Signal detection method, device and equipment and computer readable storage medium

Abstract

The embodiment of the invention discloses a signal detection method, device and equipment and a computer readable storage medium. The method comprises the following steps: acquiring an observation map of a to-be-detected signal; extracting intensity change data, frequency change data and chromatographic change data of the to-be-detected signal from the observation map; intensity characteristic information is extracted from the intensity change data, frequency characteristic information is extracted from the frequency change data, and chromatographic characteristic information is extracted from the chromatographic change data; and detecting the signal type of the to-be-detected signal based on the intensity characteristic information, the frequency characteristic information and the chromatographic characteristic information to obtain a type detection result. Through the method, the accuracy of signal type detection can be improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular, to a signal detection method, apparatus, device, and computer-readable storage medium. Background Art

[0002] A pulsar is a highly magnetized rotating compact star, such as a neutron star and a white dwarf, which can periodically emit electromagnetic pulse signals. When exploring pulsars through astronomical observations, a space telescope can collect various signals, and technicians can then search for pulsar signals from them.

[0003] In the related art, usually, according to the frequency and pulse width of the pulsar signal, a detection rule is constructed to detect the feature matching degree of each signal with the pulsar signal to determine whether the signal is emitted by a pulsar. However, interference signals may also disguise themselves as signals with similar pulsar characteristics, resulting in inaccurate judgment results for pulsar signals.

[0004] Therefore, how to improve the detection accuracy of signals is an urgent problem to be solved. Summary of the Invention

[0005] To solve the above technical problems, embodiments of this application provide a signal detection method, apparatus, device, and computer-readable storage medium.

[0006] Among them, the technical solution adopted in this application is as follows:

[0007] A signal detection method includes:

[0008] Obtaining an observation graph of the signal to be detected;

[0009] Extracting the intensity change data, frequency change data, and chromatographic change data of the signal to be detected from the observation graph;

[0010] Extracting intensity feature information from the intensity change data, frequency feature information from the frequency change data, and chromatographic feature information from the chromatographic change data;

[0011] Based on the intensity feature information, the frequency feature information, and the chromatographic feature information, detecting the signal type of the signal to be detected to obtain a type detection result.

[0012] A signal detection apparatus includes:

[0013] An obtaining unit, configured to obtain an observation graph of the signal to be detected;

[0014] A processing unit for extracting the intensity change data, frequency change data, and chromatographic change data of the signal to be detected from the observation graph;

[0015] The processing unit is further configured to extract intensity feature information from the intensity change data, extract frequency feature information from the frequency change data, and extract chromatographic feature information from the chromatographic change data;

[0016] A detection unit for detecting the signal type of the signal to be detected based on the intensity feature information, the frequency feature information, and the chromatographic feature information, and obtaining a type detection result.

[0017] In an embodiment of the present application, based on the foregoing solution, the detection unit is further configured to detect the correlation between the intensity feature information, the frequency feature information, and the chromatographic feature information; the processing unit is further configured to perform feature fusion on the intensity feature information, the frequency feature information, and the chromatographic feature information based on the correlation to obtain fused feature information; the detection unit is further configured to detect the signal type of the signal to be detected based on the fused feature information to obtain the type detection result.

[0018] In an embodiment of the present application, based on the foregoing solution, the acquisition unit is further configured to obtain the weights corresponding to the intensity feature information, the frequency feature information, and the chromatographic feature information respectively based on the correlation; the processing unit is further configured to perform weighted fusion on the intensity feature information, the frequency feature information, and the chromatographic feature information based on the weights to obtain the fused feature information.

[0019] In an embodiment of the present application, based on the foregoing solution, the detection unit is further configured to detect the probability that the signal to be detected is a pulsar signal based on the fused feature information; if the probability is greater than a preset probability threshold, the signal type of the signal to be detected is determined to be a pulsar signal type; if the probability is less than the preset probability threshold, the signal type of the signal to be detected is determined to be a radio frequency interference signal type.

[0020] In an embodiment of the present application, based on the foregoing solution, the processing unit is further configured to sample the intensity change data to obtain intensity sampling data, sample the frequency change data to obtain frequency sampling data, and sample the chromatographic change data to obtain chromatographic sampling data; perform feature extraction on the intensity sampling data to obtain the intensity feature information, perform feature extraction on the frequency sampling data to obtain the frequency feature information, and perform feature extraction on the chromatographic sampling data to obtain the chromatographic feature information.

[0021] In one embodiment of the present application, based on the foregoing solution, the processing unit is further configured to collect the signal strength value corresponding to each sampling point from the intensity change data according to a preset sampling strategy to obtain intensity sampling data; the preset sampling strategy includes a sampling strategy for setting multiple time points and a sampling strategy for setting the number of samplings; collect the signal frequency value corresponding to each sampling point from the frequency change data according to the preset sampling strategy to obtain frequency sampling data; and collect the signal chromatogram value corresponding to each sampling point from the chromatogram change data according to the preset sampling strategy to obtain chromatogram sampling data.

[0022] In one embodiment of the present application, based on the foregoing solution, the processing unit is further configured to input the intensity sampling data into an intensity feature detection model for feature extraction to obtain the intensity feature information; input the frequency sampling data into a frequency feature detection model for feature extraction to obtain frequency feature information; and input the chromatogram sampling data into a chromatogram feature detection model for feature extraction to obtain chromatogram feature information.

[0023] In one embodiment of the present application, based on the foregoing solution, the acquisition unit is further configured to acquire a plurality of signal samples and the observation map sample corresponding to each signal sample, where the signal samples include pulsar signal samples and radio frequency interference signal samples; the processing unit is further configured to extract the intensity change data sample, the frequency change data sample, and the chromatogram change data sample of the signal sample from the observation map sample of each signal sample; train a first initial model based on the intensity change data sample of each signal sample to obtain an intensity feature detection model, train a second initial model based on the frequency change data sample of each signal sample to obtain a frequency feature detection model, and train a third initial model based on the chromatogram change data sample of each signal sample to obtain a chromatogram feature detection model.

[0024] In one embodiment of the present application, based on the foregoing solution, the processing unit is further configured to sample from the change data sample to obtain a sampling data sample, where the change data sample is any one of the intensity change data sample, the frequency change data sample, and the chromatogram change data sample; input the sampling data sample into an initial model matching the sampling data sample, so that the initial model performs feature extraction on the sampling data sample to obtain predicted feature information; the initial model is any one of the first initial model, the second initial model, and the third initial model; construct a loss based on the actual feature information corresponding to the sampling data sample and the predicted feature information; the apparatus further includes a training unit configured to train the initial model based on the loss.

[0025] A signal detection device includes a processor and a memory. Computer-readable instructions are stored on the memory, and when the computer-readable instructions are executed by the processor, the above signal detection method is implemented.

[0026] A computer-readable storage medium stores computer-readable instructions thereon. When the computer-readable instructions are executed by a processor of a computer, the computer is caused to execute the above signal detection method.

[0027] A computer program product includes computer-readable instructions, and when the computer-readable instructions are executed by a processor, the above signal detection method is implemented.

[0028] In the above technical solution:

[0029] An observation diagram of a signal to be detected can be obtained, and intensity change data, frequency change data, and chromatographic change data of the signal to be detected are extracted therefrom. Furthermore, intensity feature information can be extracted from the intensity change data, frequency feature information can be extracted from the frequency change data, and chromatographic feature information can be extracted from the chromatographic change data. Finally, based on the intensity feature information, frequency feature information, and chromatographic feature information, the signal type of the signal to be detected can be detected to obtain a type detection result.

[0030] Pulsar signals and radio frequency interference signals have significant differences in intensity change data, frequency change data, and chromatographic change data. Feature extraction is performed on these three types of change data respectively, and then signal type detection is performed based on the three types of feature information extracted, which can improve the accuracy of signal type detection. In addition, this method automatically extracts change data and features, and when facing a large amount of observation data, it can detect signal types more quickly, improving the detection efficiency of signal types.

[0031] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit this application. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The drawings herein are incorporated into the specification and constitute a part of this specification, showing embodiments consistent with this application, and are used together with the specification to explain the principles of this application. Obviously, the drawings in the following description are only some embodiments of this application, and those of ordinary skill in the art can obtain other drawings based on these drawings without creative efforts. In the drawings:

[0033] Figure 1 is a signal detection system architecture related to this application;

[0034] Figure 2It is a flowchart of a signal detection method shown according to an exemplary embodiment;

[0035] Figure 3 It is a schematic diagram showing the differences in the intensity, frequency, and chromatogram changes between pulsar signals and radio frequency interference signals involved in the present application;

[0036] Figure 4 It is a flowchart of a signal detection method shown according to another exemplary embodiment;

[0037] Figure 5 It is a schematic diagram of the sampling process corresponding to the intensity change data involved in the present application;

[0038] Figure 6 It is a schematic diagram of the sampling process corresponding to the frequency change data involved in the present application;

[0039] Figure 7 It is a flowchart of a signal detection method shown according to another exemplary embodiment;

[0040] Figure 8 It is a block diagram of a signal detection device shown according to an exemplary embodiment;

[0041] Figure 9 It is a schematic diagram of the structure of a computer system of a signal detection device shown according to an exemplary embodiment. Detailed implementation

[0042] Here, the exemplary embodiments will be described in detail, and the examples are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all the embodiments of the present application. On the contrary, they are only examples of the devices and methods that are the same as some aspects of the present application as detailed in the appended claims.

[0043] The block diagrams shown in the drawings are only functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software form, or implemented in at least one hardware module or integrated circuit, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0044] The flowcharts shown in the drawings are only exemplary descriptions and do not necessarily include all the contents and operations / steps, nor do they necessarily need to be executed in the described order. For example, some operations / steps can be decomposed, while some operations / steps can be integrated or partially integrated, so the actual execution order may change according to the actual situation.

[0045] It should be noted that in this application, "a plurality of" means two or more. "And / or" describes the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. The character " / " generally represents an "or" relationship between the associated objects before and after.

[0046] Before introducing the technical solutions of the embodiments of this application, the technical terms related to the embodiments of this application will be introduced here.

[0047] The Five-hundred-meter Aperture Spherical radio Telescope (FAST) is currently the largest and most sensitive radio telescope in the world. It is located in a karst depression in Guizhou Province, China, with a diameter of 500 meters. FAST can receive radio signals from deep space and is used for astronomical observations. It is particularly significant in aspects such as radio pulsar detection, interstellar medium research, and the exploration of extraterrestrial life. With its extremely high sensitivity and wide observation range, FAST can detect very weak radio signals, helping scientists explore various celestial bodies in the universe, including pulsars, quasars, interstellar hydrogen, etc. Due to its unique design, FAST can also provide high-resolution data and has very strong signal processing capabilities, making it play a crucial role in pulsar research and cosmic exploration.

[0048] A pulsar is a rotating neutron star with an extremely high rotation speed. It usually emits electromagnetic radiation periodically within a fixed period. This radiation appears as very regular pulse signals, similar to an astronomical clock. Each rotation emits a short beam of radio waves or waves in other frequency bands, and the period usually ranges from a few milliseconds to several seconds. The characteristics of pulsar signals are high frequency, periodicity, and stability. They are generated by the strong magnetic field and rapid rotation of the neutron star. Due to the periodicity of these signals, pulsars have become important objects in astronomical research and can be used for astronomical measurements, physical experiments, and the detection of gravitational waves, etc.

[0049] Radio Frequency Interference (RFI) signals refer to non-astronomical signals from different devices or technologies in the radio frequency band used by humans, which will interfere with astronomical observations. These signals may come from radio communications, radars, satellites, and even household appliances, etc. Such signals usually have strong power and are usually non-periodic, random, and have a complex frequency distribution, so they are likely to mask pulsar signals and other astronomical signals. The existence of RFI poses a great challenge to the observations of radio astronomy, especially when conducting high-sensitivity observations.

[0050] Pulsar search tools are software tools used to analyze radio astronomy data and discover pulsar candidates from it. These tools are designed to identify signals that conform to the characteristics of pulsars through the processing and analysis of radio data. The Pulsar Exploration and Search Toolkit (PRESTO) is one of the widely used pulsar search tools at present, mainly used for preprocessing time series data, signal folding, and period detection. It helps astronomers screen out possible pulsar candidates from a large amount of observational data by analyzing the periodic changes in the signals. PRESTO can process the original observational data obtained from radio telescopes such as FAST and generate various visualization graphs such as time-frequency spectrograms and periodograms to further screen out candidate pulsars. And Your tool refers to some customized or newly developed pulsar search tools, which usually have specific optimization algorithms, machine learning functions, or efficient parallel computing capabilities, and can improve the detection accuracy and speed of pulsar signals. These pulsar search tools can not only detect pulsar signals with strong periodicity but also further eliminate radio frequency interference (RFI) signals by analyzing factors such as the dispersion and frequency characteristics of the signals, making the discovery of pulsars more accurate.

[0051] Dispersion Measure (DM) is the time delay effect caused by the different propagation speeds of electromagnetic waves with different frequencies when pulsar signals propagate through the interstellar medium. Specifically, when the radio waves emitted by a pulsar pass through the Milky Way or other interstellar media, the electromagnetic waves with lower frequencies propagate slower, while those with higher frequencies propagate faster. As a result, the different frequency components of the signal arrive at the receiving device at different times, forming a time delay phenomenon. The value of DM is calculated by measuring this time delay effect and represents the dispersion degree per unit frequency during signal propagation. The magnitude of DM is closely related to the distance between the pulsar and the Earth and the electron density of the medium it passes through. Therefore, DM can not only provide physical information about pulsars but also help astronomers infer the distance of pulsars and the characteristics of the interstellar medium.

[0052] Support Vector Machine (SVM) is a supervised learning model commonly used in classification and regression analysis, especially suitable for processing high-dimensional data. In classification problems, the goal of SVM is to find a hyperplane that separates different classes in the dataset and maximizes the margin of the classification boundary. In this way, SVM can effectively handle linear and non-linear problems, especially performing well when there are obvious intervals or boundaries between data points. The core idea of SVM is to find an optimal separation surface (hyperplane) such that the distance from the sample points on both sides of the separation surface to the hyperplane is maximized, which can enhance the generalization ability of the model and reduce overfitting. For non-linear problems, SVM maps the original data to a higher-dimensional space by using kernel functions (such as radial basis kernel function, linear kernel function, etc.), so that the linear classifier can work in this space. SVM is widely used in various fields, including text classification, image recognition, medical diagnosis, etc.

[0053] DenseNet is a deep neural network architecture, especially suitable for tasks such as image recognition. The main feature of DenseNet is that each layer in the network is directly connected to all the previous layers. The input of each layer includes not only the output of the previous layer but also the outputs of all the previous layers. This densely connected design effectively alleviates the problems of gradient disappearance and information loss in deep networks, promotes the flow of information and gradients in the network, and enables the network to be trained more easily and obtain better performance. The core advantage of DenseNet lies in its efficient information transmission mechanism. Especially when performing image processing and feature extraction, it can utilize rich feature information from each layer, thereby improving the performance and accuracy of the model. In image recognition tasks, DenseNet can effectively extract multi-scale information of images, capture complex feature patterns, and is suitable for tasks that require high-level abstraction and feature fusion.

[0054] In related technologies, usually, detection rules are constructed based on the frequency and pulse width of pulsar signals, and the feature matching degree of each signal with pulsar signals is detected to determine whether the signal is emitted by a pulsar. However, interference signals may also disguise themselves as signals with similar pulsar characteristics, resulting in inaccurate judgment results for pulsar signals.

[0055] Based on this, embodiments of the present application respectively propose a system installation method, a system installation device, a system installation equipment, a computer-readable storage medium, and a computer program product. In these embodiments, an observation map of a signal to be detected can be obtained, and intensity change data, frequency change data, and chromatographic change data of the signal to be detected can be extracted therefrom. Furthermore, intensity feature information can be extracted from the intensity change data, frequency feature information can be extracted from the frequency change data, and chromatographic feature information can be extracted from the chromatographic change data. Finally, based on the intensity feature information, frequency feature information, and chromatographic feature information, the signal type of the signal to be detected can be detected to obtain a type detection result. Pulsar signals and radio frequency interference signals have significant differences in intensity change data, frequency change data, and chromatographic change data. By respectively performing feature extraction on these three types of change data and then detecting the signal type based on the three types of feature information extracted, the accuracy of signal type detection can be improved. In addition, by automatically extracting change data and feature extraction, when facing a large amount of observation data, the signal type can be detected more quickly, improving the detection efficiency of the signal type.

[0056] Please refer to Figure 1 , Figure 1 which is a schematic diagram of a signal detection system architecture related to the present application.

[0057] Figure 1 The signal detection system architecture shown includes a data preprocessing module 110, intensity change data 120, frequency change data 130, chromatographic change data 140, a machine learning model 150, a deep learning network 160, a deep learning network 170, a feature fusion module 180, and a classification model 190.

[0058] Among them, in the preprocessing module 110, there are a data collection module 111, an anti-interference module 112, a dispersion cancellation module 113, and a single-pulse search module 114. After the single-pulse search by the single-pulse search module 114, an observation map 115 of multiple signals can be output.

[0059] Among them, the data collection module 111 can store the radio signals collected by the radio telescope terminal in the form of a voltage-time series, and these radio signals can be referred to as raw signals.

[0060] The anti-interference module 112 is used to eliminate or reduce some noise or interference in the raw signals and mitigate the influence of radio frequency interference on the search results. The anti-interference module 112 uses a variety of signal processing techniques, such as filtering, spectrum analysis, time-domain analysis, etc., to identify and remove interference signals. The core task of this module is to reduce or eliminate the influence of non-astronomical source signals on the observation data through precise algorithms and rules, thereby improving the accuracy of signal processing.

[0061] The dispersion cancellation module 113 cancels the influence of the frequency-related delay effect. When various signals pass through the interstellar medium, signals of different frequencies propagate at different speeds, causing signal dispersion, where low-frequency signals propagate slower than high-frequency signals. The dispersion cancellation module 113 calculates and cancels the dispersion effect to restore the true time-domain structure of the signal. This module uses the dispersion measure (DM) value to correct the propagation delay at each frequency point and adjust the signal timing so that each frequency component can arrive simultaneously, ensuring the precise alignment of the signal.

[0062] The single-pulse search module 114 is used to identify the original signals suspected to be pulsar signals from the preprocessed original signals and obtain the observation maps corresponding to these original signals.

[0063] Next, based on the observation maps of each original signal, the computer can extract the intensity change data 120, frequency change data 130, and chromatographic change data 140. The intensity change data can be presented in the form of an intensity change map, which can also be called a flux map, representing the change in signal intensity over time. The frequency change data can be presented in the form of a frequency change map, which can also be called a time-frequency map, characterizing the frequency distribution of the signal over time. The chromatographic change data can be presented in the form of a chromatographic change map, which can also be called a time-chromatogram, characterizing the chromatographic distribution of the signal over time.

[0064] The machine learning model 150 can perform feature extraction on the intensity change data 120 and output the intensity feature information 151. The machine information model 150 can be a support vector machine.

[0065] The deep learning network 160 can perform feature extraction on the frequency change data 130 and output the frequency feature information 164. The deep learning network 160 can adopt the DenseNet. In the deep learning network 160, there are a convolutional layer 161, a pooling layer 162, and a connection layer 163. After being processed by these three layers, the frequency change data 130 can be converted into the frequency feature information 164. Among them, the convolutional layer 161 can contain multiple convolutional layers, and the multiple convolutional layers are organized into several blocks. A transition layer can be set between every two blocks to reduce data complexity. The pooling layer 162 can be a global pooling layer, and the connection layer 163 can be a fully connected layer.

[0066] Similarly, the deep learning network 170 can perform feature extraction on the chromatographic change data 130 and output chromatographic feature information 174. The deep learning network 170 can adopt a DenseNet (Dense Convolutional Network). In the deep learning network 170, there are a convolutional layer 171, a pooling layer 172, and a connection layer 173. After being processed by these three layers, the chromatographic change data 140 can be converted into chromatographic feature information 174. Among them, the convolutional layer 171 can contain multiple convolutional layers, and the multiple convolutional layers are organized into several blocks. A transition layer can be set between every two blocks to reduce data complexity. The pooling layer 172 can be a global pooling layer, and the connection layer 173 can be a fully connected layer.

[0067] It should be noted that the deep learning network 160 and the deep learning network 170 can adopt the same model architecture, but they are two independent deep learning networks and can have different model parameters.

[0068] The feature fusion module 180 can perform fusion processing on the intensity feature information 151, the frequency feature information 164, and the chromatographic feature information 174 to generate fused feature information 181. The feature fusion module 180 can automatically learn and determine the correlation between these three types of feature information through an Attention mechanism and calculate the weight of each type of feature information. The Attention mechanism can effectively capture the importance of each feature information in the signal classification task and improve the effect of feature fusion.

[0069] In the feature fusion module 180, first, the intensity feature information 151, the frequency feature information 164, and the chromatographic feature information 174 can be respectively normalized to ensure that different feature information has the same scale. Then, a correlation matrix between them is calculated based on these feature information to measure the contribution degree of each feature information in the overall feature space. According to the correlation matrix, the weight of each type of feature information is further generated. Next, these weights are used to perform weighted fusion on the three types of feature information, and finally, the fused feature information 181 is generated.

[0070] The fused feature information 181 can effectively integrate the data characteristics from intensity changes, frequency changes, and chromatographic changes, make full use of the complementarity of different features, and thus enhance the ability to distinguish signal types.

[0071] The classification model 190 can classify the fused feature information 181 and output the classification result 191 of the signal type. The classification model 190 can be a deep learning network with a softmax output layer for making a final judgment on the type of the signal to be detected. The classification model 190 receives the fused feature information 181 and maps the fused features through a number of fully connected layers and activation function layers. Finally, the softmax output layer can calculate the probability distribution of the signal to be detected belonging to different categories.

[0072] The softmax output layer outputs two probability values: one is the probability that the signal to be detected is a pulsar signal, and the other is the probability that it is a radio frequency interference signal. If the probability of the pulsar signal is greater than the preset probability threshold, the classification model 190 will classify the signal as a pulsar signal; otherwise, it will be classified as a radio frequency interference signal.

[0073] It should be noted that the method of the embodiment of the present application can be applied to a terminal device, which can be a mobile phone, a tablet computer, a laptop computer, a handheld computer, a mobile internet device (MID), a vehicle-mounted device, an aircraft, a wearable device (such as a smart watch, a smart bracelet, a pedometer, etc.), a virtual reality device (such as a VR (Virtual Reality) device, an AR (Augmented Reality) device), and so on.

[0074] Optionally, the method of the embodiment of the present application can also be applied to a server, which can be an independent physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0075] In the specific implementation of the present application, it involves user-related data, such as the pinyin sequence, target text, and input process data input by the user. When the embodiment of the present application is applied to a specific product or technology, user permission or consent needs to be obtained, and the collection, use, and processing of relevant data need to comply with the relevant laws, regulations, and standards of relevant countries and regions.

[0076] Please refer to Figure 2 , Figure 2 is a flowchart of a signal detection method shown according to an exemplary embodiment. The method can be applicable to Figure 2The described implementation environment is executed specifically by a computer. Of course, this method can also be applied to other implementation environments, and the execution subject of this method is not restricted here.

[0077] The signal detection method will be elaborated in detail below with a computer as an exemplary execution subject. As Figure 2 shown, in an exemplary embodiment, the method at least includes the following steps:

[0078] S210. Obtain the observation graph of the signal to be detected.

[0079] The radio telescope generates original time series signal data by receiving and recording signals from celestial bodies. This data usually contains multiple parameters such as intensity and frequency, and these parameters may change over time, reflecting the physical characteristics of different celestial bodies. Due to the huge amount of observation data and the inclusion of various astronomical and non-astronomical noises, directly using the original observation data for analysis may lead to incorrect classification results or missed detections. Therefore, it is necessary to preprocess these original observation data. Among them, the original observation data can also be called FAST data, that is, the observation data collected by FAST.

[0080] The computer can preprocess the original observation data through the PRESTO tool or the Heimdall tool. The preprocessing includes using a data collection module to store the original observation data in the form of voltage time series to obtain various original signals. Then, a deinterference module and a dedispersion module are used to preprocess the original signal data. Deinterference is to eliminate or reduce the influence of radio frequency interference on the search results. Dedispersion is to eliminate the influence of the frequency-related delay effect. After these processes, a cleaner signal can be obtained, and the signal at this time can be called the preprocessed original signal.

[0081] Next, the computer can use the Your tool to perform single-pulse search on the preprocessed original signal. The single-pulse search can identify signals suspected of being pulsar signals from the preprocessed original signal and output the corresponding characteristic data, and this characteristic data can be converted into the corresponding observation graph. And the celestial bodies corresponding to these suspected pulsar signals can be called pulsar candidates.

[0082] The signal to be detected can be any one of the signals suspected of being pulsar signals obtained by single-pulse search. The computer can obtain the observation graph of the signal to be detected for subsequent further analysis.

[0083] S220. Extract the intensity change data, frequency change data, and chromatographic change data of the signal to be detected from the observation graph.

[0084] The computer can extract intensity change data, frequency change data, and chromatogram change data from the observation graph of the signal to be detected.

[0085] The intensity change data can be presented in the form of an intensity change graph, which can also be called a flux graph, representing the change of signal intensity over time. The frequency change data can be presented in the form of a frequency change graph, which can also be called a time-frequency graph, representing the frequency distribution of the signal over time. The chromatogram change data can be presented in the form of a chromatogram change graph, which can also be called a time-chromatogram, representing the chromatogram distribution of the signal over time.

[0086] The change data can be presented in the form of a change graph. The change data can be converted into a graphical form using the aforementioned Your tool to obtain a change graph.

[0087] like Figure 3 Shown is a schematic diagram showing the difference between the intensity, frequency and chromatogram change diagram of a pulsar signal and a radio frequency interference signal involved in the present application.

[0088] exist Figure 3 The top two graphs show the intensity variation of pulsar signals and radio frequency interference signals, which can show the intensity burst of the signal within a specific time. For pulsar signals, there is a spike on the graph, while radio frequency interference signals do not have this feature.

[0089] The two middle pictures are frequency change diagrams, which have been processed by de-dispersion. After de-dispersion processing, the pulsar signal will appear as an obvious vertical line on this graph, while the radio frequency interference signal does not have this feature.

[0090] The two bottom pictures are chromatogram change diagrams. The chromatogram change diagrams of pulsar signals will show a unique "butterfly shape", while the RF interference signal does not have this feature.

[0091] S230: extract intensity characteristic information from the intensity change data, extract frequency characteristic information from the frequency change data, and extract chromatogram characteristic information from the chromatogram change data.

[0092] In the signal detection method, it is first necessary to extract feature information from the intensity change data, frequency change data, and chromatographic change data of the signal to be detected. These data are usually continuous time series signals during the acquisition process. Therefore, in order to facilitate subsequent feature extraction and model processing, uniform sampling is required first to convert these time series data into data representations with a fixed size.

[0093] For the intensity change data, it is a one-dimensional time series signal that records the change of signal intensity over time. To extract features from the intensity change data, it is first necessary to sample it according to a preset sampling strategy. In this embodiment, the preset sampling strategy can use 64 sampling points, that is, uniformly sample the intensity change data within a specific time range.

[0094] From Figure 3 It can be seen that the time range of the intensity change data is from -600ms to 600ms. Then, 64 data points will be uniformly sampled within this time interval. Specifically, starting from -600ms, sampling is performed at uniform time intervals until 600ms ends. Through this process, the original intensity change data is converted into data of 64 sampling points, that is, intensity sampling data. These sampling data, as inputs, will be fed into the intensity feature detection model for feature extraction.

[0095] The frequency change data is usually two-dimensional data representing the change of signal frequency over time. In this embodiment, the preset sampling strategy can use a size of 64×64 to sample the frequency change data. Specifically, first, uniform sampling is performed within the frequency range, from Figure 3 It can be seen that the frequency range is from 1450MHz to 1150MHz. Then, 64 equally spaced samples are taken for the ordinate (frequency axis) within this frequency range. Each sampling point represents the intensity or energy of the signal at that time point and at that frequency.

[0096] At the same time, on the time axis, uniform sampling is still used to uniformly sample the frequency values corresponding to each moment within the time range, forming 64 time points. In this way, the frequency change data obtained through uniform sampling will be converted into a two-dimensional data matrix of 64×64, that is, frequency sampling data. These sampling data, as inputs, will be fed into the frequency feature detection model for feature extraction.

[0097] The chromatogram change data also belongs to two-dimensional data, which characterizes the change of signal chromatogram over time. Similarly, the chromatogram change data needs to be subjected to two-dimensional sampling processing similar to the frequency change data. For the chromatogram change data, the preset sampling strategy also uses a size of 64×64 for uniform sampling.

[0098] Uniform sampling is performed on the ordinate (chromatogram axis) of the chromatogram change data within a certain range. Assuming this range is from 0 pc / cm3 to 1400 pc / cm3 of the chromatogram intensity value, then 64 equally spaced samples are taken within this range; on the time axis, 64 sampling points are still taken at uniform time intervals. After this process, the chromatogram change data will be converted into a two-dimensional data matrix of 64×64, that is, chromatogram sampling data. These sampling data will also be input into the chromatogram feature detection model for feature extraction.

[0099] After the above sampling process, the intensity change data, frequency change data, and chromatographic change data are respectively converted into intensity sampling data, frequency sampling data, and chromatographic sampling data. Next, these sampling data will be respectively input into the corresponding feature detection models for feature extraction. The intensity sampling data will be input into the intensity feature detection model to extract intensity feature information therefrom; the frequency sampling data will be input into the frequency feature detection model to extract frequency feature information therefrom; the chromatographic sampling data will be input into the chromatographic feature detection model to extract chromatographic feature information therefrom.

[0100] Through these feature extraction processes, the intensity feature information, frequency feature information, and chromatographic feature information can provide necessary data support for subsequent signal type detection. The extracted feature information will effectively characterize the physical properties of the signal to be detected and provide a reliable basis for signal classification and judgment.

[0101] S240. Detect the signal type of the signal to be detected based on the intensity feature information, frequency feature information, and chromatographic feature information to obtain a type detection result.

[0102] During the signal detection process, after extracting the intensity feature information, frequency feature information, and chromatographic feature information of the signal to be detected through the foregoing steps, next, these feature information need to be processed to accurately detect the signal type. The computer can input the intensity feature information, frequency feature information, and chromatographic feature information into the attention mechanism.

[0103] First of all, the attention mechanism needs to detect the correlation between the intensity feature information, frequency feature information, and chromatographic feature information. In this way, the mutual relationship between these three feature information in the signal and their contribution degrees to signal type judgment can be understood. Specifically, through correlation analysis, the mutual dependence and potential connections between the three feature information can be identified. Based on the correlation analysis, the weight of each feature information can be calculated based on the degree of correlation between each feature information and the signal type. These weights represent the importance of each feature information in signal type determination.

[0104] After determining the weight of each feature information, next, the intensity feature information, frequency feature information, and chromatographic feature information can be feature fused according to the calculated weights. When feature fusing, weighted fusion can be adopted. The process of weighted fusion synthesizes each feature information according to its importance in signal type detection, thereby obtaining a fused feature information. That is, the computer can multiply the intensity feature information by its corresponding weight, add the product of the frequency feature information and its corresponding weight, and then add the product of the chromatographic feature information and its corresponding weight to obtain the fused feature information.

[0105] The fused feature information contains the information of all three types of features, and the most valuable part for signal type determination is highlighted through weighted fusion.

[0106] Finally, the computer can input the fused feature information into a classification model to detect the signal type of the signal to be detected. Through the signal type detection model, the fused feature information will be used as input to determine the type of the signal to be detected. According to the result output by the classification model, the type of the signal to be detected can be determined, such as determining whether it is a pulsar signal or a radio frequency interference signal. Finally, based on the judgment result of the classification model, the type detection result of the signal to be detected is obtained.

[0107] Through this method, the computer can obtain the observation graph of the signal to be detected, extract the intensity change data, frequency change data, and chromatographic change data of the signal to be detected from it. Furthermore, the intensity feature information can be extracted from the intensity change data, the frequency feature information can be extracted from the frequency change data, and the chromatographic feature information can be extracted from the chromatographic change data. Finally, based on the intensity feature information, frequency feature information, and chromatographic feature information, the signal type of the signal to be detected can be detected to obtain the type detection result.

[0108] There are significant differences between pulsar signals and radio frequency interference signals in terms of intensity change data, frequency change data, and chromatographic change data. By separately extracting features from these three types of change data and then detecting the signal type based on the three types of extracted feature information, the accuracy of signal type detection can be improved. In addition, through the automatic extraction of change data and feature extraction, this method can detect the signal type more quickly when facing a large amount of observation data, improving the detection efficiency of the signal type.

[0109] In an embodiment of the present application, another signal detection method is provided, and this signal detection method can be executed by a computer. As Figure 4 shown, this signal detection method may include S210 to S220, S410 to S420, and S240. That is, S410 to S420 are Figure 2 the specific implementation methods of S230 shown.

[0110] The following describes S410 to S420:

[0111] S410: Sample the intensity change data to obtain intensity sampling data, sample the frequency change data to obtain frequency sampling data, and sample the chromatographic change data to obtain chromatographic sampling data.

[0112] Specifically, S410 may include S411 to S413, which are described below:

[0113] S411. Based on a preset sampling strategy, collect the signal intensity values corresponding to each sampling point from the intensity change data to obtain intensity sampling data; the preset sampling strategy includes a sampling strategy for setting multiple time points and a sampling strategy for setting the number of samples.

[0114] First, the computer needs to sample the intensity change data. The intensity change data usually reflects the process of signal intensity changing over time. To ensure that the sampled data can accurately represent the time-domain characteristics of the intensity change data, the sampling needs to follow a preset sampling strategy.

[0115] The preset sampling strategy lies in two aspects: one is to determine the sampling time points, and the other is to determine the number of samples. In terms of time point setting, the sampling strategy stipulates multiple time points evenly distributed within the time range, so as to evenly select the corresponding signal intensity values from the intensity change data. Assuming that the time range of the signal data is from -600ms to 600ms, the preset sampling strategy divides these time periods into 64 time points, thereby obtaining 64 corresponding intensity values.

[0116] Exemplarily, as Figure 5 shown is a schematic diagram of the sampling process corresponding to the intensity change data involved in the present application. As Figure 5 can be seen, the intensity sampling data can reflect the approximate change of the intensity change data during -600ms to 600ms, evenly sampling the signal intensity values within this period to obtain the following intensity sampling data. The intensity sampling data contains 64 sampling values, and each sampling value corresponds to a serial number.

[0117] Optionally, the computer can convert the intensity change data into an intensity change graph and then sample in the intensity change graph to obtain intensity sampling data.

[0118] It should be noted that the number of samples can be set by those skilled in the art. In addition to being 64, it can also be 128, 256, etc., which are specifically determined according to different scenario requirements. The embodiments of the present application do not know.

[0119] S412. Based on a preset sampling strategy, collect the signal frequency values corresponding to each sampling point from the frequency change data to obtain frequency sampling data.

[0120] For the frequency change data, the sampling process is similar to that of the intensity change data. The frequency change data reflects the change of the signal frequency over time.

[0121] Different from the intensity data, the frequency change data is usually two-dimensional data, including two dimensions of time and frequency. In this sampling step, the preset sampling strategy will define the distribution of sampling in two dimensions: the time dimension and the frequency dimension. In the time dimension, the sampling strategy will also define multiple uniform time points to ensure the uniformity of sampling; in the frequency dimension, the sampling strategy will determine the distribution of frequency values according to the range of frequency changes and the characteristics of the signal.

[0122] Exemplarily, as Figure 6 shown is a schematic diagram of the sampling process corresponding to the frequency change data involved in this application. The frequency range of the frequency change data is from 1450 MHz to 1150 MHz. The preset sampling strategy will uniformly select 64 frequency values within this frequency range and match them one by one with the corresponding time points, thereby forming a 64×64 sampling matrix. The value of each sampling point represents the signal intensity or energy at that time point and that frequency, and thus frequency sampling data can be obtained.

[0123] S413. Based on the preset sampling strategy, collect the signal chromatogram values corresponding to each sampling point from the chromatogram change data to obtain chromatogram sampling data.

[0124] The chromatogram change data records the changes of the signal in two dimensions of time and frequency. It usually appears as a chromatogram, showing the frequency distribution and intensity distribution of the signal within a specific time window.

[0125] Similarly, the computer will use the preset sampling strategy to sample the chromatogram change data. The chromatogram change data is two-dimensional, so the sampling strategy will cover two dimensions of time and dispersion amount. In the time dimension, the sampling strategy will determine the sampling time points; in the frequency dimension, the sampling strategy will ensure that dispersion values are uniformly selected within the entire dispersion amount range.

[0126] For example, the frequency range of the chromatogram change data can be from 0 pc / cm3 to 1400 pc / cm3. The time range is from -600 ms to 600 ms. During sampling, the preset strategy will evenly distribute 64 sampling points in both the time and frequency dimensions, obtaining a 64×64 sampling matrix. Each sampling point will correspond to a chromatogram value, which represents the intensity or other relevant value of the signal at a specific time point and dispersion point.

[0127] S420. Extract feature information of intensity from the intensity sampling data, extract feature information of frequency from the frequency sampling data, and extract feature information of chromatogram from the chromatogram sampling data.

[0128] Specifically, those skilled in the art can set up an intensity feature detection model dedicated to extracting features from intensity sampling data; in addition, a frequency feature detection model can be set up, dedicated to extracting features from frequency sampling data; and a chromatographic feature detection model can also be set up, dedicated to extracting features from chromatographic sampling data.

[0129] Therefore, the computer can input the intensity sampling data into the intensity feature detection model for feature extraction to obtain intensity feature information; input the frequency sampling data into the frequency feature detection model for feature extraction to obtain frequency feature information; and input the chromatographic sampling data into the chromatographic feature detection model for feature extraction to obtain chromatographic feature information.

[0130] Through this method, intensity change data, frequency change data, and chromatographic change data are sampled to obtain corresponding sampling data, and then feature extraction is performed through a feature detection model that matches the sampling data, which can improve the accuracy of feature information extraction, thereby providing accurate and reliable data for subsequent signal type detection and improving the accuracy of signal type detection.

[0131] In an embodiment of the present application, another signal detection method is provided, and this signal detection method can be executed by a computer. As Figure 7 shown in the training process of the detection model, this training process may include S710 to S730.

[0132] The following describes S710 to S730:

[0133] S710, Obtain a plurality of signal samples and an observation map sample corresponding to each signal sample, and the signal samples include pulsar signal samples and radio frequency interference signal samples.

[0134] The computer needs to obtain signal samples for training the initial model. Signal samples usually come from actual observation data and include two main types of signals: pulsar signal samples and radio frequency interference signal samples. Pulsar signal samples refer to signals collected from astronomical observations that have periodic characteristics and are usually used to study pulsars in astrophysics. Radio frequency interference signal samples refer to signal interference caused by human factors (such as communication devices, satellite signals, etc.), which are usually non-astronomical signals but may be confused with pulsar signals in the observation data.

[0135] Each signal sample corresponds to an observation map sample, and the observation map sample refers to the visual data of the signal observed by the radio telescope in dimensions such as time, frequency, and chromatogram. The computer can obtain a plurality of signal samples containing different signal types and their corresponding observation map samples, providing sufficient sample data for subsequent model training.

[0136] S720. Extract the intensity change data sample, frequency change data sample, and chromatographic change data sample of the signal to be detected from the observed graph samples of each signal sample.

[0137] The computer needs to extract the intensity change data sample, frequency change data sample, and chromatographic change data sample of the signal sample from the observed graph sample of each signal sample. The signal sample may be a pulsar signal or a radio frequency interference signal. The computer needs to extract the corresponding intensity change data sample, frequency change data sample, and chromatographic change data sample for both types of signal samples.

[0138] S730. Train the first initial model based on the intensity change data sample of each signal sample to obtain an intensity feature detection model, train the second initial model based on the frequency change data sample of each signal sample to obtain a frequency feature detection model, and train the third initial model based on the chromatographic change data sample of each signal sample to obtain a chromatographic feature detection model.

[0139] In an embodiment of the present application, the computer needs to sample from the change data sample, and the change data sample is any one of the intensity change data sample, frequency change data sample, and chromatographic change data sample. That is to say, the computer needs to sample the intensity change data sample, frequency change data sample, and chromatographic change data sample. The specific sampling process is the same as the sampling process corresponding to the signal to be detected described above, and will not be elaborated here.

[0140] In this way, after sampling, the intensity sampling data sample, frequency sampling data sample, and chromatographic sampling data sample can be obtained. After processing the change data samples of multiple signal samples by sampling, three different data sets can be obtained, namely, the intensity sampling sample set, frequency sampling sample set, and chromatographic sampling sample set.

[0141] Next, input the intensity sampling data sample into the first initial model for training to obtain an intensity feature detection model; input the frequency sampling data sample into the second initial model for training to obtain a frequency feature detection model; input the chromatographic sampling data sample into the third initial model for training to obtain a chromatographic feature detection model.

[0142] Specifically, the computer can input the sampled data samples into an initial model that matches the sampled data samples, so that the initial model extracts features from the sampled data samples to obtain predicted feature information. The initial model can be any one of a first initial model, a second initial model, and a third initial model. Next, the computer constructs a loss based on the actual feature information and the predicted feature information corresponding to the sampled data samples, and trains the initial model based on this loss. During the training process, the computer can continuously adjust the model parameters of the initial model until the number of training iterations reaches a preset number, or until the initial model converges, to obtain the corresponding feature detection model. Among them, the preset number can be set by those skilled in the art. For example, it can be 20,000 times. The more times, the stronger the feature detection ability of the model after training.

[0143] It should be noted that any one of the first initial model, the second initial model, and the third initial model can be trained in the above manner. The three models are independent of each other and there is no data interaction.

[0144] Through this method, the computer can separately train the models corresponding to each type of variation data sample to obtain a feature detection model that matches the variation data sample. In this way, the accuracy of the feature information extracted by each feature detection model can be improved, providing an accurate data basis for subsequent signal type judgment.

[0145] Figure 8 It is a block diagram of a signal detection device shown in an embodiment of the present application. As Figure 8 shown, the signal detection device can be applied to a computer. The device includes:

[0146] A signal detection device includes:

[0147] An acquisition unit 810, configured to acquire an observation diagram of a signal to be detected;

[0148] A processing unit 820, configured to extract intensity change data, frequency change data, and chromatographic change data of the signal to be detected from the observation diagram;

[0149] The processing unit 820 is further configured to extract intensity feature information from the intensity change data, frequency feature information from the frequency change data, and chromatographic feature information from the chromatographic change data;

[0150] A detection unit 830, configured to detect the signal type of the signal to be detected based on the intensity feature information, the frequency feature information, and the chromatographic feature information, to obtain a type detection result.

[0151] In one embodiment of the present application, based on the foregoing solution, the detection unit 830 is further configured to detect the correlation between the intensity feature information, the frequency feature information, and the chromatographic feature information; the processing unit 820 is further configured to perform feature fusion on the intensity feature information, the frequency feature information, and the chromatographic feature information based on the correlation to obtain the fused feature information; the detection unit 830 is further configured to detect the signal type of the signal to be detected based on the fused feature information to obtain a type detection result.

[0152] In one embodiment of the present application, based on the foregoing solution, the acquisition unit 810 is further configured to obtain the weights corresponding to the intensity feature information, the frequency feature information, and the chromatographic feature information respectively based on the correlation; the processing unit 820 is further configured to perform weighted fusion on the intensity feature information, the frequency feature information, and the chromatographic feature information based on the weights to obtain the fused feature information.

[0153] In one embodiment of the present application, based on the foregoing solution, the detection unit 830 is further configured to detect the probability that the signal to be detected is a pulsar signal based on the fused feature information; if the probability is greater than a preset probability threshold, the signal type of the signal to be detected is determined to be a pulsar signal type; if the probability is less than the preset probability threshold, the signal type of the signal to be detected is determined to be a radio frequency interference signal type.

[0154] In one embodiment of the present application, based on the foregoing solution, the processing unit 820 is further configured to sample the intensity change data to obtain intensity sampling data, sample the frequency change data to obtain frequency sampling data, and sample the chromatographic change data to obtain chromatographic sampling data; extract features from the intensity sampling data to obtain intensity feature information, extract features from the frequency sampling data to obtain frequency feature information, and extract features from the chromatographic sampling data to obtain chromatographic feature information.

[0155] In one embodiment of the present application, based on the foregoing solution, the processing unit 820 is further configured to collect the signal intensity value corresponding to each sampling point from the intensity change data based on a preset sampling strategy to obtain intensity sampling data; the preset sampling strategy includes a sampling strategy for setting multiple time points and a sampling strategy for setting the sampling quantity; collect the signal frequency value corresponding to each sampling point from the frequency change data based on the preset sampling strategy to obtain frequency sampling data; and collect the signal chromatographic value corresponding to each sampling point from the chromatographic change data based on the preset sampling strategy to obtain chromatographic sampling data.

[0156] In one embodiment of the present application, based on the foregoing solution, the processing unit 820 is further configured to input the intensity sampling data into an intensity feature detection model for feature extraction to obtain intensity feature information; input the frequency sampling data into a frequency feature detection model for feature extraction to obtain frequency feature information; and input the chromatographic sampling data into a chromatographic feature detection model for feature extraction to obtain chromatographic feature information.

[0157] In one embodiment of the present application, based on the foregoing solution, the acquisition unit 810 is further configured to acquire a plurality of signal samples and an observation map sample corresponding to each signal sample, where the signal samples include pulsar signal samples and radio frequency interference signal samples; the processing unit 820 is further configured to extract an intensity change data sample, a frequency change data sample, and a chromatographic change data sample of the signal sample from the observation map sample of each signal sample; train a first initial model based on the intensity change data sample of each signal sample to obtain an intensity feature detection model, train a second initial model based on the frequency change data sample of each signal sample to obtain a frequency feature detection model, and train a third initial model based on the chromatographic change data sample of each signal sample to obtain a chromatographic feature detection model.

[0158] In one embodiment of the present application, based on the foregoing solution, the processing unit 820 is further configured to sample from the change data sample to obtain a sampling data sample, where the change data sample is any one of the intensity change data sample, the frequency change data sample, and the chromatographic change data sample; input the sampling data sample into an initial model matching the sampling data sample, so that the initial model performs feature extraction on the sampling data sample to obtain predicted feature information; the initial model is any one of the first initial model, the second initial model, and the third initial model; construct a loss based on the actual feature information and the predicted feature information corresponding to the sampling data sample; the device further includes a training unit configured to train the initial model based on the loss.

[0159] It should be noted that the device provided in the foregoing embodiment and the method provided in the foregoing embodiment belong to the same concept, and the specific manners in which each module and unit perform operations have been described in detail in the method embodiment.

[0160] An embodiment of the present application further provides a signal detection device, including: at least one processor; a memory configured to store at least one program, and when the at least one program is executed by the at least one processor, enable the electronic device to implement the signal detection method as described above.

[0161] Figure 9 It is a schematic structural diagram of a computer system suitable for implementing the signal detection device of the embodiment of the present application.

[0162] It should be noted thatFigure 9 The computer system 900 of the illustrated electronic device is merely an example and should not impose any limitations on the functions and scope of use of the embodiments of the present application.

[0163] As Figure 9 shown, the computer system 900 includes a central processing unit (CPU) 901, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 902 or the program loaded from the storage section 908 into the random access memory (RAM) 903, such as executing the methods in the above embodiments. In the RAM 903, various programs and data required for system operation are also stored. The CPU 901, ROM 902, and RAM 903 are connected to each other via a bus 904. An input / output (I / O) interface 905 is also connected to the bus 904.

[0164] The following components are connected to the I / O interface 905: an input section 906 including a keyboard, a mouse, etc.; an output section 907 including, for example, a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 908 including a hard disk, etc.; and a communication section 909 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 909 performs communication processing via a network such as the Internet. A drive 910 is also connected to the I / O interface 905 as required. A removable medium 911, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 910 as required so that a computer program read from it can be installed into the storage section 908 as required.

[0165] Specifically, according to the embodiments of the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments of the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication section 909 and / or installed from the removable medium 911. When the computer program is executed by the central processing unit (CPU) 901, various functions defined in the system of the present application are executed.

[0166] It should be noted that the computer-readable medium shown in the embodiments of the present application can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. The computer-readable medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the computer-readable medium may include, but are not limited to: an electrical connection having at least one wire, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, the computer-readable medium can be any tangible medium that contains or stores a program, and this program can be used by or in conjunction with an instruction execution system, apparatus, or device. In the present application, the computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable computer program. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer-readable signal medium can also be any computer-readable medium other than the computer-readable storage medium, and this computer-readable medium can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium can be transmitted using any appropriate medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0167] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present application. Among them, each block in the flowchart or block diagram can represent a module, a program segment, or a part of code, and the above module, program segment, or part of code contains at least one executable instruction for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in an order different from that marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram or flowchart, as well as the combination of blocks in the block diagram or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0168] The units involved in the embodiments described in this application can be implemented in software or in hardware, and the described units can also be provided in a processor. Among them, the names of these units do not constitute a limitation on the units themselves in some cases.

[0169] Another aspect of this application also provides a computer-readable medium, on which a computer program is stored. When the computer program is executed by a processor, the signal detection method as described above is implemented. The computer-readable medium can be included in the electronic device described in the above embodiments, or can exist alone without being assembled into the electronic device.

[0170] Another aspect of this application also provides a computer program product or a computer program. The computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable medium. The processor of the computer device reads the computer instructions from the computer-readable medium, and the processor executes the computer instructions, so that the computer device executes the signal detection methods provided in the above various embodiments.

[0171] The above content is only a preferred exemplary embodiment of this application and is not used to limit the implementation of this application. Those of ordinary skill in the art can easily make corresponding adaptations or modifications according to the main concept and spirit of this application. Therefore, the protection scope of this application should be subject to the protection scope required by the claims.

Claims

1. A signal detection method, characterized in that: include: Obtaining an observation graph of a signal to be detected; Extracting intensity change data, frequency change data, and chromatogram change data of the signal to be detected from the observation graph; Extracting intensity characteristic information from the intensity change data, extracting frequency characteristic information from the frequency change data, and extracting chromatogram characteristic information from the chromatogram change data; Based on the intensity characteristic information, the frequency characteristic information, and the chromatogram characteristic information, the signal type of the signal to be detected is detected to obtain a type detection result.

2. The method according to claim 1, characterized in that The detecting the signal type of the signal to be detected based on the intensity characteristic information, the frequency characteristic information, and the chromatogram characteristic information to obtain a type detection result includes: Detecting the correlation between the intensity characteristic information, the frequency characteristic information, and the chromatogram characteristic information; Based on the correlation, feature fusion is performed on the intensity feature information, the frequency feature information, and the chromatogram feature information to obtain fused feature information; The signal type of the signal to be detected is detected based on the fused feature information to obtain the type detection result.

3. The method according to claim 2, characterized in that The step of performing feature fusion on the intensity feature information, the frequency feature information, and the chromatogram feature information based on the correlation to obtain fused feature information includes: Based on the correlation, obtaining weights corresponding to the intensity characteristic information, the frequency characteristic information, and the chromatogram characteristic information respectively; The intensity feature information, the frequency feature information, and the chromatogram feature information are weighted and fused based on the weight to obtain the fused feature information.

4. The method according to claim 2, characterized in that: The detecting the signal type of the signal to be detected based on the fused feature information to obtain the type detection result includes: Detecting the probability that the signal to be detected is a pulsar signal based on the fused characteristic information; If the probability is greater than a preset probability threshold, determining the signal type of the signal to be detected as a pulsar signal type; If the probability is less than the preset probability threshold, the signal type of the signal to be detected is determined to be a radio frequency interference signal type.

5. The method according to claim 1, characterized in that The extracting of intensity characteristic information from the intensity change data, extracting of frequency characteristic information from the frequency change data, and extracting of chromatogram characteristic information from the chromatogram change data include: Sampling the intensity change data to obtain intensity sampling data, sampling the frequency change data to obtain frequency sampling data, and sampling the chromatogram change data to obtain chromatogram sampling data; The intensity characteristic information is obtained by performing feature extraction on the intensity sampling data, the frequency characteristic information is obtained by performing feature extraction on the frequency sampling data, and the chromatogram characteristic information is obtained by performing feature extraction on the chromatogram sampling data.

6. The method according to claim 5, characterized in that The sampling of the intensity change data to obtain intensity sampling data, the sampling of the frequency change data to obtain frequency sampling data, and the sampling of the chromatogram change data to obtain chromatogram sampling data include: Based on a preset sampling strategy, the signal strength value corresponding to each sampling point is collected from the intensity change data to obtain intensity sampling data; the preset sampling strategy includes a sampling strategy for setting a plurality of time points and a sampling strategy for setting a sampling number; Based on the preset sampling strategy, the signal frequency value corresponding to each sampling point is collected from the frequency change data to obtain frequency sampling data; and Based on a preset sampling strategy, the signal chromatogram value corresponding to each sampling point is collected from the chromatogram change data to obtain chromatogram sampling data.

7. The method according to claim 5, characterized in that The extracting features of the intensity sampling data to obtain the intensity characteristic information, extracting features of the frequency sampling data to obtain the frequency characteristic information, and extracting features of the chromatogram sampling data to obtain the chromatogram characteristic information include: Inputting the intensity sampling data into an intensity feature detection model to perform feature extraction to obtain the intensity feature information; Inputting the frequency sampling data into a frequency feature detection model to extract features and obtain frequency feature information; The chromatogram sampling data is input into a chromatogram feature detection model for feature extraction to obtain chromatogram feature information.

8. The method according to claim 1, characterized in that Extracting intensity characteristic information from the intensity change data, extracting frequency characteristic information from the frequency change data, and extracting chromatogram characteristic information from the chromatogram change data through a detection model; The training process of the detection model includes: Acquire a plurality of signal samples and an observation graph sample corresponding to each signal sample, wherein the signal samples include a pulsar signal sample and a radio frequency interference signal sample; Extracting an intensity change data sample, a frequency change data sample, and a color spectrum change data sample of the signal sample from the observation graph sample of each signal sample; A first initial model is trained based on the intensity change data samples of each signal sample to obtain an intensity feature detection model, a second initial model is trained based on the frequency change data samples of each signal sample to obtain a frequency feature detection model, and a third initial model is trained based on the chromatographic change data samples of each signal sample to obtain a chromatographic feature detection model.

9. The method according to claim 8, characterized in that The method further comprises: Sampling from the change data samples to obtain a sampled data sample, wherein the change data sample is any one of the intensity change data sample, the frequency change data sample, and the chromatogram change data sample; Inputting the sampled data sample into an initial model matching the sampled data sample, so that the initial model extracts features of the sampled data sample to obtain prediction feature information; the initial model is any one of the first initial model, the second initial model and the third initial model; Constructing a loss based on the actual feature information corresponding to the sampled data sample and the predicted feature information; The initial model is trained based on the loss.

10. A signal detection device, characterized in that: include: An acquisition unit, used for acquiring an observation graph of a signal to be detected; A processing unit, used for extracting the intensity change data, frequency change data, and chromatogram change data of the signal to be detected from the observation graph; The processing unit is further used to extract intensity characteristic information from the intensity change data, extract frequency characteristic information from the frequency change data, and extract chromatogram characteristic information from the chromatogram change data; The detection unit is used to detect the signal type of the signal to be detected based on the intensity characteristic information, the frequency characteristic information, and the chromatogram characteristic information to obtain a type detection result.

11. A signal detection device, characterized in that: include: a memory storing computer-readable instructions; A processor reads the computer readable instructions stored in the memory to execute the method according to any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that: Computer-readable instructions are stored thereon, and when the computer-readable instructions are executed by a processor of a computer, the computer is caused to execute the method according to any one of claims 1 to 9.

13. A computer program product comprising computer instructions, characterized in that: When the computer instructions are executed by a processor, the method according to any one of claims 1 to 9 is implemented.

Images