present application provides a spectrum detection device, which comprises:

[0049] The artificial synapse component 100 comprises a plurality of artificial synapse devices 110 for responding to different wavelengths of incident light and outputting corresponding induced currents; as a photoelectric conversion core element, the light sensing behavior of the synapse device mainly depends on the functional layer material, which is configured to produce differential carrier migration behavior under irradiation of incident light of different wavelengths, thereby producing a response curve related to the wavelength of the incident light;

[0050] Among them, the artificial synapse device has long-term potentiation (LTP) and long-term depression (LTD) characteristics. That is, when the device is subjected to strong stimulation, such as high-intensity light irradiation, the synaptic response weight will continue to rise, which is equivalent to remembering the strong signal, and the response weight corresponds to the conductive ability of the artificial synapse device 110, such as the size of the current; when the device is subjected to continuous weak stimulation or random stimulation without target, the synaptic response weight will continue to decrease, which is equivalent to ignoring weak signals or invalid signals.

[0051] The LTP and LTD characteristics of the artificial synapse are used to realize intelligent filtering of noise suppression and effective signal reservation, and to improve the signal-to-noise ratio of photoelectric conversion. Random light noise such as ambient stray light is a persistent weak invalid stimulus, which triggers LTD after irradiating the artificial synapse device 110, corresponding to the low synaptic weight of the artificial synapse, and suppresses the weight of the weak response generated by random light noise such as ambient stray light; the target light such as specific wavelength and high intensity as strong stimulus will trigger the LTP characteristics, so that the effective response weight is maintained or increased, so that the artificial synapse device 110 itself can filter out the background noise without additional high-pass filter design.

[0052] The signal processing assembly 200 is connected with a plurality of artificial synapse devices 110, and is used to receive the induced current and perform amplification, analog-to-digital conversion and analysis to reconstruct the spectral information of the incident light.

[0053] The artificial synapse assembly 100 is directly irradiated by the light to be detected, and the corresponding induced current is output by the artificial synapse assembly 100; the induced current is transmitted to the signal processing assembly 200, and after processing, the spectral information of the light to be detected can be output. Therefore, the entire detection process does not need to rely on complex scanning components, so that the detection architecture can be simplified, the power consumption can be reduced, and the integration application can be facilitated.

[0054] In some embodiments, the artificial synapse device 110 includes a planar structure and a vertical structure, and the vertical structure includes a first electrode, a second electrode and a functional layer, the functional layer is arranged between the first electrode and the second electrode, and the first electrode and / or the second electrode is connected with the signal processing assembly 200 to guide the induced current to the signal processing assembly 200.

[0055] Optionally, the planar structure includes a first electrode, a second electrode and a functional layer, the first electrode and the second electrode are arranged at both ends of the functional layer respectively, and the first electrode and / or the second electrode is connected with the signal processing assembly 200 to guide the induced current to the signal processing assembly 200. Through the double-electrode structure of the planar or vertical type, the miniaturization design requirement can be met.

[0056] Specifically, after the light to be detected irradiates the functional layer, the functional layer generates an induced current, and the current is guided to the signal processing assembly 200 through the first electrode and the second electrode or one of them, so that the electrode is used as a current transmission carrier to generate differentiated carrier migration behavior under irradiation of light at different wavelengths, form a wavelength-related response curve, and the current is stably transmitted through the electrode to avoid signal loss, thereby ensuring the stability of photoelectric conversion and the effectiveness of current transmission, and providing conditions for subsequent signal processing.

[0057] In some embodiments, the material of the functional layer is selected from at least one of a metal oxide, a two-dimensional material, a perovskite material, or a heterojunction thereof. The metal oxide can be TiO2, HfO2, etc., and the two-dimensional material can be, for example, MoS2, Bi2Se3, etc. The material selection is flexible and can adapt to the performance requirements of different application scenarios. The energy band structure of the selected material can be controlled by photoexcitation and ion migration, that is, when the functional layer is irradiated by incident light of different wavelengths, the behaviors of carrier migration, defect state trapping, oxygen vacancy ionization or deionization, etc. inside the material are different, forming a differentiated response curve to have wavelength-specific response capability, which can accurately distinguish different wavelengths of light and ensure the recognition accuracy of the artificial synapse device 110 to different wavelengths of light, providing a differentiated signal basis for spectral analysis.

[0058] In some embodiments, the signal processing assembly 200 includes a current amplification module 210, an analog-to-digital conversion module 220, a data processing module 230, and a spectral output module 240.

[0059] The current amplification module 210 is connected to the plurality of artificial synapse devices 110 and is configured to amplify the induced current to an analog current signal in a processable range. Specifically, the weak induced current of the artificial synapse device 110 is amplified by the amplification circuit to increase the current amplitude, which meets the input threshold requirement of subsequent analog-to-digital conversion, avoids the loss of weak current signal or the inability to be recognized by the subsequent module, and further ensures the integrity and processability of signal transmission.

[0060] The analog-to-digital conversion module 220 is connected to the current amplification module 210 and is configured to convert the amplified analog current signal into a digital signal. Analog signals are susceptible to interference, while digital signals have strong anti-interference performance. Through analog-to-digital conversion, the signal format is standardized, which not only improves the stability and anti-interference ability of the signal, but also provides an adaptive format for the algorithm analysis of the data processing module 230, reducing the analysis error.

[0061] The data processing module 230 is connected to the analog-to-digital conversion module 220 and is configured to analyze the wavelength distribution and intensity information of the incident light based on the digital signal. Based on a pre-set three-dimensional response matrix of “wavelength-intensity-current”, a mapping relationship between the digital signal and the spectral information is established using a nonlinear fitting or neural network algorithm, low-frequency background noise is filtered out, and spectral core features are retained. In this way, the spectral information is accurately reconstructed, and the signal-to-noise ratio is further improved through the high-pass filtering function of the synapse without the need for additional filter components.

[0062] Exemplarily, the current amplification module adopts a low-noise operational amplifier-based transimpedance amplifier to linearly convert the weak current output by the artificial synapse device into an analog voltage signal; the analog-to-digital conversion module adopts a 12-bit or higher successive approximation analog-to-digital converter to convert the voltage signal into a digital signal for analysis by the data processing module; the data processing module is responsible for running a spectral reconstruction algorithm to analyze the digital current signal into a wavelength-intensity distribution, and an embedded microcontroller is used for signal processing, and a spectral reconstruction model based on a fully connected neural network trained is internally solidified in the memory to analyze the digital current sequence into a wavelength-intensity distribution in real time; the spectral output module uploads the spectral data to the upper computer through a USB digital interface, or drives an OLED display screen through an interface to visualize the curve, thereby realizing the process flow from photoelectric response to spectral information reconstruction.

[0063] Optionally, the nonlinear fitting is to construct a parameterized mathematical model with current value as input and wavelength and intensity as output according to the known response curve of the artificial synapse device to light of different wavelengths; the neural network algorithm is used to regard the spectral reconstruction problem as a regression problem from multi-dimensional current signal to spectral vector for end-to-end learning.

[0064] The spectral output module 240 is connected with the data processing module 230, and is used for outputting the obtained wavelength distribution and intensity information. In actual configuration, the wavelength distribution and intensity information analyzed by the data processing module 230 are output in the form of visualized spectral graph or transmissible numerical report.

[0065] The application also provides a spectral detection method, which comprises the spectral detection device as described in the above embodiments, and the detection method comprises the following steps.

[0066] S100, irradiating a to-be-detected light to an artificial synapse assembly; the to-be-detected light is directly irradiated to the artificial synapse assembly to cover multiple artificial synapse devices in the artificial synapse assembly, since the functional layer of the artificial synapse device is sensitive to light, there is no need to pre-split light, and a corresponding response can be generated.

[0067] S200, after irradiation, adjusting the original response weight of the artificial synapse device according to the irradiation intensity of the to-be-detected light; the preset threshold is set based on the intensity difference between the target light and the noise light, and is used for distinguishing the effective signal and the interference signal.

[0068] Specifically, the artificial synapse assembly is built-in with a threshold judgment mechanism to detect the irradiation intensity of the incident light in real time and compare it with the preset threshold. Thus, in some embodiments, in the S200, after irradiation, the original response weight of the artificial synapse device is adjusted according to the irradiation intensity of the to-be-detected light, and specifically comprises:

[0069] S210, comparing the light intensity of the to-be-detected light with a preset threshold value;

[0070] S220, if the light intensity of the to-be-detected light is equal to the preset threshold value, maintaining the original response weight;

[0071] S230, if the light intensity of the to-be-detected light is greater than the preset threshold value, increasing the original response weight;

[0072] S240, if the light intensity of the to-be-detected light is less than the preset threshold value, decreasing the original response weight. If the light intensity of the to-be-detected light is greater than or equal to the preset threshold value, the original response weight is increased or maintained. The response weight corresponds to the conductive capacity of the artificial synapse device. When the light intensity is detected to reach or exceed the preset threshold value, the response weight of the artificial synapse component is increased or maintained. By using the long-term potentiation (LTP) characteristics of the artificial synapse, strong stimulation causes the response weight of the synapse to continuously increase, which is equivalent to remembering effective signals in actual configuration, so as to strengthen the response signal of the target light and ensure that the effective signal is not suppressed and the signal strength is improved.

[0073] If the light intensity of the to-be-detected light is less than the preset threshold value, the original response weight is decreased. When the light intensity is detected to be lower than the preset threshold value, it is determined that it is noise light such as environmental stray light, and the response weight of the artificial synapse component is decreased. By using the long-term depression (LTD) characteristics of the artificial synapse, weak stimulation causes the response weight of the synapse to continuously decrease, which is equivalent to ignoring invalid signals in actual configuration, so as to automatically suppress noise signals and improve the signal-to-noise ratio, without the need for additional high-pass filter components, thereby simplifying the circuit design.

[0074] Therefore, clear judgment basis is provided for subsequent weight adjustment, so as to realize preliminary screening of signals.

[0075] S300, outputting a corresponding induced current signal based on the adjusted original response weight.

[0076] In some embodiments, in the S300, the corresponding induced current signal is outputted based on the adjusted original response weight, specifically including:

[0077] S310, comparing the adjusted original response weight with a preset response weight. Since the response weight is positively correlated with the conductive capacity on the artificial synapse device, the weight change directly affects the size of the induced current, and the processing direction of the current signal is determined by comparison.

[0078] S320, if the adjusted original response weight is greater than the preset response weight, the induced current signal is enhanced; when the adjusted original response weight is higher than the preset response weight, the effective signal is strengthened, the induced current intensity output by the artificial synapse device is increased, the current characteristics of the effective signal are highlighted, and subsequent module identification and processing are facilitated.

[0079] S330, if the adjusted original response weight is less than the preset response weight, the induced current signal is suppressed. When the adjusted original response weight is lower than the preset response weight, the noise signal is suppressed, the induced current intensity output by the artificial synapse device is reduced, the interference of the noise signal is weakened, the signal purity is further improved, and the subsequent analysis error is reduced.

[0080] Therefore, by comparing the adjusted original response weight with the preset response weight and adjusting, the pertinence of signal processing is ensured.

[0081] S400, the induced current signal is processed to obtain a digital current signal;

[0082] Optionally, in some embodiments, in S400, the induced current signal is processed to obtain a digital current signal, specifically comprising:

[0083] The induced current may still be at a weak level after the weight adjustment, and thus, in S410, the induced current signal is amplified to obtain an analog current signal; the current amplification module receives the induced current output by the artificial synapse component, and amplifies it to a preset processable range through an amplification circuit. After amplification, the input requirement of the analog-to-digital conversion module is met, and signal loss is avoided.

[0084] Since the analog signal is susceptible to electromagnetic interference, in S420, the analog current signal is converted into a digital current signal through analog-to-digital conversion. The analog-to-digital conversion module receives the amplified analog current signal and converts it into a standardized digital current signal. The digital signal has strong stability and can improve the signal anti-interference ability.

[0085] S500, the digital current signal is input into a preset spectrum reconstruction model for analysis to obtain spectrum information of the to-be-detected light. The data processing module inputs the digital current signal into the preset spectrum reconstruction model to analyze the wavelength distribution and intensity information of the incident light through the preset spectrum reconstruction algorithm.

[0086] Optionally, the spectrum reconstruction model is trained based on a three-dimensional response matrix of wavelength, light intensity and current, and can inversely deduce the corresponding spectrum information through the digital signal. Specifically, in some embodiments, before S500, the digital current signal is input into the preset spectrum reconstruction model for analysis to obtain the spectrum information of the to-be-detected light, further comprising:

[0087] S600, the artificial synapse component is irradiated with a plurality of light signals of known wavelengths and light intensities in sequence;

[0088] S610, the test current generated by the artificial synapse component in response to the light signals of known wavelengths and light intensities is collected and recorded; the signal processing component collects and stores the test current generated by the artificial synapse component in response to each standard light signal in real time, wherein the test current is the response current for testing and recording.

[0089] S620, based on the recorded test current and corresponding standard spectrum data, a spectrum training data set is constructed; the one-to-one correspondence between the known light parameters and the response current is recorded to provide input data for model training. The recorded response current is paired with the corresponding standard spectrum data, i.e. the real spectrum corresponding to the known light parameters, to form a spectrum training data set, which provides conditions for improving model accuracy.

[0090] S630, using the spectrum training data set, a spectrum reconstruction model is trained by a preset algorithm. A standard monochromatic light source such as a tunable laser or an LED array is used to irradiate the artificial synapse component to cover a target wavelength range of, for example, 400-800 nm; after irradiation, the induced current at each wavelength is collected, and the corresponding standard spectrum data is recorded to construct a "wavelength-light intensity-current" three-dimensional response matrix as a training data set; a convolutional neural network (CNN) or a support vector machine (SVM) is used for model training, and the loss function is optimized to convergence; the trained model is solidified into the memory of the data processing module for calling during actual detection. Exemplarily, using the spectrum training data set, the model parameters can also be iteratively optimized by a nonlinear fitting or a neural network algorithm until the model analytical accuracy meets the preset requirements, for example, the training is iterated until the evaluation index of the model on the validation set is stable within the preset threshold for N consecutive iterations, or the maximum number of iterations is reached. At this time, it is considered that the model has converged and the accuracy meets the requirements, and the training is stopped and the final model parameters are saved. Thus, by establishing a stable spectrum reconstruction model, the accuracy of spectrum analysis in the subsequent detection process is ensured. Therefore, by obtaining diversified training samples, the model is ensured to adapt to light signals of different wavelengths and intensities, and the generalization ability is improved.

[0091] The preferred embodiments of the present application are described in detail above, but the present disclosure is not limited to the described embodiments, and those skilled in the art can make various equivalent modifications or replacements without departing from the spirit of the present application. These equivalent modifications or replacements are all included in the scope defined by the claims of the present disclosure.

Claims

1. A method of spectral exploration, characterized by: A method for operating a spectral detection device, the spectral detection device comprising: An artificial synapse assembly comprising a plurality of artificial synapse devices for responding to different wavelengths of incident light and outputting corresponding induced currents; A signal processing assembly connected to the plurality of artificial synapse devices for receiving the induced currents and amplifying, analog-to-digital converting, and analyzing to reconstruct spectral information of the incident light; the spectral detection method comprising: Irradiating the artificial synapse assembly with light to be detected; After irradiation, adjusting the original response weight of the artificial synapse device according to the irradiance of the light to be detected; Based on the adjusted original response weight, outputting the corresponding induced current signal; Processing the induced current signal to obtain a digital current signal; Inputting the digital current signal into a preset spectral reconstruction model for analysis to obtain the spectral information of the light to be detected.

2. The method of spectral exploration of claim 1, wherein: The method further comprises: Comparing the intensity of the light to be detected with a preset threshold; If the intensity of the light to be detected is equal to the preset threshold, maintaining the original response weight; If the intensity of the light to be detected is greater than the preset threshold, increasing the original response weight; If the intensity of the light to be detected is less than the preset threshold, decreasing the original response weight.

3. The method of spectral exploration of claim 1, wherein: The method further comprises: Comparing the adjusted original response weight with a preset response weight; If the adjusted original response weight is greater than the preset response weight, enhancing the induced current signal to output the corresponding induced current signal; If the adjusted original response weight is less than the preset response weight, suppressing the induced current signal to output the corresponding induced current signal.

4. The method of spectral exploration of claim 1, wherein: The method further comprises: Amplifying the induced current signal to obtain an analog current signal; Analog-to-digital converting the analog current signal to obtain a digital current signal.

5. The method of spectral exploration of claim 1, wherein: The method further comprises: Irradiating the artificial synapse assembly with a plurality of known wavelength and irradiance light signals in sequence; Collecting and recording the test current generated by the artificial synapse assembly in response to the known wavelength and irradiance light signals; Based on the recorded test current and corresponding standard spectral data, constructing a spectral training dataset; Using the spectral training dataset, training a spectral reconstruction model using a preset algorithm.

6. The method of spectral exploration of claim 1, wherein: The artificial synapse device comprises a first electrode, a second electrode, and a functional layer, the functional layer being disposed between the first electrode and the second electrode, the first electrode and / or the second electrode being connected to the signal processing assembly for leading the induced current to the signal processing assembly.

7. The method of spectral exploration of claim 1, wherein: The artificial synapse device comprises a first electrode, a second electrode, and a functional layer, the first electrode and the second electrode being disposed at both ends of the functional layer, the first electrode and / or the second electrode being connected to the signal processing assembly for leading the induced current to the signal processing assembly.

8. The method of spectral exploration according to claim 6 or 7, characterized in that: The material of the functional layer is selected from at least one of a metal oxide, a two-dimensional material, a perovskite material, or a heterojunction thereof.

9. The method of spectral exploration of claim 1, wherein: The signal processing assembly comprises: a current amplification module connected with the plurality of artificial synapse devices, configured to amplify the induced current to an analog current signal in a processable range; an analog-to-digital conversion module connected with the current amplification module, configured to convert the amplified analog current signal into a digital signal; a data processing module connected with the analog-to-digital conversion module, configured to analyze wavelength distribution and intensity information of the incident light according to the digital signal; a spectrum output module connected with the data processing module, configured to output the obtained wavelength distribution and intensity information.

Images