Learning model, signal processing device, flyer, and teacher data generation method

Abstract

An object such as a ground object can be detected using RAW data. The learning model learns using teacher data which has, as an input, a feature amount of a first synthetic aperture radar reception signal based on a reflected electromagnetic wave obtained by reflecting an electromagnetic wave irradiated on the learning region, and which has, as an output, first object information of an object in the learning region. The computer is enabled to realize a function of outputting second object information of the object in the detection area in response to a characteristic quantity of a second synthetic aperture radar reception signal based on the input of the characteristic quantity of the second synthetic aperture radar reception signal of a reflected electromagnetic wave in which the electromagnetic wave irradiated on the detection area is reflected.

Description

Technical Field

[0001] The present invention relates to a learning model, a signal processing device, a flying object and a program. Background Art

[0002] Observation of the Earth's surface, both on land and at sea, is widely performed using artificial satellites, aircraft, and drones. Satellite-based observation methods include those that acquire radar images, or SAR images, using synthetic aperture radar (SAR), and those that acquire and combine optical and SAR images. Patent Document 1 describes object detection using raw data (RAW data) obtained from SAR without SAR imaging.

[0003] Prior art literature

[0004] Patent Literature

[0005] Patent Document 1: Japanese Patent Application Laid-Open No. 2023-000897 Summary of the Invention

[0006] Various methods have been considered for detecting objects such as ground objects using RAW data. Therefore, an object of the present invention is to provide a learning model, a signal processing device, a flying object, and a method for generating teacher data that can detect objects such as ground objects using RAW data.

[0007] For a learning model involved in one embodiment of the present invention, teacher data is used for learning so that a computer can implement the following functions: output second object information of an object in a detection area in response to an input of a characteristic quantity of a second synthetic aperture radar reception signal based on a reflected electromagnetic wave obtained by reflecting an electromagnetic wave irradiating a detection area, wherein the teacher data is data that takes as input the characteristic quantity of a first synthetic aperture radar reception signal based on a reflected electromagnetic wave obtained by reflecting an electromagnetic wave irradiating a learning area and outputs the first object information of an object in a learning area.

[0008] According to this embodiment, since object information can be output based on the feature amount of the synthetic aperture radar reception signal, objects such as ground objects can be detected using RAW data.

[0009] In the above manner, the characteristic quantity of the first synthetic aperture radar received signal may include the frequency spectrum of the first synthetic aperture radar received signal, and the characteristic quantity of the second synthetic aperture radar received signal may include the frequency spectrum of the second synthetic aperture radar received signal.

[0010] In the above manner, the characteristic quantity of the first synthetic aperture radar received signal may include the signal strength of the first synthetic aperture radar received signal, and the characteristic quantity of the second synthetic aperture radar received signal may include the signal strength of the second synthetic aperture radar received signal.

[0011] In the above manner, the characteristic quantity of the first synthetic aperture radar received signal may include the phase of the first synthetic aperture radar received signal, and the characteristic quantity of the second synthetic aperture radar received signal may include the phase of the second synthetic aperture radar received signal.

[0012] According to these methods, object information can be output based on the spectrum, signal intensity, or phase as a feature quantity of the synthetic aperture radar reception signal, and thus objects such as ground objects can be detected using RAW data.

[0013] In the above embodiment, the first SAR received signal corresponds to a portion of the learning area whose feature quantity satisfies a predetermined condition. The feature quantity of the first SAR received signal may include a feature quantity based on an object in the learning area.

[0014] This method reduces the amount of data required for learning because the received signal, which serves as the basis for the received signal feature quantities used as teacher data for the learning model, corresponds to an area where the feature quantities satisfy specified conditions. Furthermore, since the received signal feature quantities include feature quantities based on objects within the learning area, it is possible to detect objects such as ground objects using RAW data while reducing the amount of data.

[0015] In the above embodiment, the second SAR reception signal is a signal corresponding to a portion of the detection area whose feature quantity satisfies a predetermined condition. The feature quantity of the second SAR reception signal may include a feature quantity based on an object in the detection area.

[0016] According to this method, the feature values ​​of the second synthetic aperture radar received signal used in inference using the learning model can also be based on signals corresponding to a portion of the area where the feature values ​​meet predetermined conditions, thereby reducing the amount of data required for inference. Furthermore, since the feature values ​​of the second synthetic aperture radar received signal include feature values ​​based on objects within the learning area, it is possible to detect objects such as ground objects using RAW data while reducing the amount of data.

[0017] In the above embodiment, the first object information includes information indicating the attributes of the object in the learning area, and the second object information includes information indicating the attributes of the object in the detection area.

[0018] According to this aspect, since information indicating the attributes of the object can be obtained, the object can be detected while obtaining more detailed information about the object.

[0019] An information processing device according to another embodiment of the present invention includes a storage unit storing a learning model, a signal acquisition unit acquiring a second synthetic aperture radar received signal, and an estimation unit inputting the second synthetic aperture radar received signal into the learning model to infer second object information.

[0020] Another embodiment of the present invention relates to a flying object comprising a storage unit storing a learning model, a signal acquisition unit for acquiring a second synthetic aperture radar received signal, an estimating unit for inputting the second synthetic aperture radar received signal into the learning model and inferring second object information, and a signal output unit for outputting an output signal based on the second object information to the outside.

[0021] Another embodiment of the present invention involves a teacher data generation method that includes causing a computer to perform the following operations: obtaining a first synthetic aperture radar received signal based on a reflected electromagnetic wave obtained by reflecting an electromagnetic wave irradiating a learning area; extracting a feature quantity of the learning signal corresponding to an area in the learning area where an object in the learning area exists from the first synthetic aperture radar received signal; obtaining object information of the object; and storing a group of the feature quantity of the learning signal and the object information as teacher data for learning a learning model for causing the computer to implement the following functions: outputting second object information of the object in the detection area in response to the feature quantity of the second synthetic aperture radar received signal in response to the feature quantity of the second synthetic aperture radar received signal in response to the feature quantity of the second synthetic aperture radar received signal.

[0022] Effects of the Invention

[0023] According to the present invention, it is possible to detect objects such as ground objects using RAW data. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 This is a block diagram of the observation system according to this embodiment.

[0025] Figure 2 It is a diagram for explaining the learning model involved in this embodiment.

[0026] Figure 3 This is a diagram for explaining the learning of the learning model according to this embodiment.

[0027] Figure 4 This is a flowchart illustrating an example of processing in a flying object involved in this embodiment.

[0028] Figure 5 This is a diagram illustrating an example of inference using the learning model according to this embodiment.

[0029] Figure 6This is a flowchart illustrating another example of processing in a flying object involved in this embodiment.

[0030] Figure 7 This is a diagram illustrating another example of inference using the learning model according to this embodiment.

[0031] Figure 8 This is a flowchart illustrating another example of processing in a flying object involved in this embodiment.

[0032] Figure 9 This is a diagram illustrating another example of the learning model according to this embodiment.

[0033] Figure 10 This is a diagram for explaining an example of a method for generating teacher data used for learning the learning model according to the present embodiment.

[0034] Figure 11 This is a flowchart illustrating another example of processing in a flying object involved in this embodiment.

[0035] Figure 12 This is a diagram illustrating another example of inference using the learning model according to this embodiment.

[0036] Figure 13 is a block diagram showing another embodiment of the observation system.

[0037] Explanation of symbols

[0038] 10…Observation system; 100…Flying object; 101…Communication antenna; 102…Radar device; 103…Signal processing device; 104…Storage unit; 1041…Learning model; 105…Control unit; 1051…Signal acquisition unit; 1052…Inference unit; 1053…Signal output unit; 200…Observation device; 201…Communication unit; 202…Signal processing unit. DETAILED DESCRIPTION

[0039] The preferred embodiment of the present invention will be described with reference to the accompanying drawings. In the drawings, the parts denoted by the same reference numerals have the same or similar configurations.

[0040] Figure 1is a block diagram of an observation system 10 according to this embodiment. Observation system 10 includes an aerial object 100 and an observation device 200. The aerial object 100 is located in space above the Earth, and the observation device 200 is located on the Earth. In this embodiment, the aerial object 100 observes a detection area D on the Earth's surface using radar, and an observation signal O processed by the aerial object 100 is transmitted to the observation device 200. Observation signal O is, for example, a received signal acquired by the aerial object 100 (described later), or a signal corresponding to the received signal and indicating information about an object, such as an object on the ground or a ship at sea, in the detection area D (i.e., object information).

[0041] Aircraft 100 includes a communication antenna 101, a radar device 102, and a signal processing device 103. Aircraft 100 is an artificial satellite capable of acquiring and processing received signals, located in outer space and orbiting the Earth. Alternatively, aircraft 100 may be a geostationary satellite. Furthermore, aircraft 100 may be an aircraft, helicopter, drone, or other device capable of being positioned above the Earth.

[0042] The communication antenna 101 is an antenna for communicating between the flying object 100 and an external device installed on the earth or in outer space.

[0043] Radar device 102 irradiates electromagnetic waves EM1, such as microwaves, toward a detection area D on the Earth's surface and acquires reflected electromagnetic waves EM2, which are reflected by observation objects in detection area D. Radar device 102 is, for example, a synthetic aperture radar (SAR). The reflected electromagnetic waves EM2 are processed and stored as received signals (RAW data) based on changes in the electromagnetic waves, which can be processed by aircraft 100. The received signals are stored, for example, as complex signals corresponding to the coordinates of detection area D. Radar device 102 transmits observation signals O to observation device 200 via communication antenna 101.

[0044] The radar device 102 includes a processor for controlling the acquisition process of the received signal and a storage device for storing a program required for the control.

[0045] The signal processing device 103 is an information processing device that processes the reception signal acquired by the radar device 102. The signal processing device 103 is a computer having a storage area such as a memory, and performs predetermined processing by causing a processor to execute a program stored in the storage area.

[0046] The signal processing device 103 includes a storage unit 104 and a control unit 105. The storage unit 104 is, for example, a semiconductor memory such as a RAM or an optical disk. The storage unit 104 stores various information used in the processing in the signal processing device 103.

[0047] The storage unit 104 stores a learning model 1041. The learning model 1041 is a program that learns by taking a received signal as input and outputting object information corresponding to a feature value of the received signal. Details of the feature value and the learning model 1041 in this embodiment will be described later.

[0048] The control unit 105 performs signal processing in the signal processing device 103. The control unit 105 also controls the transmission of the processing results of the received signal via the flying object 100. The control unit 105 includes a signal acquisition unit 1051, an estimation unit 1052, and a signal output unit 1053.

[0049] The signal acquisition unit 1051 acquires a received signal from the radar device 102 .

[0050] The estimation unit 1052 inputs the feature amount of the reception signal acquired by the signal acquisition unit 1051 into the learning model 1041 , and acquires the object information from the learning model 1041 .

[0051] The signal output unit 1053 outputs the object information acquired by the estimation unit 1052 as the observation signal O to the observation device 200 via the communication antenna 101. Alternatively, the signal output unit 1053 may output the received signal corresponding to the object information as the observation signal O together with the object information.

[0052] Observation device 200 transmits a control signal to flying object 100 to control observation of detection area D by flying object 100, thereby acquiring observation signal O from flying object 100. Observation device 200 includes a communication unit 201 that includes a control antenna and a control unit for antenna-based communication. Information is transmitted and received with flying object 100 via communication unit 201.

[0053] The signal processing unit 202 processes the observation signal O from the flying object 100. Based on the observation signal O acquired from the flying object 100, the signal processing unit 202 performs processing to visualize the observation result in the detection area D, for example, using an image.

[0054] Reference Figure 2 and Figure 3 , the learning of the learning model 1041 involved in this embodiment is explained.

[0055] Figure 2This figure schematically illustrates the learning and inference of the learning model 1041. The learning model 1041 learns using the learning data LD1 as teacher data. The learning data LD includes a set of feature quantities of a received signal R0 (a first synthetic aperture radar received signal) obtained by irradiating a certain area (the learning area) with electromagnetic waves, and object information TI0 (first object information) corresponding to the feature quantities of the received signal R0. The learning model 1041 learns using the feature quantities of the received signal R0 as input and the object information TI0 as output. Here, the feature quantities in this embodiment are information including the spectrum, signal strength, or phase of the received signal R0.

[0056] Figure 3 Shown are a received signal R0 and a feature amount of the received signal when a learning area D0 in which two objects (object O1 and object O2) are observed.

[0057] If learning area D0 is observed, received signal R0 is obtained. Based on received signal R0, the characteristic value (spectrum, signal strength, or phase) of received signal R0 is calculated. Information processing is performed to associate object information TI0 with the characteristic value of received signal R0, allowing the user to understand the object information TI0 corresponding to received signal R0.

[0058] For example, a curve is generated based on the feature value of received signal R0, and the user can associate the object information TI0 with the curve. Alternatively, a learning model that associates received signal R0 with object information TI0 can be used, and the object information can be associated with the feature value of received signal R0 by a computer. Furthermore, received signal R0 can be converted to information that is user-interpretable by a predetermined conversion process for SAR imaging. Alternatively, the object information TI0 associated with the feature value can be information obtained from another device observing the learning area D0 without undergoing the aforementioned imaging process. For example, in the case of ship detection, the object information TI0 can be information obtained using a ship's automatic identification system (AIS), in addition to information obtained from processing received signal R0.

[0059] The object information TI0 is information associated with the characteristic quantity (spectrum, signal strength, or phase) of the received signal R0. The received signal R0 is, for example, a complex number signal obtained as I(t)+jQ(t) (j is an imaginary unit and t is a time). When the received signal R0 is expressed in polar coordinates, it can be expressed as A(t)e jθ(t)At this point, the signal strength of received signal R0 is A(t), and the phase of received signal R0 is θ(t). The signal strength and phase are obtained for each coordinate of detection area D0. Furthermore, by performing frequency conversion on received signal R0 in the time domain, the spectrum X(ω) of received signal R0 can be obtained.

[0060] The learning model 1041 learns by outputting object information in response to input of a feature quantity calculated from the received signal R0. For example, the learning model 1041 calculates a feature quantity based on the received signal R0 in the control unit 105, uses the feature quantity as input, and outputs object information. Alternatively, the learning model 1041 can learn by inputting the received signal R0 and outputting object information. In this case, the learning model 1041 can also calculate the feature quantity based on the received signal R0.

[0061] Object information T10 includes probability distribution information about the probability of an object existing in each pixel in the learning area. Object information TI0 also includes, for example, the number of objects present in the learning area and information about their attributes. Object attributes refer to various types of object-related information, such as object type (building, mobile object, terrain, natural object, etc.), object size, and object movement speed. Object attribute information can also include the probability of an object being classified as belonging to a specific attribute.

[0062] The learning model 1041 uses the learning data LD1 prepared as described above to learn using a common machine learning method such as a neural network method. The learning model 1041 may be a single learning model or a combination of multiple learning models.

[0063] When the characteristic value of the received signal R1 (second synthetic aperture radar received signal) obtained by the flying object 100 based on the electromagnetic wave irradiated by the flying object 100 to the detection area is input into the learned learning model 1041, the learning model 1041 outputs the object information TI1 (second object information).

[0064] Reference Figure 4 and Figure 5 , the processing performed on the flying object 100 is described. Figure 5 In the processing, the object information output by the learning model 1041 is set as information corresponding to the spectrum of the received signal R1.

[0065] exist Figure 4 In step S401, the radar device 102 irradiates the electromagnetic wave EM1 with respect to the detection area D. The timing of irradiation may be a timing controlled by the observation device 200 or a timing pre-specified in the flying object 100. Figure 5 As shown, objects O3, O4, and O5 exist in the detection area D1.

[0066] In step S402 , the signal acquisition unit 1051 acquires a reception signal R1 based on the reflected electromagnetic wave EM2 detected by the radar device 102 from the radar device 102 .

[0067] In step S403, the estimation unit 1052 inputs the feature quantity of the reception signal R1 into the learning model 1041. At this time, the estimation unit 1052 performs an operation to calculate the feature quantity (for example, spectrum) of the reception signal R1.

[0068] In step S404, the estimation unit 1052 obtains the object information MD1 corresponding to the feature amount of the received signal R1 from the learning model 1041. As an example, the object information includes Figure 5 The following figure shows object information TI1a and object information TI1b. Object information TI1a is a probability distribution of the probability of objects existing in the detection area. Object information TI1a shows areas A1, A2, and A3 as areas with high probability of objects existing. Object information TI1b indicates the number of objects or their attributes in the detection area. In this example, three objects were detected.

[0069] In step S405, the signal output unit 1053 outputs an output signal based on the object information TI1 to the observation device 200. The output signal based on the object information TI1 is a signal that transmits all or part of the object information TI1. Alternatively, the output signal may be a signal that transmits information resulting from information processing performed on the object information.

[0070] Reference Figure 6 and Figure 7 , which illustrates the case where the object information output by the learning model 1041 is based on the information of the signal strength of the received signal R1. Figure 7 As shown, it is assumed that in the detection area D1 Figure 5 The same situation exists for objects O3, O4 and O5.

[0071] exist Figure 6 In step S601 , the radar device 102 irradiates the detection area D1 with electromagnetic waves EM1 . In step S602 , the signal acquisition unit 1051 acquires a reception signal R1 based on the reflected electromagnetic waves EM2 detected by the radar device 102 from the radar device 102 .

[0072] In step S603, the estimation unit 1052 inputs the signal strength of the received signal R1 as a feature quantity of the received signal R1 into the learning model 1041. The received signal R1 is, for example, a signal that can be generated. Figure 7The image IG1 shown in FIG. 1 can take various forms in response to the signal. For example, the image IG1 depicts the signal strength obtained by detecting each object, and the change in signal strength can also be expressed as a pattern. The estimation unit 1052 obtains the object information MD1 in response to the signal strength of the received signal R1 from the learning model 1041. As an example, the object information is as follows: Figure 7 As shown, Figure 5 The illustrated case is the same and includes object information TI1a and object information TI1b.

[0073] In step S605 , the signal output unit 1053 outputs an output signal based on the object information TI1 to the observation device 200 .

[0074] Reference Figure 8 , which illustrates the case where the object information output by the learning model 1041 is based on the phase information of the received signal R1. Figure 8 In step S801, the radar device 102 irradiates the detection area D1 with electromagnetic waves EM1. In step S602, the signal acquisition unit 1051 acquires a received signal R1 from the radar device 102 based on the reflected electromagnetic waves EM2 detected by the radar device 102. In step S803, the estimation unit 1052 inputs the phase of the received signal R1 into the learning model 1041. In step S804, the estimation unit 1052 acquires object information MD1 corresponding to the signal strength of the received signal R1 from the learning model 1041. In step S605, the signal output unit 1053 outputs an output signal based on the object information TI1 to the observation device 200.

[0075] As described so far, when inferring object information using the learning model 1041, the learning model 1041 learns to output object information using the spectrum, intensity, or phase as a feature quantity of the received signal as input, and inference is performed using the learning model 1041. In this case, the spectrum, intensity, or phase can be combined as the feature quantity of the received signal.

[0076] refer to Figure 9 and Figure 10 , describing other learning methods of the learning model 1041 involved in this embodiment.

[0077] Figure 9This diagram schematically illustrates the learning and inference of learning model 1041. Learning model 1041 learns using learning data LD2 as teacher data. Learning data LD2 includes a set of received signal R0a (first synthetic aperture radar received signal) obtained by irradiating a certain area (learning area) with electromagnetic waves, whose feature quantity satisfies a predetermined condition, and object information TI0 (first object information) corresponding to the feature quantity of received signal R0a.

[0078] The specified conditions for feature quantities are conditions for extracting received signals containing feature quantities based on objects within the learning area and can be changed depending on the purpose of observation, etc. For example, when extracting received signal R0a from received signal R0, it is possible to extract received signals whose signal strength meets the specified conditions. For example, the specified condition for signal strength is that the signal strength is greater than a specified threshold. Alternatively, for example, the specified condition for spectrum is that the frequency is within a specified frequency band. Furthermore, for example, the specified condition for phase is that the phase is within a specified range.

[0079] The learning model 1041 performs learning by taking the feature amount of the reception signal R0a as input and outputting the object information TI0 as output.

[0080] refer to Figure 10 , explaining the generation of learning data LD2. Figure 10 In the example of FIG. 1 , a case is shown where the learning data LD2 is generated using the observation results of the learning area D0 where two objects (object O1 and object O2) exist.

[0081] When learning area D0 is observed, received signal R0 is obtained. When received signal R0 is visualized by plotting it as a curve or converting it into a SAR image, for example, an image IG0 is obtained. Image IG0 can be correlated with the signal strength of received signal R0. Within received signal R0, received signals with signal strengths greater than a predetermined threshold are extracted. For example, received signal R0a is extracted, including received signal R0a1 corresponding to a portion of image IG0 and received signal R0a2 corresponding to another portion of image IG0. Visualizing received signal R0a1 yields image IG2, while visualizing received signal R0a2 yields image IG3. Images IG2 and IG3 can be correlated with the signal strengths of received signals R0a1 and R0a2, respectively.

[0082] Target information TI0 based on image IG0 is associated with the feature values ​​of each received signal R0a, including received signal R0a1 and received signal R0a2. As an example, object information includes object information TI0a and object information TI0b. Object information TI0a is a probability distribution of the probability of objects existing in the detection area. In object information TI0a, areas A4 and A5 are indicated as areas with a high probability of objects existing. Object information TI0b is the number of objects or their attributes in the detection area. In this example, two objects are detected. Learning data LD2 is generated as a collection of received signal R0a and object information TI0.

[0083] The learning model 1041 is learned by a general machine learning method such as a neural network method using the learning data LD2 prepared as described above.

[0084] Reference Figure 11 and Figure 12 , describing the processing performed by flying object 100.

[0085] exist Figure 11 In step S1101, the radar device 102 irradiates the detection area D1 with electromagnetic waves EM1. Figure 12 As shown, objects O3, O4, and O5 exist in the detection area D1.

[0086] In step S1102 , the signal acquisition unit 1051 acquires, from the radar device 102 , a reception signal R1 based on the reflected electromagnetic wave EM2 detected by the radar device 102 .

[0087] In step S1103, the estimation unit 1052 extracts the received signal R1a that satisfies the predetermined signal strength condition from the received signal R1. The extracted feature quantity of the received signal R1a is, for example, Figure 12 As shown, it corresponds to an image IG4 obtained by cutting out a portion of the image IG1 based on the reception signal R1.

[0088] In step S1104 , the estimation unit 1052 inputs the reception signal R1 a to the learning model 1041 .

[0089] In step S1105, the estimation unit 1052 obtains the object information MD1 based on the feature value of the received signal R1a from the learning model 1041. As an example, the object information includes Figure 12 The object information TI1a and object information TI1b are shown. Figure 5Similarly, as shown in FIG, areas A1, A2, and A3 are shown as areas with a high probability of the presence of objects. Object information TI1b is the number of objects in the detection area or the attributes of the objects. In this example, three objects are detected.

[0090] In step S1106 , the signal output unit 1053 outputs an output signal based on the object information TI1 to the observation device 200 .

[0091] By being able to output object information based on the feature quantities of all received signals based on the feature quantities of a portion of the received signals, the amount of data used as input for inference by learning model 1041 can be reduced. Furthermore, since the computation time using learning model 1041 can be shortened, the real-time nature of observations can be improved. Furthermore, since the computational load of learning model 1041 can be reduced, object information can be acquired using a more lightweight processing device.

[0092] Figure 13 FIG. 1 shows a block diagram of an observation system 10A as another embodiment. Figure 13 As shown, the signal processing device 103 may be included in the observation device 200A. The control unit 1301 of the aircraft 100A transmits the received signal acquired by the radar device 102 to the observation device 200A. The above-mentioned processing may also be performed by the signal processing device 103 of the observation device 200A.

[0093] The embodiments described above are intended to facilitate understanding of the present invention and are not intended to limit the present invention. The various elements and conditions of the embodiments are not limited to the elements shown in the examples and may be appropriately changed. In addition, different structures may be partially replaced or combined.

Claims

1. A learning model, Performing learning using teacher data, the teacher data being data that takes as input a feature amount of a first synthetic aperture radar reception signal based on a reflected electromagnetic wave resulting from an electromagnetic wave irradiated onto a learning area and reflects and outputs first object information of an object within the learning area; and The computer is caused to realize the function of outputting second object information of an object in the detection area in response to input of a feature amount of a second synthetic aperture radar reception signal based on a reflected electromagnetic wave obtained by reflecting the electromagnetic wave irradiated into the detection area.

2. The learning model according to claim 1, wherein The characteristic quantity of the first synthetic aperture radar received signal includes the frequency spectrum of the first synthetic aperture radar received signal. The characteristic quantity of the second synthetic aperture radar received signal includes a frequency spectrum of the second synthetic aperture radar received signal.

3. The learning model according to claim 1, wherein: The characteristic quantity of the first synthetic aperture radar received signal includes the signal strength of the first synthetic aperture radar received signal. The characteristic quantity of the second synthetic aperture radar received signal includes the signal strength of the second synthetic aperture radar received signal.

4. The learning model according to claim 1, wherein: The characteristic quantity of the first synthetic aperture radar received signal includes the phase of the first synthetic aperture radar received signal. The characteristic quantity of the second synthetic aperture radar received signal includes a phase of the second synthetic aperture radar received signal.

5. The learning model according to claim 1, wherein: The first synthetic aperture radar received signal is a signal corresponding to a portion of the learning area where the feature quantity satisfies a predetermined condition. The feature quantity of the first synthetic aperture radar received signal includes a feature quantity based on an object in the learning area.

6. The learning model according to claim 5, wherein: The second synthetic aperture radar received signal is a signal corresponding to a portion of the detection area where the characteristic quantity satisfies the predetermined condition. The feature quantity of the second synthetic aperture radar received signal includes a feature quantity based on an object in the detection area.

7. The learning model according to claim 1, wherein: The first object information includes information indicating attributes of the object in the learning area. The second object information includes information indicating attributes of the object in the detection area.

8. A signal processing device comprising: a storage unit storing the learning model according to any one of claims 1 to 7; a signal acquiring unit, configured to acquire the second synthetic aperture radar received signal; and The inference unit inputs the second synthetic aperture radar reception signal into the learning model and infers the second object information.

9. A flying object having: a storage unit storing the learning model according to any one of claims 1 to 7; a signal acquisition unit, configured to acquire the second synthetic aperture radar received signal; an inference unit that inputs the second synthetic aperture radar received signal into the learning model and infers the second object information; and The signal output unit outputs an output signal based on the second object information to the outside.

10. A method for generating teacher data, comprising causing a computer to perform the following operations: acquiring a first synthetic aperture radar received signal based on a reflected electromagnetic wave resulting from reflection of the electromagnetic wave irradiated onto the learning area; extracting, from the first synthetic aperture radar received signal, a learning signal corresponding to an area in the learning area where the object in the learning area exists; acquiring object information of the object; The group of the characteristic quantity of the learning signal and the object information is stored as teacher data for learning a learning model for enabling a computer to implement the following functions: in response to the input of the characteristic quantity of a second synthetic aperture radar receiving signal of a reflected electromagnetic wave based on the reflection of the electromagnetic wave irradiated to the detection area, second object information of the object in the detection area is output in response to the characteristic quantity of the second synthetic aperture radar receiving signal.

Images